Improvements¶

• [logical_range_filter] Improved search plan for large data.

  • Normally, logical_range_filter is faster than logical_select. However, it had been slower than logical_select in the below case.

    • If Groonga can’t get the number of required records easily, it has the feature that switches index search from sequential search. (Normally, logical_range_filter uses a sequential search when records of search target are many.)

    • The search process for it is almost the same as logical_select if the above switching occurred. So, logical_range_filter is severalfold slower than logical_select in the above case if the search target is large data. Because logical_range_filter executes sort after the search.

  • If we search for large data, Groonga easily use sequential search than until now since this release.

  • Therefore, logical_range_filter will improve performance. Because the case of the search process almost the same as logical_select decreases.

• [httpd] Updated bundled nginx to 1.19.1.

• Modify how to install into Debian GNU/Linux.

  • We modify to use groonga-apt-source instead of groonga-archive-keyring. Because the lintian command recommends using apt-source if a package that it puts files under the /etc/apt/sources.lists.d/.

    • The lintian command is the command which checks for many common packaging errors.

    • Please also refer to the following URL for the details about installation procedures.

      • https://groonga.org/docs/install/debian.html

• [logical_select] Added a support for highlight_html and highlight_full .

• Added support for recycling the IDs of records that are deleted when an array without value space delete.[GitHub#mroonga/mroonga#327][Reported by gaeeyo]

  • If an array that doesn’t have value space is deleted, deleted IDs are never recycled.

  • Groonga had used large storage space by large ID. Because it uses large storage space by itself.

    • For example, large ID is caused after many adds and deletes like Mroonga’s mroonga_operations

• [select] Improved performance of full-text-search without index.

• [Function] Improved performance for calling of function that all arguments a variable reference or literal.

• [Indexing] Improved performance of offline index construction by using token column. [GitHub#1126][Patched by naoa]

• Improved performance for "_score = func(...)".

  • The performance when the _score value calculate by using only function like "_score = func(...)" improved.

Fixes¶

• Fixed a bug that garbage may be included in response after response send error.

  • It may occur if a client didn’t read all responses and closed the connection.

Thanks¶

• gaeeyo

• naoa

Improvements¶

• [select] Added support for storing reference in table that we specify with --load_table.

  • --load_table is a feature that stores search results to the table in a prepared.

    • If the searches are executed multiple times, we can cache the search results by storing them to this table.

    • We can shorten the search times that the search after the first time by using this table.

  • We can store a reference to other tables into the key of this table as below since this release.

    • We can make a smaller size of this table. Because we only store references without store column values.

    • If we search against this table, we can search by using indexes for reference destination.

    table_create Logs TABLE_HASH_KEY ShortText
    column_create Logs timestamp COLUMN_SCALAR Time
    
    table_create Times TABLE_PAT_KEY Time
    column_create Times logs_timestamp COLUMN_INDEX Logs timestamp
    
    table_create LoadedLogs TABLE_HASH_KEY Logs
    
    load --table Logs
    [
    {
      "_key": "2015-02-03:1",
      "timestamp": "2015-02-03 10:49:00"
    },
    {
      "_key": "2015-02-03:2",
      "timestamp": "2015-02-03 12:49:00"
    },
    {
      "_key": "2015-02-04:1",
      "timestamp": "2015-02-04 00:00:00"
    }
    ]
    
    select \
      Logs \
      --load_table LoadedLogs \
      --load_columns "_key" \
      --load_values "_key" \
      --limit 0
    
    select \
      --table LoadedLogs \
      --filter 'timestamp >= "2015-02-03 12:49:00"'
    [
      [
        0,
        0.0,
        0.0
      ],
      [
        [
          [
            2
          ],
          [
            [
              "_id",
              "UInt32"
            ],
            [
              "_key",
              "ShortText"
            ],
            [
              "timestamp",
              "Time"
            ]
          ],
          [
            2,
            "2015-02-03:2",
            1422935340.0
          ],
          [
            3,
            "2015-02-04:1",
            1422975600.0
          ]
        ]
      ]
    ]
    

• [select] Improved sort performance on below cases.

  • When many sort keys need ID resolution.

    • For example, the following expression needs ID resolution.

      • --filter true --sort_keys column

    • For example, the following expression doesn’t need ID resolution. Because the _score pseudo column exists in the result table not the source table.

      • --filter true --sort_keys _score

  • When a sort target table has a key.

    • Therefore, TABLE_NO_KEY isn’t supported this improvement.

• [select] Improved performance a bit on below cases.

  • A case of searching for many records matches.

  • A case of drilldown for many records.

• [aggregator] Added support for score accessor for target. [GitHub#1120][Patched by naoa]

  • For example, we can _score subject to aggregator_* as below.

    table_create Items TABLE_HASH_KEY ShortText
    column_create Items price COLUMN_SCALAR UInt32
    column_create Items tag COLUMN_SCALAR ShortText
    
    load --table Items
    [
    {"_key": "Book",  "price": 1000, "tag": "A"},
    {"_key": "Note",  "price": 1000, "tag": "B"},
    {"_key": "Box",   "price": 500,  "tag": "B"},
    {"_key": "Pen",   "price": 500,  "tag": "A"},
    {"_key": "Food",  "price": 500,  "tag": "C"},
    {"_key": "Drink", "price": 300,  "tag": "B"}
    ]
    
    select Items \
      --filter true \
      --drilldowns[tag].keys tag \
      --drilldowns[tag].output_columns _key,_nsubrecs,score_mean \
      --drilldowns[tag].columns[score_mean].stage group \
      --drilldowns[tag].columns[score_mean].type Float \
      --drilldowns[tag].columns[score_mean].flags COLUMN_SCALAR \
      --drilldowns[tag].columns[score_mean].value 'aggregator_mean(_score)'
    [
      [
        0,
        0.0,
        0.0
      ],
      [
        [
          [
            6
          ],
          [
            [
              "_id",
              "UInt32"
            ],
            [
              "_key",
              "ShortText"
            ],
            [
              "price",
              "UInt32"
            ],
            [
              "tag",
              "ShortText"
            ]
          ],
          [
            1,
            "Book",
            1000,
            "A"
          ],
          [
            2,
            "Note",
            1000,
            "B"
          ],
          [
            3,
            "Box",
            500,
            "B"
          ],
          [
            4,
            "Pen",
            500,
            "A"
          ],
          [
            5,
            "Food",
            500,
            "C"
          ],
          [
            6,
            "Drink",
            300,
            "B"
          ]
        ],
        {
          "tag": [
            [
              3
            ],
            [
              [
                "_key",
                "ShortText"
              ],
              [
                "_nsubrecs",
                "Int32"
              ],
              [
                "score_mean",
                "Float"
              ]
            ],
            [
              "A",
              2,
              1.0
            ],
            [
              "B",
              3,
              1.0
            ],
            [
              "C",
              1,
              1.0
            ]
          ]
        }
      ]
    ]
    

• [Indexing] Improved performance of offline index construction on VC++ version.

• [select] Use null instead NaN, Infinity, and -Infinity when Groonga outputs result for JSON format.

  • Because JSON doesn’t support them.

• [select] Add support fot aggregating standard deviation value.

  • For example, we can calculate a standard deviation for every group as below.

    table_create Items TABLE_HASH_KEY ShortText
    column_create Items price COLUMN_SCALAR UInt32
    column_create Items tag COLUMN_SCALAR ShortText
    
    load --table Items
    [
    {"_key": "Book",  "price": 1000, "tag": "A"},
    {"_key": "Note",  "price": 1000, "tag": "B"},
    {"_key": "Box",   "price": 500,  "tag": "B"},
    {"_key": "Pen",   "price": 500,  "tag": "A"},
    {"_key": "Food",  "price": 500,  "tag": "C"},
    {"_key": "Drink", "price": 300,  "tag": "B"}
    ]
    
    select Items \
      --drilldowns[tag].keys tag \
      --drilldowns[tag].output_columns _key,_nsubrecs,price_sd \
      --drilldowns[tag].columns[price_sd].stage group \
      --drilldowns[tag].columns[price_sd].type Float \
      --drilldowns[tag].columns[price_sd].flags COLUMN_SCALAR \
      --drilldowns[tag].columns[price_sd].value 'aggregator_sd(price)' \
      --output_pretty yes
    [
      [
        0,
        1594339851.924836,
        0.002813816070556641
      ],
      [
        [
          [
            6
          ],
          [
            [
              "_id",
              "UInt32"
            ],
            [
              "_key",
              "ShortText"
            ],
            [
              "price",
              "UInt32"
            ],
            [
              "tag",
              "ShortText"
            ]
          ],
          [
            1,
            "Book",
            1000,
            "A"
          ],
          [
            2,
            "Note",
            1000,
            "B"
          ],
          [
            3,
            "Box",
            500,
            "B"
          ],
          [
            4,
            "Pen",
            500,
            "A"
          ],
          [
            5,
            "Food",
            500,
            "C"
          ],
          [
            6,
            "Drink",
            300,
            "B"
          ]
        ],
        {
          "tag": [
            [
              3
            ],
            [
              [
                "_key",
                "ShortText"
              ],
              [
                "_nsubrecs",
                "Int32"
              ],
              [
                "price_sd",
                "Float"
              ]
            ],
            [
              "A",
              2,
              250.0
            ],
            [
              "B",
              3,
              294.3920288775949
            ],
            [
              "C",
              1,
              0.0
            ]
          ]
        }
      ]
    ]
    

    • We can also calculate sample standard deviation to specifing aggregate_sd(target, {"unbiased": true}).

• [Windows] Dropped Visual Studio 2013 support.

Fixes¶

• [Groonga HTTP server] Fixed a bug that a request can’t halt even if we execute shutdown?mode=immediate when the response was halted by error occurrence.

• Fixed a crash bug when an error occurs while a request.

  • It only occurs when we use Apache Arrow Format.

  • Groonga crashes when we send request to Groonga again after the previous request was halted by error occurrence.

• [between] Fixed a crash bug when temporary table is used.

  • For example, if we specify a dynamic column in the first argument for between, Groonga had crashed.

• Fixed a bug that procedure created by plugin is freed unexpectedly.

  • It’s only occurred in reference count mode.

  • It’s not occurred if we don’t use plugin_register.

  • It’s not occurred in the process that executes plugin_register.

  • It’s occurred in the process that doesn’t execute plugin_register.

• Fixed a bug that normalization error occurred while static index construction by token_column. [GitHub#1122][Reported by naoa]

Thanks¶

• naoa

Improvements¶

• [Tables] Added support for registering 400M records into a hash table.

• [select] Improve scorer performance when the _score doesn’t get recursively values.

  • Groonga get recursively value of _score when search result is search target.

  • For example, the search targets of slices are search result. Therefore, if we use slice in a query, this improvement doesn’t ineffective.

• [Log] Improved that we output drilldown keys in query-log.

• [reference_acquire], [reference_release] Added new commands for reference count mode.

  • If we need to call multiple load in a short time, auto close by the reference count mode will degrade performance.

  • We can avoid the performance degrading by calling reference_acquire before multiple load and calling reference_release after multiple load. Between reference_acquire and reference_release, auto close is disabled.

    • Because reference_acquire acquires a reference of target objects.

  • We can must call reference_release after you finish performance impact operations.

  • If we don’t call reference_release, the reference count mode doesn’t work.

• [select] Added support for aggregating multiple groups on one time drilldown.

  • We came to be able to calculate sum or arithmetic mean every different multiple groups on one time drilldown as below.

    table_create Items TABLE_HASH_KEY ShortText
    column_create Items price COLUMN_SCALAR UInt32
    column_create Items quantity COLUMN_SCALAR UInt32
    column_create Items tag COLUMN_SCALAR ShortText
    
    load --table Items
    [
    {"_key": "Book",  "price": 1000, "quantity": 100, "tag": "A"},
    {"_key": "Note",  "price": 1000, "quantity": 10,  "tag": "B"},
    {"_key": "Box",   "price": 500,  "quantity": 15,  "tag": "B"},
    {"_key": "Pen",   "price": 500,  "quantity": 12,  "tag": "A"},
    {"_key": "Food",  "price": 500,  "quantity": 111, "tag": "C"},
    {"_key": "Drink", "price": 300,  "quantity": 22,  "tag": "B"}
    ]
    
    select Items \
      --drilldowns[tag].keys tag \
      --drilldowns[tag].output_columns _key,_nsubrecs,price_sum,quantity_sum \
      --drilldowns[tag].columns[price_sum].stage group \
      --drilldowns[tag].columns[price_sum].type UInt32 \
      --drilldowns[tag].columns[price_sum].flags COLUMN_SCALAR \
      --drilldowns[tag].columns[price_sum].value 'aggregator_sum(price)' \
      --drilldowns[tag].columns[quantity_sum].stage group \
      --drilldowns[tag].columns[quantity_sum].type UInt32 \
      --drilldowns[tag].columns[quantity_sum].flags COLUMN_SCALAR \
      --drilldowns[tag].columns[quantity_sum].value 'aggregator_sum(quantity)'
    [
      [
        0,
        0.0,
        0.0
      ],
      [
        [
          [
            6
          ],
          [
            [
              "_id",
              "UInt32"
            ],
            [
              "_key",
              "ShortText"
            ],
            [
              "price",
              "UInt32"
            ],
            [
              "quantity",
              "UInt32"
            ],
            [
              "tag",
              "ShortText"
            ]
          ],
          [
            1,
            "Book",
            1000,
            100,
            "A"
          ],
          [
            2,
            "Note",
            1000,
            10,
            "B"
          ],
          [
            3,
            "Box",
            500,
            15,
            "B"
          ],
          [
            4,
            "Pen",
            500,
            12,
            "A"
          ],
          [
            5,
            "Food",
            500,
            111,
            "C"
          ],
          [
            6,
            "Drink",
            300,
            22,
            "B"
          ]
        ],
        {
          "tag": [
            [
              3
            ],
            [
              [
                "_key",
                "ShortText"
              ],
              [
                "_nsubrecs",
                "Int32"
              ],
              [
                "price_sum",
                "UInt32"
              ],
              [
                "quantity_sum",
                "UInt32"
              ]
            ],
            [
              "A",
              2,
              1500,
              112
            ],
            [
              "B",
              3,
              1800,
              47
            ],
            [
              "C",
              1,
              500,
              111
            ]
          ]
        }
      ]
    ]
    

• [groonga executable file] Added support for --pid-path in standalone mode.

  • Because --pid-path had been ignored in standalone mode in before version.

• [io_flush] Added support for reference count mode.

• [logical_range_filter], [logical_count] Added support for reference count mode.

• [Groonga HTTP server] We didn’t add header after the last chunk.

  • Because there is a possibility to exist that the HTTP client ignores header after the last chunk.

• [vector_slice] Added support for a vector that has the value of the Float32 type. [GitHub#1112 patched by naoa]

• Added support for parallel offline index construction using token column.

  • We came to be able to construct an offline index on parallel threads from data that are tokenized in advance.

  • We can tune parameters of parallel offline construction by the following environment variables

    • GRN_TOKEN_COLUMN_PARALLEL_CHUNK_SIZE : We specify how many records are processed per thread.

      • The default value is 1024 records.

    • GRN_TOKEN_COLUMN_PARALLEL_TABLE_SIZE_THRESHOLD : We specify how many source records are required for parallel offline construction.

      • The default value is 102400 records.

• [select] Improved performance for load_table on the reference count mode.

Fixes¶

• Fixed a bug that the database of Groonga was broken when we search by using the dynamic columns that don’t specify a --filter and stridden over shard.

• Fixed a bug that Float32 type had not displayed on a result of schema command.

• Fixed a bug that we count in surplus to _nsubrecs when the reference uvector hasn’t element.

Thanks¶

• naoa

Improvements¶

• We came to be able to construct an inverted index from data that are tokenized in advance.

  • The construct of an index is speeded up from this.

  • We need to prepare token column to use this improvement.

  • token column is an auto generated value column like an index column.

  • token column value is generated from source column value by tokenizing the source column value.

  • We can create a token column by setting the source column as below.

    table_create Terms TABLE_PAT_KEY ShortText \
      --normalizer NormalizerNFKC121 \
      --default_tokenizer TokenNgram
    
    table_create Notes TABLE_NO_KEY
    column_create Notes title COLUMN_SCALAR Text
    
    # The last "title" is the source column.
    column_create Notes title_terms COLUMN_VECTOR Terms title
    

• [select] We came to be able to specify a vector for the argument of a function.

  • For example, flags options of query can describe by a vector as below.

    select \
      --table Memos \
      --filter 'query("content", "-content:@mroonga", \
                      { \
                        "expander": "QueryExpanderTSV", \
                        "flags": ["ALLOW_LEADING_NOT", "ALLOW_COLUMN"] \
                      })'
    

• [select] Added a new stage result_set for dynamic columns.

  • This stage generates a column into a result set table. Therefore, it is not generated if query or filter doesn’t exist

    • Because if query or filter doesn’t exist, Groonga doesn’t make a result set table.

  • We can’t use _value for the stage. The result_set stage is for storing value by score_column.

• [vector_slice] Added support for weight vector that has weight of Float32 type. [GitHub#1106 patched by naoa]

• [select] Added support for filtered stage and output stage of dynamic columns on drilldowns. [GitHub#1101 patched by naoa][GitHub#1100 patched by naoa]

  • We can use filtered and output stage of dynamic columns on drilldowns as with drilldowns[Label].stage filtered and drilldowns[Label].stage output.

• [select] Added support for Float type value in aggregating on drilldown.

  • We can aggregate max value, min value, and sum value for Float type value using MAX, MIN, and SUM.

• [query] [geo_in_rectangle] [geo_in_circle] Added a new option score_column for query(), geo_in_rectangle(), and geo_in_circle().

  • We can store a score value by condition using score_column.

  • Normally, Groonga calculate a score by adding scores of all conditions. However, we sometimes want to get a score value by condition.

  • For example, if we want to only use how near central coordinate as score as below, we use score_column.

  table_create LandMarks TABLE_NO_KEY
  column_create LandMarks name COLUMN_SCALAR ShortText
  column_create LandMarks category COLUMN_SCALAR ShortText
  column_create LandMarks point COLUMN_SCALAR WGS84GeoPoint
  
  table_create Points TABLE_PAT_KEY WGS84GeoPoint
  column_create Points land_mark_index COLUMN_INDEX LandMarks point
  
  load --table LandMarks
  [
    {"name": "Aries"      , "category": "Tower"     , "point": "11x11"},
    {"name": "Taurus"     , "category": "Lighthouse", "point": "9x10" },
    {"name": "Gemini"     , "category": "Lighthouse", "point": "8x8"  },
    {"name": "Cancer"     , "category": "Tower"     , "point": "12x12"},
    {"name": "Leo"        , "category": "Tower"     , "point": "11x13"},
    {"name": "Virgo"      , "category": "Temple"    , "point": "22x10"},
    {"name": "Libra"      , "category": "Tower"     , "point": "14x14"},
    {"name": "Scorpio"    , "category": "Temple"    , "point": "21x9" },
    {"name": "Sagittarius", "category": "Temple"    , "point": "43x12"},
    {"name": "Capricorn"  , "category": "Tower"     , "point": "33x12"},
    {"name": "Aquarius"   , "category": "mountain"  , "point": "55x11"},
    {"name": "Pisces"     , "category": "Tower"     , "point": "9x9"  },
    {"name": "Ophiuchus"  , "category": "mountain"  , "point": "21x21"}
  ]
  
  select LandMarks \
    --sort_keys 'distance' \
    --columns[distance].stage initial \
    --columns[distance].type Float \
    --columns[distance].flags COLUMN_SCALAR \
    --columns[distance].value 0.0 \
    --output_columns 'name, category, point, distance, _score' \
    --limit -1 \
    --filter 'geo_in_circle(point, "11x11", "11x1", {"score_column": distance}) && category == "Tower"'
  [
    [
      0,
      1590647445.406149,
      0.0002503395080566406
    ],
    [
      [
        [
          5
        ],
        [
          [
            "name",
            "ShortText"
          ],
          [
            "category","ShortText"
          ],
          [
            "point",
            "WGS84GeoPoint"
          ],
          [
            "distance",
            "Float"
          ],
          [
            "_score",
            "Int32"
          ]
        ],
        [
          "Aries",
          "Tower",
          "11x11",
          0.0,
          1
        ],
        [
          "Cancer",
          "Tower",
          "12x12",
          0.0435875803232193,
          1
        ],
        [
          "Leo",
          "Tower",
          "11x13",
          0.06164214760065079,
          1
        ],
        [
          "Pisces",
          "Tower",
          "9x9",
          0.0871751606464386,
          1
        ],
        [
          "Libra",
          "Tower",
          "14x14",
          0.1307627409696579,
          1
        ]
      ]
    ]
  ]
  

  • The sort by _score is meaningless in the above example. Because the value of _score is all 1 by category == "Tower". However, we can sort distance from central coordinate using score_column.

• [Windows] Groonga came to be able to output backtrace when it occurs error even if it doesn’t crash.

• [Windows] Dropped support for old Windows.

  • Groonga for Windows come to require Windows 8 (Windows Server 2012) or later from 10.0.3.

• [select] Improved sort performance when sort keys were mixed referable sort keys and the other sort keys.

  • We improved sort performance if mixed referable sort keys and the other and there are referable keys two or more.

    • Referable sort keys are sort keys that except below them.

      • Compressed columns

      • _value against the result of drilldown that is specified multiple values to the key of drilldown.

      • _key against patricia trie table that has not the key of ShortText type.

      • _score

  • The more sort keys that except string, a decrease in the usage of memory for sort.

• [select] Improved sort performance when sort keys are all referable keys case.

• [select] Improve scorer performance as a _socre = column1*X + column2*Y + ... case.

  • This optimization effective when there are many + or * in _score.

  • At the moment, it has only effective against + and *.

• [select] Added support for phrase near search.

  • We can search phrase by phrase by a near search.

    • Query syntax for near phrase search is *NP"Phrase1 phrase2 ...".

    • Script syntax for near phrase search is column *NP "phrase1 phrase2 ...".

    • If the search target phrase includes space, we can search for it by surrounding it with " as below.

      table_create Entries TABLE_NO_KEY
      column_create Entries content COLUMN_SCALAR Text
      
      table_create Terms TABLE_PAT_KEY ShortText \
        --default_tokenizer 'TokenNgram("unify_alphabet", false, \
                                        "unify_digit", false)' \
        --normalizer NormalizerNFKC121
      column_create Terms entries_content COLUMN_INDEX|WITH_POSITION Entries content
      
      load --table Entries
      [
      {"content": "I started to use Groonga. It's very fast!"},
      {"content": "I also started to use Groonga. It's also very fast! Really fast!"}
      ]
      
      select Entries --filter 'content *NP "\\"I started\\" \\"use Groonga\\""' --output_columns 'content'
      [
        [
          0,
          1590469700.715882,
          0.03997230529785156
        ],
        [
          [
            [
              1
            ],
            [
              [
                "content",
                "Text"
              ]
            ],
            [
              "I started to use Groonga. It's very fast!"
            ]
          ]
        ]
      ]
      

• [Vector column] Added support for float32 weight vector.

  • We can store weight as float32 instead of uint32.

  • We need to add WEIGHT_FLOAT32 flag when execute column_create to use this feature.

    column_create Records tags COLUMN_VECTOR|WITH_WEIGHT|WEIGHT_FLOAT32 Tags
    

  • However, WEIGHT_FLOAT32 flag isn’t available with COLUMN_INDEX flag for now.

• Added following APIs

  • Added grn_obj_is_xxx functions. For more information as below.

    • grn_obj_is_weight_vector(grn_ctx *ctx, grn_obj *obj)

      • It returns as a bool whether the object is a weight vector.

    • grn_obj_is_uvector(grn_ctx *ctx, grn_obj *obj)

      • It returns as a bool whether the object is a uvector.

        • uvector is a vector that size of elements for vector are fixed.

    • grn_obj_is_weight_uvector(grn_ctx *ctx, grn_obj *obj)

      • It returns as a bool whether the object is a weight uvector.

  • Added grn_type_id_size(grn_ctx *ctx, grn_id id).

    • It returns the size of Groonga data type as a size_t.

  • Added grn_selector_data_get_xxx functions. For more information as below.

    • These functions return selector related data.

      • These functions are supposed to call in selector. If they are called except in selector, they return NULL.

        • grn_selector_data_get(grn_ctx *ctx)

          • It returns all information that relating calling selector as grn_selector_data * structure.

        • grn_selector_data_get_selector(grn_ctx *ctx, grn_selector_data *data)

          • It returns selector itself as grn_obj *.

        • grn_selector_data_get_expr(grn_ctx *ctx, grn_selector_data *data)

          • It returns selector is used --filter condition and --query condition as grn_obj *.

        • grn_selector_data_get_table(grn_ctx *ctx, grn_selector_data *data)

          • It returns target table as grn_obj *

        • grn_selector_data_get_index(grn_ctx *ctx, grn_selector_data *data)

          • It returns index is used by selector as grn_obj *.

        • grn_selector_data_get_args(grn_ctx *ctx, grn_selector_data *data, size_t *n_args)

          • It returns arguments of function that called selector as grn_obj **.

        • grn_selector_data_get_result_set(grn_ctx *ctx, grn_selector_data *data)

          • It returns result table as grn_obj *.

        • grn_selector_data_get_op(grn_ctx *ctx, grn_selector_data *data)

          • It returns how to perform the set operation on existing result set as grn_operator.

  • Added grn_plugin_proc_xxx functions. For more information as below.

    • grn_plugin_proc_get_value_operator(grn_ctx *ctx, grn_obj *value, grn_operator default_operator, const char *context)

      • It returns the operator of a query as a grn_operator.

        • For example, && is returned as a GRN_OP_AND.

    • grn_plugin_proc_get_value_bool(grn_ctx *ctx, grn_obj *value, bool default_value, const char *tag)

      • It returns the value that is specified true or false like with_transposition argument of the below function as a bool (bool is the data type of C language).

        fuzzy_search(column, query, {"max_distance": 1, "prefix_length": 0, "max_expansion": 0, "with_transposition": true})
        

  • Added grn_proc_options_xxx functions. For more information as below.

    • query() only uses them for now.

      • grn_proc_options_parsev(grn_ctx *ctx, grn_obj *options, const char *tag, const char *name, va_list args)

        • This function execute parse options.

        • We had to implement parsing to options ourselves until now, however, we can parse them by just call this function from this version.

      • grn_proc_options_parse(grn_ctx *ctx, grn_obj *options, const char *tag, const char *name, ...)

        • It calls grn_proc_options_parsev(). Therefore, features of this function same grn_proc_options_parsev().

        • It only differs in the interface compare with grn_proc_options_parsev().

  • Added grn_text_printfv(grn_ctx *ctx, grn_obj *bulk, const char *format, va_list args)

    • grn_text_vprintf is deprecated from 10.0.3. We use grn_text_printfv instead.

  • Added grn_type_id_is_float_family(grn_ctx *ctx, grn_id id).

    • It returns whether grn_type_id is GRN_DB_FLOAT32 or GRN_DB_FLOAT or not as a bool.

  • Added grn_dat_cursor_get_max_n_records(grn_ctx *ctx, grn_dat_cursor *c).

    • It returns the number of max records the cursor can have as a size_t. (This API is for the DAT table)

  • Added grn_table_cursor_get_max_n_records(grn_ctx *ctx, grn_table_cursor *cursor).

    • It returns the number of max records the cursor can have as a size_t.

    • It can use against all table type (TABLE_NO_KEY, TABLE_HASH_KEY, TABLE_DAT_KEY, and TABLE_PAT_KEY).

  • Added grn_result_set_add_xxx functions. For more information as below.

    • grn_result_set_add_record(grn_ctx *ctx, grn_hash *result_set, grn_posting *posting, grn_operator op)

      • It adds a record into the table of result sets.

      • grn_ii_posting_add_float is deprecated from 10.0.3. We use grn_rset_add_records() instead.

    • grn_result_set_add_table(grn_ctx *ctx, grn_hash *result_set, grn_obj *table, double score, grn_operator op)

      • It adds a table into the result sets.

    • grn_result_set_add_table_cursor(grn_ctx *ctx, grn_hash *result_set, grn_table_cursor *cursor, double score, grn_operator op)

      • It adds records that a table cursor has into the result sets.

  • Added grn_vector_copy(grn_ctx *ctx, grn_obj *src, grn_obj *dest).

    • It copies a vector object. It returns whether success copy a vector object.

  • Added grn_obj_have_source(grn_ctx *ctx, grn_obj *obj).

    • It returns whether the column has a source column as a bool.

  • Added grn_obj_is_token_column(grn_ctx *ctx, grn_obj *obj).

    • It returns whether the column is a token column as a bool.

  • Added grn_hash_add_table_cursor(grn_ctx *ctx, grn_hash *hash, grn_table_cursor *cursor, double score).

    • It’s for bulk result set insert. It’s faster than inserting records by grn_ii_posting_add().

Fixes¶

• Fixed a crash bug if the modules (tokenizers, normalizers, and token filters) are used at the same time from multiple threads.

• Fixed precision of Float32 value when it outputted.

  • The precision of it changes to 8-digit to 7-digit from 10.0.3.

• Fixed a bug that Groonga used the wrong cache when the query that just the parameters of dynamic column different was executed. [GitHub#1102 patched by naoa]

Thanks¶

• naoa

Improvements¶

• Added support for uvector for time_classify_* functions. [GitHub#1089][Patched by naoa]

  • uvector is a vector that size of elements for vector are fixed.

  • For example, a vector that has values of Time type as elements is a uvector.

• Improve sort performance if sort key that can’t refer value with zero-copy is mixed.

  • Some sort key (e.g. _score) values can’t be referred with zero-copy.

  • If there is at least one sort key that can’t be referable is included, all sort keys are copied before.

  • With this change, we just copy sort keys that can’t be referred. Referable sort keys are just referred without a copy.

  • However, this change may have performance regression when all sort keys are referable.

• Added support for loading weight vector as a JSON string.

  • We can load weight vector as a JSON string as below example.

    table_create Tags TABLE_PAT_KEY ShortText
    table_create Data TABLE_NO_KEY
    column_create Data tags COLUMN_VECTOR|WITH_WEIGHT Tags
    column_create Tags data_tags COLUMN_INDEX|WITH_WEIGHT Data tags
    load --table Data
    [
      {"tags": "{\"fruit\": 10, \"apple\": 100}"},
      {"tags": "{\"fruit\": 200}"}
    ]
    

• Added support for Float32 type.

  • Groonga already has a Float type. However, it is double precision floating point number. Therefore if we only use single precision floating point number, it is not efficient.

  • We can select more suitable type by adding a Float32 type.

• Added following APIs

  • grn_obj_unref(grn_ctx *ctx, grn_obj *obj)

    • This API is only used on the reference count mode (The reference count mode is a state of GRN_ENABLE_REFERENCE_COUNT=yes.).

      • It calls grn_obj_unlink() only on the reference count mode. It doesn’t do anything except when the reference count mode.

      • We useful it when we need only call grn_obj_unlink() on the referece count mode.

      • Because as the following example, we don’t write condition that whether the reference count mode or not.

        • The example if we don’t use grn_obj_unref().

          if (grn_enable_reference_count) {
           grn_obj_unlink(ctx, obj);
          }
          

        • The example if we use grn_obj_unref().

          grn_obj_ubref(ctx, obj);
          

  • grn_get_version_major(void)

  • grn_get_version_minor(void)

  • grn_get_version_micro(void)

    • They return Groonga’s major, minor, and micor version numbers as a uint32_t.

  • grn_posting_get_record_id(grn_ctx *ctx, grn_posting *posting)

  • grn_posting_get_section_id(grn_ctx *ctx, grn_posting *posting)

  • grn_posting_get_position(grn_ctx *ctx, grn_posting *posting)

  • grn_posting_get_tf(grn_ctx *ctx, grn_posting *posting)

  • grn_posting_get_weight(grn_ctx *ctx, grn_posting *posting)

  • grn_posting_get_weight_float(grn_ctx *ctx, grn_posting *posting)

  • grn_posting_get_rest(grn_ctx *ctx, grn_posting *posting)

    • They return information on the posting list.

    • These APIs return value as a uint32_t except grn_posting_get_weight_float.

    • grn_posting_get_weight_float returns value as a float.

    • grn_posting_get_section_id(grn_ctx *ctx, grn_posting *posting)

      • Section id is the internal representation of the column name.

      • If column name store in posting list as a string, it is a large amount of information and it use waste capacity.

      • Therefore, Groonga compresses the amount of information and use storage capacity is small by storing column name in the posting list as a number called section id.

    • grn_posting_get_tf(grn_ctx *ctx, grn_posting *posting)

      • tf of grn_posting_get_tf is Term Frequency score.

    • grn_posting_get_weight_float(grn_ctx *ctx, grn_posting *posting)

      • It returns weight of token as a float.

      • We suggest using this API when we get a weight of token after this.

        • Because we modify the internal representation of the weight from uint32_t to float in the near future.

Fixes¶

• Fixed a bug that Groonga for 32bit on GNU/Linux may crash.

• Fixed a bug that unrelated column value may be cleared. [GtiHub#1087][Reported by sutamin]

• Fixed a memory leak when we dumped records with dump command.

• Fixed a memory leak when we specified invalid value into output_columns.

• Fixed a memory leak when we executed snippet function.

• Fixed a memory leak when we filled the below conditions.

  • If we used dynamic columns on the initial stage.

  • If we used slices argument with select command.

• Fixed a memory leak when we deleted tables with logical_table_remove.

• Fixed a memory leak when we use the reference count mode.

  • The reference count mode is a GRN_ENABLE_REFERENCE_COUNT=yes state.

  • This mode is experimental. Performance may degrade by this mode.

• Fixed a bug that Groonga too much unlink _key accessor when we load data for apache arrow format.

Thanks¶

• sutamin

• naoa

Fixes¶

• Added a missing runtime(vcruntime140_1.dll) in package for Windows VC++ version.

Improvements¶

• [httpd] Updated bundled nginx to 1.17.9.

• [httpd] Added support for specifying output type as an extension.

  • For example, we can write load.json instead of load?output_type=json.

• [Log] Outputted a path of opened or closed file into a log of dump level on Linux.

• [Log] Outputted a path of closed file into a log of debug level on Windows.

• Added following API and macros

  • grn_timeval_from_double(grn_ctx, double)

    • This API converts double type to grn_timeval type.

    • It returns value of grn_timeval type.

  • GRN_TIMEVAL_TO_NSEC(timeval)

    • This macro converts value of grn_timeval type to nanosecond as the value of uint64_t type.

  • GRN_TIME_USEC_TO_SEC(usec)

    • This macro converts microsecond to second.

• Deprecated the following macro.

  • GRN_OBJ_FORMAT_FIN(grn_ctx, grn_obj_format)

    • We grn_obj_format_fin(grn_ctx, grn_obj_format) use instead since 10.0.0.

• [logical_range_filter],[dump] Added support for stream output.

  • This feature requires command_version 3 or later. The header content is outputted after the body content.

  • Currently, this feature support only dump and logical_range_filter.

  • logical_range_filter always returns the output as a stream on command_version 3 or later.

  • This feature has the following limitations.

    • -1 is only allowed for negative limit

    • MessagePack output isn’t supported

  • We a little changed the response contents of JSON by this modify.

    • The key order differs from before versions as below.

      • The key order in before versions.

        {
          "header": {...},
          "body": {...}
        }
        

      • The key order in this version(10.0.0).

        {
          "body": {...},
          "header": {...}
        }
        

  • Disabled caches of dump and logical_range_filter when they execute on command_version 3.

    • Because of dump and logical_range_filter on command_version 3 returns stream since 10.0.0, Groonga can not cache the whole response.

• [logical_range_filter] Added support for outputting response as Apache Arrow format.

  • Supported data type as below.

    • UInt8

    • Int8

    • UInt16

    • Int16

    • UInt32

    • Int32

    • UInt64

    • Int64

    • Time

    • ShortText

    • Text

    • LongText

    • Vector of Int32

    • Reference vector

• Supported Ubuntu 20.04 (Focal Fossa).

• Dropped Ubuntu 19.04 (Disco Dingo).

  • Because this version has been EOL.

Improvements¶

• [tools] Added a script for copying only files of specify tables or columns.

  • This script name is copy-related-files.rb.

  • This script is useful if we want to extract specifying tables or columns from a huge database.

  • Related files of specific tables or columns may need for reproducing fault.

  • If we difficult to offer a database whole, we can extract related files of target tables or columns by this tool.

• [shutdown] Accept /d/shutdown?mode=immediate immediately even when all threads are used.

  • This feature can only use on the Groonga HTTP server.

• Unused objects free immediately by using GRN_ENABLE_REFERENCE_COUNT=yes.

  • This feature is experimental. Performance degrade by this feature.

  • If we load to span many tables, we can expect to keep in the usage of memory by this feature.

• [CentOS] We prepare groonga-release by version.

  • Note that the little modification how to install.

• [Debian GNU/Linux] We use groonga-archive-keyring for adding the Groonga apt repository.

  • We can easy to add the Groonga apt repository to our system by this improvement.

  • groonga-archive-keyring includes all information for using the Groonga apt repository. Thus, we need not be conscious of changing of repository information or PGP key by installing this package.

  • groonga-archive-keyring is deb package. Thus, we can easy to update by apt update.

Improvements¶

• [load] Added support for Apache Arrow format data.

  • If we use Apache Arrow format data, we may reduce parse cost. Therefore, data might be loading faster than other formats.

  • Groonga can also directly input data for Apache Arrow format from other data analysis systems by this improvement.

  • However, Apache Arrow format can use in the HTTP interface only. We can’t use it in the command line interface.

• [load] Added how to load Apache Arrow format data in the document.

• [load] Improve error message.

  • Response of load command includes error message also.

  • If we faile data load, Groonga output error detail of load command by this Improvement.

• [httpd] Updated bundled nginx to 1.17.7.

• [Groonga HTTP server] Added support for sending command parameters by body of HTTP request.

  • We must set application/x-www-form-urlencoded to Content-Type for this case.

• [Groonga HTTP server] Added how to use HTTP POST in the document.

Improvements¶

• Improved the performance of the “&&” operation.

  • For example, the performance of condition expression such as the following is increased.

  • ( A || B ) && ( C || D ) && ( E || F) …

• [TokenMecab] Added a new option use_base_form

  • We can search using the base form of a token by this option.

  • For example, if we search “支えた” using this option, “支える” is hit also.

Fixes¶

• Fix a bug that when the accessor is index, performance decreases.

  • For example, it occurs with the query include the following conditions.

    • accessor @ query

    • accessor == query

• Fixed a bug the estimated size of a search result was overflow when the buffer is big enough. [PGroonga#GitHub#115][Reported by Albert Song]

• Improved a test(1) portability. [GitHub#1065][Patched by OBATA Akio]

• Added missing tools.

  • Because index-column-diff-all.sh and object-inspect-all.sh had not bundled in before version.

Thanks¶

• Albert Song

• OBATA Akio

Improvements¶

• [Log] Improved that output the sending time of response into query-log.

• [status] Added that the number of current jobs in the status command response.

• [groonga-httpd] Added support for $request_time in log.

  • In the previous version, even if we specified the $request_time in the log_format directive, it was always 0.00.

  • If we specify the $request_time, groonga-httpd output the correct time form this version.

• [groonga-httpd] Added how to set the $request_time in the document.

• Supported Ubuntu 19.10 (Eoan Ermine)

• Supported CentOS 8 (experimental)

  • The package for CentOS 8 can’t use a part of features(e.g. we can’t use TokenMecab and can’t cast to int32 vector from JSON string) for lacking some packages for development.

• [tools] Added a script for executing the index_column_diff command simply.

  • This script name is index-column-diff-all.sh.

  • This script extracts index columns form Groonga’s database and execute the index_column_diff command to them.

• [tools] Added a script for executing object_inspect against all objects.

  • This script name is object-inspect-all.sh.

Fixes¶

• Fixed a bug that Groonga crash when we specify the value as the first argument of between.[GitHub#1045][Reported by yagisumi]

Thanks¶

• yagisumi

Improvements¶

• [log_reopen] Added a supplementary explanation when we use groonga-httpd with 2 or more workers.

• Improved that Groonga ignores the index being built.

  • We can get correct search results even if the index is under construction.

  • However, the search is slow because of Groonga out of use the index to search in this case.

• [sub_filter] Added a feature that sub_filter executes a sequential search when Groonga is building indexes for the target column or the target column hasn’t indexed.

  • sub_filter was an error if the above situation in before version.

  • From this version, sub_filter returns search results if the above situation.

  • However if the above situation, sub_filter is slow. Because it is executed as a sequential search.

• [CentOS] Dropped 32-bit package support on CentOS 6.

Fixes¶

• [logical_range_filter] Fixed a bug that exception about closing the same object twice occurs when we have enough records and the number of records that unmatch filter search criteria is more than the estimated value of it.

Improvements¶

• [httpd] Updated bundled nginx to 1.17.3.

  • Contains security fix for CVE-2019-9511, CVE-2019-9513, and CVE-2019-9516.

Fixes¶

• Fixed a bug that Groonga crash when posting lists were huge.

  • However, this bug almost doesn’t occur by general data. Because posting lists don’t grow bigger so much by them.

• Fixed a bug that returns an empty result when we specify initial into a stage of a dynamic column and search for using index. [GitHub#683]

• Fixed a bug that the configure phase didn’t detect libedit despite installing it. [GitHub#1030][Patched by yu]

• Fixed a bug that --offset and --limit options didn’t work with --sort_keys and --slices options. [clear-code/redmine_full_text_search#70][Reported by a9zawa]

• Fixed a bug that search result is empty when the result of select command is huge. [groonga-dev,04770][Reported by Yutaro Shimamura]

• Fixed a bug that doesn’t use a suitable index when prefix search and suffix search. [GitHub#1007, PGroonga#GitHub#96][Reported by oknj]

Thanks¶

• oknj

• Yutaro Shimamura

• yu

• a9zawa

Improvements¶

• Added support for Debian 10 (buster).

Fixes¶

• [select] Fixed a bug that search is an error when occurring search escalation.

• [select] Fixed a bug that may return wrong search results when we use nested equal condition.

• [geo_distance_location_rectangle] Fixed an example that has wrong load format. [GitHub#1023][Patched by yagisumi]

• [Let’s create micro-blog] Fixed an example that has wrong search results. [GutHub#1024][Patched by yagisumi]

Thanks¶

• yagisumi

Improvements¶

• [logical_range_filter] Improved that only apply an optimization when the search target shard is large enough.

  • This feature reduces that duplicate search result between offset when we use same sort key.

  • Large enough threshold is 10000 records by default.

• [Normalizers] Added new option unify_to_katakana for NormalizerNFKC100.

  • This option normalize hiragana to katakana.

  • For example, ゔぁゔぃゔゔぇゔぉ is normalized to ヴァヴィヴヴェヴォ.

• [select] Added drilldowns support as a slices parameter.

• [select] Added columns support as a slices parameter.

• [select] Improved that we can refer _score in the initial stage for slices parameter.

• [highlight_html], [snippet_html] Improved that extract a keyword also from an expression of before executing a slices when we specify the slices parameter.

• Improved that collect scores also from an expression of before executing a slices when we specify the slices parameter.

• Stopped add 1 in score automatically when add posting to posting list.

  • grn_ii_posting_add is backward incompatible changed by this change. * Caller must increase the score to maintain compatibility.

• Added support for index search for nested equal like XXX.YYY.ZZZ == AAA.

• Reduce rehash interval when we use hash table.

  • This feature improve performance for output result.

• Improved to we can add tag prefix in the query log.

  • We become easy to understand that it is filtered which the condition.

• Added support for Apache Arrow 1.0.0.

  • However, It’s not released this version yet.

• Added support for Amazon Linux 2.

Fixes¶

• Fixed a bug that vector values of JSON like "[1, 2, 3]" are not indexed.

• Fixed wrong parameter name in table_create tests. [GitHub#1000][Patch by yagisumi]

• Fixed a bug that drilldown label is empty when a drilldown command is executed by command_version=3. [GitHub#1001][Reported by yagisumi]

• Fixed build error for Windows package on MinGW.

• Fixed install missing COPYING for Windows package on MinGW.

• Fixed a bug that don’t highlight when specifing non-text query as highlight target keyword.

• Fixed a bug that broken output of MessagePack format of [object_inspect]. [GitHub#1009][Reported by yagisumi]

• Fixed a bug that broken output of MessagePack format of index_column_diff. [GitHub#1010][Reported by yagisumi]

• Fixed a bug that broken output of MessagePack format of [suggest]. [GitHub#1011][Reported by yagisumi]

• Fixed a bug that allocate size by realloc isn’t enough when a search for a table of patricia trie and so on. [Reported by Shimadzu Corporation]

  • Groonga may be crashed by this bug.

• Fix a bug that groonga.repo is removed when updating 1.5.0 from groonga-release version before 1.5.0-1. [groonga-talk:429][Reported by Josep Sanz]

Thanks¶

• yagisumi

• Shimadzu Corporation

• Josep Sanz

Improvements¶

• Added support for array literal with multiple elements.

• Added support equivalence operation of a vector.

• [logical_range_filter] Increase outputting logs into query log.

  • logical_range_filter command comes to output a log for below timing.

    • After filtering by logical_range_filter.

    • After sorting by logical_range_filter.

    • After applying dynamic column.

    • After output results.

  • We can see how much has been finished this command by this feature.

• [Tokenizers] Added document for TokenPattern description.

• [Tokenizers] Added document for TokenTable description.

• [Tokenizers] Added document for TokenNgram description.

• [grndb] Added output operation log into groonga.log

  • grndb command comes to output execution result and execution process.

• [grndb] Added support for checking empty files.

  • We can check if the empty files exist by this feature.

• [grndb] Added support new option --since

  • We can specify a scope of an inspection.

• [grndb] Added document about new option --since

• Bundle RapidJSON

  • We can use RapidJson as Groonga’s JSON parser partly. (This feature is partly yet)

  • We can more exactly parse JSON by using this.

• Added support for casting to int32 vector from JSON string.

  • This feature requires RapidJSON.

• [query] Added default_operator.

  • We can customize operator when “keyword1 keyword2”.

  • “keyword1 Keyword2” is AND operation in default.

  • We can change “keyword1 keyword2“‘s operator except AND.

Fixes¶

• [optimizer] Fix a bug that execution error when specified multiple filter conditions and like xxx.yyy=="keyword".

• Added missing LICENSE files in Groonga package for Windows(VC++ version).

• Added UCRT runtime into Groonga package for Windows(VC++ version).

• [Window function] Fix a memory leak.

  • This occurs when multiple windows with sort keys are used. [Patched by Takashi Hashida]

Thanks¶

• Takashi Hashida

Improvements¶

• [select] Added more query logs.

  • select command comes to output a log for below timing.

    • After sorting by drilldown.

    • After filter by drilldown.

  • We can see how much has been finished this command by this feature.

• [logical_select] Added more query logs.

  • logical_select command comes to output a log for below timing.

    • After making dynamic columns.

    • After grouping by drilldown.

    • After sorting by drilldown.

    • After filter by drilldown.

    • After sorting by logical_select.

  • We can see how much has been finished this command by this feature.

• [logical_select] Improved performance of sort a little when we use limit option.

• [index_column_diff] Improved performance.

  • We have greatly shortened the execution speed of this command.

• [index_column_diff] Improved ignore invalid reference.

• [index_column_diff] Added support for duplicated vector element case.

• [Normalizers] Added a new Normalizer NormalizerNFKC121 based on Unicode NFKC (Normalization Form Compatibility Composition) for Unicode 12.1.

• [TokenFilters] Added a new TokenFilter TokenFilterNFKC121 based on Unicode NFKC (Normalization Form Compatibility Composition) for Unicode 12.1.

• [grndb] Added a new option --log-flags

  • We can specify output items of a log as with groonga executable file.

  • See [groonga executable file] to know about supported log flags.

• [snippet_html] Added a new option for changing a return value when no match by search.

• [plugin_unregister] Added support full path of Windows.

• Added support for multiline log message.

  • The multiline log message is easy to read by this feature.

• Output key in Groonga’s log when we search by index.

• [match_columns parameter] Added a document for indexes with weight.

• [logical_range_filter] Added a explanation for order parameter.

• [object_inspect] Added an explanation for new statistics INDEX_COLUMN_VALUE_STATISTICS_NEXT_PHYSICAL_SEGMENT_ID and INDEX_COLUMN_VALUE_STATISTICS_N_PHYSICAL_SEGMENTS.

• Dropped Ubuntu 14.04 support.

Fixes¶

• [index_column_diff] Fixed a bug that too much remains are reported.

• Fixed a build error when we use --without-onigmo option. [GitHub#951] [Reported by Tomohiro KATO]

• Fixed a vulnerability of “CVE: 2019-11675”. [Reported by Wolfgang Hotwagner]

• Removed extended path prefix \\?\ at Windows version of Groonga. [GitHub#958] [Reported by yagisumi]

  • This extended prefix causes a bug that plugin can’t be found correctly.

Thanks¶

• Tomohiro KATO

• Wolfgang Hotwagner

• yagisumi

Improvements¶

• [column_create] Added a new flag INDEX_LARGE for index column.

  • We can make an index column has space that two times of default by this flag.

  • However, note that it also uses two times of memory usage.

  • This flag useful when index target data are large.

  • Large data must have many records (normally at least 10 millions records) and at least one of the following features.

    • Index targets are multiple columns

    • Index table has tokenizer

• [object_inspect] Added a new statistics next_physical_segment_id and max_n_physical_segments for physical segment information.

  • We can confirm usage of index column space and max value of index column space by this information.

• [logical_select] Added support for window function over shard.

• [logical_range_filter] Added support for window function over shard.

• [logical_count] Added support for window function over shard.

• We provided a package for Windows made from VC++.

• [io_flush] Added a new option --recursive dependent

  • We can all of the specified flush target object, child objects, corresponding table of an index column and corresponding index column are flush target objects.

Fixes¶

• Fixed “unknown type name ‘bool’” compilation error in some environments.

• Fixed a bug that incorrect output number over Int32 by command of execute via mruby (e.g. logical_select, logical_range_filter, logical_count, etc.). [GitHub#936] [Patch by HashidaTKS]

Thanks¶

• HashidaTKS

Improvements¶

• Added support to acccept null for vector value.

  • You can use select … –columns[vector].flags COLUMN_VECTOR –columns[vector].value “null”

• [dump] Translated document into English.

• Added more checks and logging for invalid indexes. It helps to clarify the index related bugs.

• Improved an explanation about GRN_TABLE_SELECT_ENOUGH_FILTERED_RATIO behavior in news at Release 8.0.6 - 2018-08-29.

• [select] Added new argument --load_table, --load_columns and --load_values.

  • You can store a result of select in a table that specifying --load_table.

  • --load_values option specifies columns of result of select.

  • --load_columns options specifies columns of table that specifying --load_table.

  • In this way, you can store values of columns that specifying with --load_values into columns that specifying with --load_columns.

• [select] Added documentation about load_table, load_columns and load_values.

• [load] Added supoort to display a table of load destination in a query log.

  • A name of table of load destination display as string in [] as below.

  • :000000000000000 load(3): [LoadedLogs][3]

• Added a new API:

  • grn_ii_get_flags()

  • grn_index_column_diff()

  • grn_memory_get_usage()

• Added index_column_diff command to check broken index column. If you want to log progress of command execution, set log level to debug.

Fixes¶

• [snippet_html] Changed to return an empty vector for no match.

  • In such a case, an empty vector [] is returned instead of null.

• Fixed a warning about possibility of counting threads overflow. In real world, it doesn’t affect user because enourmous number of threads is not used. [GitHub#904]

• Fixed build error on macOS [GitHub#909] [Reported by shiro615]

• Fixed a stop word handling bug.

  • This bug occurs when we set the first token as a stop word in our query.

  • If this bug occurs, our search query isn’t hit.

• [Global configurations] Fixed a typo about parameter name of grn_lock_set_timeout.

• Fixed a bug that deleted records may be matched because of updating indexes incorrectly.

  • It may occure when large number of records is added or deleted.

• Fixed a memory leak when logical_range_filter returns no records. [GitHub#911] [Patch by HashidaTKS]

• Fixed a bug that query will not match because of loading data is not normalized correctly. [PGroonga#GitHub#93, GitHub#912,GitHub#913] [Reported by kamicup and dodaisuke]

  • This bug occurs when load data contains whitespace after KATAKANA and unify_kana option is used for normalizer.

• Fixed a bug that an indexes is broken during updating indexes.

  • It may occurs when repeating to add large number of records or delete them for a long term.

• Fixed a crash bug that allocated working area is not enough size when updating indexes.

Thanks¶

• shiro615

• HashidaTKS

• kamicup

• dodaisuke

Improvements¶

• [Tokenizers] Added a new tokenizer TokenPattern.

  • You can extract tokens by regular expression.

    • This tokenizer extracts only token that matches the regular expression.

  • You can also specify multiple patterns of regular expression.

• [Tokenizers] Added a new tokenizer TokenTable.

  • You can extract tokens by a value of columns of existing a table.

• [dump] Added support for dumping binary data.

• [select] Added support for similer search against index column.

  • If you have used multi column index, you can similar search against all source columns by this feature.

• [Normalizers] Added new option remove_blank for NormalizerNFKC100.

  • This option remove white spaces.

• [groonga executable file] Improve display of thread id in log.

  • Because It was easy to confuse thread id and process id on Windows version, it made clear which is a thread id or a process id.

Improvements¶

• [logical_select] Added new argument --load_table, --load_columns and --load_values.

  • You can store a result of logical_select in a table that specifying --load_table.

  • --load_values option specifies columns of result of logical_select.

  • --load_columns options specifies columns of table that specifying --load_table.

  • In this way, you can store values of columns that specifying with --load_values into columns that specifying with --load_columns.

• Improve error log when update error of index.

  • Added more information in the log.

    • For example, output source buffer and chunk when occur merge of posting lists error.

    • Also, outputting the log a free space size of a buffer and request size of a buffer when occurs error of allocating a buffer.

• [groonga executable file] Added a new option --log-flags.

  • We can specify output items of a log of the Groonga.

  • We can output as below items.

    • Timestamp

    • Log message

    • Location(the location where the log was output)

    • Process id

    • Thread id

  • We can specify prefix as below.

    • +

      • This prefix means that “add the flag”.

    • -

      • This prefix means that “remove the flag”.

    • No prefix means that “replace existing flags”.

  • Specifically, we can specify flags as below.

    • none

      • Output nothing into the log.

    • time

      • Output a timestamp into the log.

    • message

      • Output log messages into the log.

    • location

      • Output the location where the log was output( a file name, a line and a function name) and process id.

    • process_id

      • Output a process id into the log.

    • pid

      • This flag is an alias of process_id.

    • thread_id

      • Output thread id into the log.

    • all

      • This flag specifies all flags except none and default flags.

    • default

      • Output a timestamp and log messages into the log.

  • We can also specify multiple log flags by separating flags with |.

Fixes¶

• Fixed a memory leak when occurs index update error.

• [Normalizers] Fixed a bug that stateless normalizers and stateful normalizers return wrong results when we use them at the same time.

  • Stateless normalizers are below.

    • unify_kana

    • unify_kana_case

    • unify_kana_voiced_sound_mark

    • unify_hyphen

    • unify_prolonged_sound_mark

    • unify_hyphen_and_prolonged_sound_mark

    • unify_middle_dot

  • Stateful normalizers are below.

    • unify_katakana_v_sounds

    • unify_katakana_bu_sound

    • unify_to_romaji

Improvements¶

• [httpd] Updated bundled nginx to 1.15.8.

Fixes¶

• Fixed a bug that unlock against DB is always executed after flush when after execute a io_flush command.

• Fixed a bug that reindex command doesn’t finish when execute a reindex command against table that has record that has not references.

Improvements¶

• [Tokenizers] Improved that output a tokenizer name in error message when create tokenizer fail.

• [Tokenizers][TokenDelimit] Supported that customizing delimiter of a token.

  • You can use token other than whitespace as a token of delimiter.

• [Tokenizers][TokenDelimit] Added new option pattern.

  • You can specify delimiter with regular expression by this option.

• [Tokenizers] Added force_prefix_search value to each token information.

  • “force_prefix” is kept for backward compatibility.

• [Token filters] Added built-in token filter TokenFilterNFKC100.

  • You can convert katakana to hiragana like NormalizerNFKC100 with a unify_kana option.

• [Token filters][TokenFilterStem] Added new option algorithm.

  • You can also stem language other than English(French, Spanish, Portuguese, Italian, Romanian, German, Dutch, Swedish, Norwegian, Danish, Russian, Finnish) by this option.

• [Token filters][TokenFilterStopWord] Added new option column.

  • You can specify stop word in a column other than is_stop_word column by this option.

• [dump] Supported output options of token filter options.

  • If you specify a tokenizer like TokenNgram or TokenMecab etc that has options, you can output these options with table_list command.

• [truncate] Supported a table that it has token filter option.

  • You can truncate even a tabel that it has token filter like TokenFilterStem or TokenStopWord that has options.

• [schema] Support output of options of token filter.

• [Normalizers] Added new option for NormalizerNFKC100 that unify_to_romaji option.

  • You can normalize hiragana and katakana to romaji by this option.

• [query-log][show-condition] Supported “func() > 0” case.

• [Windows] Improved that ensure flushing on unmap.

• Improved error message on opening input file error.

• [httpd] Updated bundled nginx to 1.15.7.

  • contains security fix for CVE-2018-16843 and CVE-2018-16844.

Fixes¶

• Fixed a memory leak when evaluating window function.

• [groonga-httpd] Fixed bug that log content may be mixed.

• Fixed a bug that generates invalid JSON when occurs error of slice on output_columns.

• Fixed a memory leak when getting nested reference vector column value.

• Fixed a crash bug when outputting warning logs of index corruption.

• Fix a crash bug when temporary vector is reused in expression evaluation.

  • For example, crash when evaluating an expression that uses a vector as below.

  _score = _score + (vector_size(categories) > 0)

• Fix a bug that hits a value of vector columns deleted by a delete command.[GitHub PGroonga#85][Reported by dodaisuke]

Thanks¶

• dodaisuke

Improvements¶

• [table_list] Supported output options of default tokenizer.

  • If you specify a tokenizer like TokenNgram or TokenMecab etc that has options, you can output these options with table_list command.

• [select] Supported normalizer options in sequential match with record @ 'query'.

• [truncate] Supported a table that it has tokenizer option.

  • You can truncate even a tabel that it has tokenizer like TokenNgram or TokenMecab etc that has options.

• [Tokenizers][TokenMecab] Added new option target_class

  • This option searches a token of specifying a part-of-speech. For example, you can search only a noun.

  • This option can also specify subclasses and exclude or add specific part-of-speech of specific using + or -. So, you can also search except a pronoun as below.

    'TokenMecab("target_class", "-名詞/代名詞", "target_class", "+")'

• [io_flush] Supported locking of a database during a io_flush.

  • Because Groonga had a problem taht is a crash when deleteing a table of a target of a io_flush during execution of a io_flush.

• [cast_loose] Added a new function cast_loose.

  • This function cast to a type to specify. If a value to specify can’t cast, it become to a default value to specify.

• Added optimize the order of evaluation of a conditional expression.(experimental)

  • You can active this feature by setting environment value as below.

    GRN_EXPR_OPTIMIZE=yes

• Supported (?-mix:XXX) form for index searchable regular expression. [groonga-dev,04683][Reported by Masatoshi SEKI]

  • (?-mix:XXX) form treats the same as XXX.

• [httpd] Updated bundled nginx to 1.15.5.

• Supported Ubuntu 18.10 (Cosmic Cuttlefish)

Fixes¶

• Fixed a bug that the Groonga GQTP server may fail to accept a new connection. [groonga-dev,04688][Reported by Yutaro Shimamura]

  • It’s caused when interruption client process without using quit.

Thanks¶

• Masatoshi SEKI

• Yutaro Shimamura

Improvements¶

• [Tokenizers][TokenMecab] support outputting metadata of Mecab.

  • Added new option include_class for TokenMecab.

    This option outputs class and subclass in Mecab’s metadata.

  • Added new option include_reading for TokenMecab.

    This option outputs reading in Mecab’s metadata.

  • Added new option include_form for TokenMecab.

    This option outputs inflected_type, inflected_form and base_form in Mecab’s metadata.

  • Added new option use_reading for TokenMecab.

    This option supports a search by kana.

    This option is useful for countermeasure of orthographical variants because it searches with kana.

• [plugin] Groonga now can grab plugins from multiple directories.

  You can specify multiple directories to GRN_PLUGINS_PATH separated with “:” on non Windows, “;” on Windows.

  GRN_PLUGINS_PATH has high priority than the existing GRN_PLUGINS_DIR. Currently, this option is not supported Windows.

• [Tokenizers][TokenNgram] Added new option unify_alphabet for TokenNgram.

  If we use unify_alphabet as false, TokenNgram uses bigram tokenize method for ASCII character.

• [Tokenizers][TokenNgram] Added new option unify_symbol for TokenNgram.

  TokenNgram("unify_symbol", false) is same behavior of TokenBigramSplitSymbol.

• [Tokenizers][TokenNgram] Added new option unify_digit for TokenNgram.

  If we use unify_digit as false, If we set false, TokenNgram uses bigram tokenize method for digits.

• [httpd] Updated bundled nginx to 1.15.4.

Fixes¶

• Fixed wrong score calculations on some cases.

  • It’s caused when adding, multiplication or division numeric to a bool value.

  • It’s caused when comparing a scalar and vector columns using != or ==.

Improvements¶

• [Tokenizers][TokenMecab] add chunked_tokenize and chunk_size_threshold options.

• [optimizer] support estimation for query family expressions. It will generate more effective execution plan with query family expressions such as column @ query, column @~ pattern and so on.

• [optimizer] plug-in -> built-in It’s disabled by default for now. We can enable it by defining GRN_EXPR_OPTIMIZE=yes environment variable or using expression_rewriters table as before.

• Enable sequential search for enough filtered case by default. If the current result is enough filtered, sequential search is faster than index search. If the current result has only 1% records of all records in a table and less than 1000 records, sequential search is used even when index search is available.

  Cullently, this optimization is applied when search by ==, >, <, >=, or <=.

  When a key of a table that has columns specified by the filter is ShortText, you must set NormalizerAuto to normalizer of the table to apply this optimization.

  You can disable this feature by GRN_TABLE_SELECT_ENOUGH_FILTERED_RATIO=0.0 environment variable.

• [load] improve error message. Table name is included.

• [load] add lock_table option. If --lock_table yes is specified, load locks the target table while updating columns and applying --each. This option avoids load and delete conflicts but it’ll reduce load performance.

• [vector_find] avoid to crash with unsupported modes

Fixes¶

• [index] fix a bug that offline index construction for text vector with HASH_KEY. It creates index with invalid section ID.

• Fix a bug that --match_columns 'index[0] || index[9]' uses wrong section.

• [highlighter] fix a wrong highlight bug It’s caused when lexicon is hash table and keyword is less than N of N-gram.

• [mruby] fix a bug that real error is hidden. mruby doesn’t support error propagation by no argument raise. https://github.com/mruby/mruby/issues/290

• [Tokenizers][TokenNgram loose]: fix a not found bug when query has only loose types. highlight_html() with lexicon was also broken.

• Fix a bug that text->number cast ignores trailing garbage. “0garbage” should be cast error.

• Fix an optimization bug for reference_column >= 'key_value' case

Improvements¶

• [Script syntax] Added complementary explain about similar search against Japanese documents. [GitHub#858][Patch by Yasuhiro Horimoto]

• [time_classify_day_of_week] Added a new API: time_classify_day_of_week().

• Suppressed a warning with -fstack-protector. Suggested by OBATA Akio.

• Added a new API: time_format_iso8601().

• Exported a struct grn_raw_string.

• Added a new API: grn_obj_clear_option_values(). It allows you to clear option values on remove (for persistent) / close (for temporary.)

• [log] Reported index column name for error message [ii][update][one].

• [httpd] Updated bundled nginx to 1.15.2.

• [Ubuntu] Dropped Ubuntu 17.10 (Artful Aardvark) support. It has reached EOL at July 19, 2018.

• [Debian GNU/Linux] Dropped jessie support. Debian’s security and release team will no longer produce updates for jessie.

Fixes¶

• Fixed returning wrong result after unfinished /d/load data by POST.

• Fixed wrong function call around KyTea.

• [grndb] Added a missing label for the --force-truncate option.

• Fixed crash on closing of a database, when a normalizer provided by a plugin (ex. groonga-normalizer-mysql) is used with any option.

• Fixed a bug that normalizer/tokenizer options may be ignored. It’s occurred when the same object ID is reused.

Improvements¶

• [log] Add sub error for error message [ii][update][one].

• Added a new API: grn_highlighter_clear_keywords().

• Added a new predicate: grn_obj_is_number_family_bulk().

• Added a new API: grn_plugin_proc_get_value_mode().

• [vector_find] Added a new function vector_find().

• Suppress memcpy warnings in msgpack.

• Updated mruby from 1.0.0 to 1.4.1.

• [doc][grn_obj] Added API reference for grn_obj_is_index_column().

• [windows] Suppress printf format warnings.

• [windows] Suppress warning by msgpack.

• [grn_obj][Plugin] Added encoding converter. rules:

  • grn_ctx::errbuf: grn_encoding

  • grn_logger_put: grn_encoding

  • mruby: UTF-8

  • path: locale

• [mrb] Added LocaleOutput.

• [windows] Supported converting image path to grn_encoding.

• [Tokenizers][TokenMecab] Convert error message encoding.

• [window_sum] Supported dynamic column as a target column.

• [doc][grn_obj] Added API reference for grn_obj_is_vector_column().

• [column_create] Added more validations.

  • 1: Full text search index for vector column must have WITH_SECTION flag. (Note that TokenDelmit with WITH_POSITION without WITH_SECTION is permitted. It’s useful pattern for tag search.)

  • 2: Full text search index for vector column must not be multi column index. detail: https://github.com/groonga/groonga/commit/08e2456ba35407e3d5172f71a0200fac2a770142

• [grndb] Disabled log check temporarily. Because it’s not completed yet.

Fixes¶

• [sub_filter] Fixed too much score with a too filtered case.

• Fixed build error if KyTea is installed.

• [grndb] Fixed output channel.

• [query-log][show-condition] Maybe fixed a crash bug.

• [highlighter][lexicon] Fixed a not highlighted bug. The keyword wasn’t highlighted if keyword length is less than N (“N”-gram. In many cases, it’s Bigram so “less than 2”).

• [windows] Fixed a base path detection bug. If system locale DLL path includes 0x5c (\ in ASCII) such as “U+8868 CJK UNIFIED IDEOGRAPH-8868” in CP932, the base path detection is buggy.

• [Tokenizers][TokenNgram] Fixed wrong first character length. It’s caused for “PARENTHESIZED IDEOGRAPH” characters such as “U+3231 PARENTHESIZED IDEOGRAPH STOCK”.

Improvements¶

• [highlight_html] Support highlight of results of the search by NormalizerNFKC100 or TokenNgram.

• [Tokenizers] Added new option for TokenNgram that report_source_location option . This option used when highlighting with highlight_html use a lexicon.

• [Normalizers] Added new option for NormalizerNFKC100 that unify_middle_dot option. This option normalizes middle dot. You can search with or without ・ (middle dot) and regardless of ・ position.

• [Normalizers] Added new option for NormalizerNFKC100 that unify_katakana_v_sounds option. This option normalizes ヴァヴィヴヴェヴォ (katakana) to バビブベボ (katakana). For example, you can search バイオリン (violin) in ヴァイオリン (violin).

• [Normalizers] Added new option for NormalizerNFKC100 that unify_katakana_bu_sound option. This option normalizes ヴァヴィヴゥヴェヴォ (katakana) to ブ (katakana). For example, you can search セーブル (katakana) and セーヴル in セーヴェル (katakana).

• [sub_filter] Supported sub_filter optimization for the too filter case. this optimize is valid when records are enough narrowed down before sub_filter execution as below.

• [groonga-httpd] Made all workers context address to unique. context address is #{ID} of below query log.

  #{TIME_STAMP}|#{MESSAGE}

  #{TIME_STAMP}|#{ID}|>#{QUERY}

  #{TIME_STAMP}|#{ID}|:#{ELAPSED_TIME} #{PROGRESS}

  #{TIME_STAMP}|#{ID}|<#{ELAPSED_TIME} #{RETURN_CODE}

• [delete] Added new options that limit. You can limit the number of delete records as below example. delete --table Users --filter '_key @^ "b"' --limit 4

• [httpd] Updated bundled nginx to 1.14.0.

Fixes¶

• [logical_select] Fixed memory leak when an error occurs in filtered dynamic columns.

• [logical_count] Fixed memory leak on initial dynamic column error.

• [logical_range_filter] Fixed memory leak when an error occurs in dynamic column evaluation.

• [Tokenizers] Fixed a bug that the wrong source_offset when a loose tokenizing such as loose_symbol option.

• [Normalizers] Fixed a bug that FULLWIDTH LATIN CAPITAL LETTERs such as U+FF21 FULLWIDTH LATIN CAPITAL LETTER A aren’t normalized to LATIN SMALL LETTERs such as U+0061 LATIN SMALL LETTER A. If you have been used NormalizerNFKC100 , you must recreate your indexes.

Improvements¶

• [grndb][--force-truncate] Improved grndb recover --force-truncate option that it can be truncated even if locks are left on the table.

• [logical_range_filter] Added sort_keys option.

• Added a new function time_format(). You can specify time format against a column of Time type. You can specify with use format of strftime .

• [Tokenizers] Support new tokenizer TokenNgram. You can change its behavior dynamically via options. Here is a list of available options:

  • n : “N” of Ngram. For example, “3” for trigram.

  • loose_symbol : Tokenize keywords including symbols, to be searched by both queries with/without symbols. For example, a keyword “090-1111-2222” will be found by any of “09011112222”, “090”, “1111”, “2222” and “090-1111-2222”.

  • loose_blank : Tokenize keywords including blanks, to be searched by both queries with/without blanks. For example, a keyword “090 1111 2222” will be found by any of “09011112222”, “090”, “1111”, “2222” and “090 1111 2222”.

  • remove_blank : Tokenize keywords including blanks, to be searched by queries without blanks. For example, a keyword “090 1111 2222” will be found by any of “09011112222”, “090”, “1111” or “2222”. Note that the keyword won’t be found by a query including blanks like “090 1111 2222”.

• [Normalizers] Support new normalizer “NormalizerNFKC100” based on Unicode NFKC (Normalization Form Compatibility Composition) for Unicode 10.0.

• [Normalizers] Support options for “NormalizerNFKC51” and “NormalizerNFKC100” normalizers. You can change their behavior dynamically. Here is a list of available options:

  • unify_kana : Same pronounced characters in all of full-width Hiragana, full-width Katakana and half-width Katakana are regarded as the same character.

  • unify_kana_case : Large and small versions of same letters in all of full-width Hiragana, full-width Katakana and half-width Katakana are regarded as the same character.

  • unify_kana_voiced_sound_mark : Letters with/without voiced sound mark and semi voiced sound mark in all of full-width Hiragana, full-width Katakana and half-width Katakana are regarded as the same character.

  • unify_hyphen : The characters like hyphen are regarded as the hyphen.

  • unify_prolonged_sound_mark : The characters like prolonged sound mark are regarded as the prolonged sound mark.

  • unify_hyphen_and_prolonged_sound_mark : The characters like hyphen and prolonged sound mark are regarded as the hyphen.

• [dump] Support output of tokenizer’s options and normalizer’s options. Groonga 8.0.1 and earlier versions cannot import dump including options for tokenizers or normalizers generated by Groonga 8.0.2 or later, and it will occurs error due to unsupported information.

• [schema] Support output of tokenizer’s options and normalizer’s options. Groonga 8.0.1 and earlier versions cannot import schema including options for tokenizers or normalizers generated by Groonga 8.0.2 or later, and it will occurs error due to unsupported information.

• Supported Ubuntu 18.04 (Bionic Beaver)

Fixes¶

• Fixed a bug that unexpected record is matched with space only query. [groonga-dev,04609][Reported by satouyuzh]

• Fixed a bug that wrong scorer may be used. It’s caused when multiple scorers are used as below. --match_columns 'title || scorer_tf_at_most(content, 2.0)'.

• Fixed a bug that it may also take so much time to change “thread_limit”.

Thanks¶

• satouyuzh

Improvements¶

• [Log] Show filter conditions in query log. It’s disabled by default. To enable it, you need to set an environment variable GRN_QUERY_LOG_SHOW_CONDITION=yes.

• Install *.pdb into the directory where *.dll and *.exe are installed.

• [logical_count] Support filtered stage dynamic columns.

• [logical_count] [post_filter] Added a new filter timing. It’s executed after filtered stage columns are generated.

• [logical_select] [post_filter] Added a new filter timing. It’s executed after filtered stage columns are generated.

• Support LZ4/Zstd/zlib compression for vector data.

• Support alias to accessor such as _key.

• [logical_range_filter] Optimize window function for large result set. If we find enough matched records, we don’t apply window function to the remaining windows.

  TODO: Disable this optimization for small result set if its overhead is not negligible. The overhead is not evaluated yet.

• [select] Added match_escalation parameter. You can force to enable match escalation by --match_escalation yes. It’s stronger than --match_escalation_threshold 99999....999 because --match_escalation yes also works with SOME_CONDITIONS && column @ 'query'. --match_escalation_threshold isn’t used in this case.

  The default is --match_escalation auto. It doesn’t change the current behavior.

  You can disable match escalation by --match_escalation no. It’s the same as --match_escalation_threshold -1.

• [httpd] Updated bundled nginx to 1.13.10.

Fixes¶

• Fixed memory leak that occurs when a prefix query doesn’t match any token. [GitHub#820][Patch by Naoya Murakami]

• Fixed a bug that a cache for different databases is used when multiple databases are opened in the same process.

• Fixed a bug that a wrong index is constructed. This occurs only when the source of a column is a vector column and WITH_SECTION isn’t specified.

• Fixed a bug that a constant value can overflow or underflow in comparison (>,>=,<,<=,==,!=).

Thanks¶

• Naoya Murakami

Improvements¶

• [select] Added --drilldown_adjuster and --drilldowns[LABEL].adjuster. You can adjust score against result of drilldown.

• [Online index construction] Changed environment variable name GRN_II_REDUCE_EXPIRE_ENABLE to GRN_II_REDUCE_EXPIRE_THRESHOLD.

  GRN_II_REDUCE_EXPIRE_THRESHOLD=0 == GRN_II_REDUCE_EXPIRE_ENABLE=no. GRN_II_REDUCE_EXPIRE_THRESHOLD=-1 uses ii->chunk->max_map_seg / 2 as threshold. GRN_II_REDUCE_EXPIRE_THRESHOLD > 0 uses MIN(ii->chunk->max_map_seg / 2, GRN_II_REDUCE_EXPIRE_THRESHOLD) as threshold. GRN_II_REDUCE_EXPIRE_THRESHOLD=32 is the default.

• [between] Accept between() without borders. If the number of arguments passed to between() is 3, the 2nd and 3rd arguments are handled as the inclusive edges. [GitHub#685]

Fixes¶

• Fixed a memory leak for normal hash table. [GitHub:mroonga/mroonga#190][Reported by fuku1]

• Fix a memory leak for normal array.

• [select] Stopped to cache when output_columns uses not stable function.

• [Windows] Fixed wrong value report on WSASend error.

Thanks¶

• fuku1

Improvements¶

• [Ubuntu] Dropped Ubuntu 17.04 (Zesty Zapus) support. It has reached EOL at Jan 13, 2018.

• Added quorum match support. You can use quorum match in both script syntax and query syntax. [groonga-talk,385][Suggested by 付超群]

  TODO: Add documents for quorum match syntax and link to them.

• Added custom similarity threshold support in script syntax. You can use custom similarity threshold in script syntax.

  TODO: Add document for the syntax and link to it.

• [grndb][--force-lock-clear] Added --force-lock-clear option. With this option, grndb forces to clear locks of database, tables and data columns. You can use your database again even if locks are remained in database, tables and data columns.

  But this option very risky. Normally, you should not use it. If your database is broken, your database is still broken. This option just ignores locks.

• [load] Added surrogate pairs support in escape syntax. For example, \uD83C\uDF7A is processed as 🍺.

• [Windows] Changed to use sparse file on Windows. It reduces disk space and there are no performance demerit.

• [Online index construction] Added GRN_II_REDUCE_EXPIRE_THRESHOLD environment variable to control when memory maps are expired in index column. It’s -1 by default. It means that expire timing is depends on index column size. If index column is smaller, expire timing is more. If index column is larger, expire timing is less.

  You can use the previous behavior by 0. It means that Groonga always tries to expire.

• [logical_range_filter] [post_filter] Added a new filter timing. It’s executed after filtered stage generated columns are generated.

Fixes¶

• Reduced resource usage for creating index for reference vector. [GitHub#806][Reported by Naoya Murakami]

• [table_create] Fixed a bug that a table is created even when token_filters is invalid. [GitHub#266]

Thanks¶

• 付超群

• Naoya Murakami

Improvements¶

• [load] Improved the load’s query-log format. Added detail below items in the load’s query-log.

  • outputs number of loaded records.

  • outputs number of error records and columns.

  • outputs number of total records.

• [logical_count] Improved the logical_count’s query-log format. Added detail below items in the logical_count’s query-log.

  • outputs number of count.

• [logical_select] Improve the logical_select’s query-log format. Added detail below items in the logical_select’s query-log.

  • log N outputs.

  • outputs plain drilldown.

  • outputs labeled drilldown.

  • outputs selected in each shard.

  • use “[…]” for target information.

• [delete] Improved the delete’s query-log format. Added detail below items in the delete’s query-log.

  • outputs number of deleted and error records.

  • outputs number of rest number of records.

• [Groonga HTTP server] The server executed by groonga -s ensure stopping by C-c.

• Used NaN and Infinity, -Infinity instead of Lisp representations(#<nan> and #i1/0, #-i1/0).

• Supported vector for drilldown calc target.

• Partially supported keyword extraction from regexp search. It enables highlight_html and snippet_html for regexp search. [GitHub#787][Reported by takagi01]

• [bulk] Reduced the number of realloc(). grn_bulk_*() API supports it.

  It improves performance for large output case on Windows. For example, it causes 100x faster for 100MB over output.

  Because realloc() is heavy on Windows.

• Enabled GRN_II_OVERLAP_TOKEN_SKIP_ENABLE only when its value is “yes”.

• Deprecated GRN_NGRAM_TOKENIZER_REMOVE_BLANK_DISABLE. Use GRN_NGRAM_TOKENIZER_REMOVE_BLANK_ENABLE=no instead.

• Added new function index_column_source_records. It gets source records of index column.[Patch by Naoya Murakami]

• [select] Supported negative “offset” for “offset + size - limit” >= 0

• Added grn_column_cache. It’ll improve performance for getter of fixed size column value.

• [groonga executable file] Added --listen-backlog option. You can customize listen(2)’s backlog by this option.

• [httpd] Updated bundled nginx to 1.13.8.

Fixes¶

• Fixed a memory leak in highlight_full

• Fixed a crash bug by early unlink It’s not caused by instruction in grn_expr_parse() but it’s caused when libgroonga user such as Mroonga uses the following instructions:

  1. grn_expr_append_const("_id")

  2. grn_expr_append_op(GRN_OP_GET_VALUE)

Thanks¶

• takagi01

• Naoya Murakami

Improvements¶

• Supported newer version of Apache Arrow. In this release, 0.8.0 or later is required for Apache Arrow support.

• [sharding] Added new API for dynamic columns.

  • Groonga::LabeledArguments

• [sharding] Added convenient Table#select_all method.

• [logical_range_filter] Supported dynamic columns. Note that initial and filtered stage are only supported.

• [logical_range_filter] Added documentation about cache parameter and dynamic columns.

• [logical_count] Supported dynamic columns. Note that initial stage is only supported.

• [logical_count] Added documentation about named parameters.

• [select] Supported --match_columns _key without index.

• [in_values] Supported to specify more than 126 values. [GitHub#760] [GitHub#781] [groonga-dev,04449] [Reported by Murata Satoshi]

• [httpd] Updated bundled nginx to 1.13.7.

Fixes¶

• [httpd] Fixed build error when old Groonga is already installed. [GitHub#775] [Reported by myamanishi3]

• [in_values] Fixed a bug that in_values with too many arguments can cause a crash. This bug is found during supporting more than 126 values. [GitHub#780]

• [cmake] Fixed LZ4 and MessagePack detection. [Reported by Sergei Golubchik]

• [Offline index construction] Fixed a bug that offline index construction for vector column consumes unnecessary resources. If you have a log of elements in one vector column and many records, Groonga will crash. [groonga-dev,04533][Reported by Toshio Uchiyama]

Thanks¶

• Murata Satoshi

• myamanishi3

• Sergei Golubchik

• Toshio Uchiyama

Improvements¶

• [windows] Supported backtrace on crash. This feature not only function call history but also source filename and number of lines can be displayed as much as possible. This feature makes problem solving easier.

• Supported ( ) (empty block) only query (--query "( )") for QUERY_NO_SYNTAX_ERROR. In the previous version, it caused an error. [GitHub#767]

• Supported (+) (only and block) only query (--query "(+)") for QUERY_NO_SYNTAX_ERROR. In the previous version, it caused an error. [GitHub#767]

• Supported ~foo (starting with “~”) query (--query "~y") for QUERY_NO_SYNTAX_ERROR. In the previous version, it caused an error. [GitHub#767]

• Modified log level of expired from info to debug. 2017-10-29 14:05:34.123456|i| <0000000012345678:0> expired i=000000000B123456 max=10 (2/2) This message is logged when memory mapped area for index is unmapped. Thus, this log message is useful information for debugging, in other words, as it is unnecessary information in normal operation, we changed log level from info to debug.

• Supported Ubuntu 17.10 (Artful Aardvark)

Fixes¶

• [dat] Fixed a bug that large file is created unexpectedly in the worst case during database expansion process. This bug may occurs when you create/delete index columns so frequently. In 7.0.7 release, a related bug was fixed - “table_create command fails when there are many deleted keys”, but it turns out that it is not enough in the worst case.

• [logical_select] Fixed a bug that when offset and limit were applied to multiple shards at the same time, there is a case that it returns a fewer number of records unexpectedly.

Improvements¶

• Supported + only query (--query "+") for QUERY_NO_SYNTAX_ERROR. In the previous version, it caused an error.

• [httpd] Updated bundled nginx to 1.13.5.

• [dump] Added the default argument values to the syntax section.

• [Command version] Supported --default-command-version 3.

• Supported caching select result with function call. Now, most of existing functions supports this feature. There are two exception, when now() and rand() are used in query, select result will not cached. Because of this default behavior change, new APIs are introduced.

  • grn_proc_set_is_stable()

  • grn_proc_is_stable()

  Note that if you add a new function that may return different result with the same argument, you must call grn_proc_is_stable(ctx, proc, GRN_FALSE). If you don’t call it, select result with the function call is cached and is wrong result for multiple requests.

Fixes¶

• [windows] Fixed to clean up file handle correctly on failure when database_unmap is executed. There is a case that critical section is not initialized when request is canceled before executing database_unmap. In such a case, it caused a crach bug.

• [Tokenizers] Fixed document for wrong tokenizer names. It should be TokenBigramIgnoreBlankSplitSymbolAlpha and TokenBigramIgnoreBlankSplitSymbolAlphaDigit.

• Changed not to keep created empty file on error.

  In the previous versions, there is a case that empty file keeps remain on error.

  Here is the senario to reproduce:

  1. creating new file by grn_fileinfo_open succeeds

  2. mapping file by DO_MAP() is failed

  In such a case, it causes an another error such as “already file exists” because of the file which isn’t under control. so these file should be removed during cleanup process.

• Fixed a bug that Groonga may be crashed when search process is executed during executing many updates in a short time.

• [table_create] Fixed a bug that table_create failed when there are many deleted keys.

Improvements¶

• Supported prefix match search using multiple indexes. (e.g. --query "Foo*" --match_columns "TITLE_INDEX_COLUMN||BODY_INDEX_COLUMN").

• [window_count] Supported window_count function to add count data to result set. It is useful to analyze or filter additionally.

• Added the following API

  • grn_obj_get_disk_usage():

  • GRN_EXPR_QUERY_NO_SYNTAX_ERROR

  • grn_expr_syntax_expand_query_by_table()

  • grn_table_find_reference_object()

• [object_inspect] Supported to show disk usage about specified object.

• Supported falling back query parse feature. It is enabled when QUERY_NO_SYNTAX_ERROR flag is set to query_flags. (this feature is disabled by default). If this flag is set, query never causes syntax error. For example, “A +” is parsed and escaped automatically into “A +”. This behavior is useful when application uses user input directly and doesn’t want to show syntax error to user and in log.

• Supported to adjust score for term in query. “>”, “<”, and “~” operators are supported. For example, “>Groonga” increments score of “Groonga”, “<Groonga” decrements score of “Groonga”. “~Groonga” decreases score of matched document in the current search result. “~” operator doesn’t change search result itself.

• Improved performance to remove table. thread_limit=1 is not needed for it. The process about checking referenced table existence is done without opening objects. As a result, performance is improved.

• [httpd] Updated bundled nginx to 1.13.4.

Fixes¶

• [dump] Fixed a bug that the 7-th unnamed parameter for –sort_hash_table option is ignored.

• [schema] Fixed a typo in command line parameter name. It should be source instead of sources. [groonga-dev,04449] [Reported by murata satoshi]

• [ruby_eval] Fixed crash when ruby_eval returned syntax error. [GitHub#751] [Patch by ryo-pinus]

Thanks¶

• murata satoshi

• ryo-pinus

Improvements¶

• [httpd] Updated bundled nginx to 1.13.3. Note that this version contains security fix for CVE-2017-7529.

• [load] Supported to load the value of max UInt64. In the previous versions, max UInt64 value is converted into 0 unexpectedlly.

• Added the following API

  • grn_window_get_size() [GitHub#725] [Patch by Naoya Murakami]

• [math_abs] Supported math_abs() function to calculate absolute value. [GitHub#721]

• Supported to make grn_default_logger_set_path() and grn_default_query_logger_set_path() thread safe.

• [windows] Updated bundled pcre library to 8.41.

• [normalize] Improved not to output redundant empty string "" on error. [GitHub#730]

• [functions/time] Supported to show error message when division by zero was happened. [GitHub#733] [Patch by Naoya Murakami]

• [windows] Changed to map ERROR_NO_SYSTEM_RESOURCES to GRN_RESOURCE_TEMPORARILY_UNAVAILABLE. In the previous versions, it returns rc=-1 as a result code. It is not helpful to investigate what actually happened. With this fix, it returns rc=-12.

• [functions/min][functions/max] Supported vector column. Now you need not to care scalar column or vector column to use. [GitHub#735] [Patch by Naoya Murakami]

• [dump] Supported --sort_hash_table option to sort by _key for hash table. Specify --sort_hash_table yes to use it.

• [between] Supported to specify index column. [GitHub#740] [Patch by Naoya Murakami]

• [load] Supported Apache Arrow 0.5.0 or later.

• [How to analyze error messages] Added howto article to analyze error message in Groonga.

• [Debian GNU/Linux] Updated required package list to build from source.

• [Ubuntu] Dropped Ubuntu 16.10 (Yakkety Yak) support. It has reached EOL at July 20, 2017.

Fixes¶

• Fixed to construct correct fulltext indexes against vector column which type belongs to text family (`ShortText and so on). This fix resolves that fulltext search doesn’t work well against text vector column after updating indexes. [GitHub#494]

• [thread_limit] Fixed a bug that deadlock occurs when thread_limit?max=1 is requested at once.

• [groonga-httpd] Fixed a mismatch path of pid file between default one and restart command assumed. This mismatch blocked restarting groonga-httpd. [GitHub#743] [Reported by sozaki]

Thanks¶

• Naoya Murakami

Improvements¶

• Added physical create/delete operation logs to identify problem for troubleshooting. [GitHub#700,#701]

• [in_records] Improved performance for fixed sized column. It may reduce 50% execution time.

• [grndb] Added --log-path option. [GitHub#702,#703]

• [grndb] Added --log-level option. [GitHub#706,#708]

• Added the following API

  • grn_operator_to_exec_func()

  • grn_obj_is_corrupt()

• Improved performance for “FIXED_SIZE_COLUMN OP CONSTANT”. Supported operators are: ==, !=, <, >, <= and >=.

• Improved performance for “COLUMN OP VALUE && COLUMN OP VALUE && …”.

• [grndb] Supported corrupted object detection with grndb check.

• [io_flush] Supported --only_opened option which enables to flush only opened database objects.

• [grndb] Supported to detect/delete orphan “inspect” object. The orphaned “inspect” object is created by renamed command name from inspect to object_inspect.

Fixes¶

• [rpm][centos] Fixed unexpected macro expansion problem with customized build. This bug only affects when rebuilding Groonga SRPM with customized additional_configure_options parameter in spec file.

• Fixed missing null check for grn_table_setoperation(). There is a possibility of crash bug when indexes are broken. [GitHub#699]

Thanks¶

Improvements¶

• [select] Add document about Full text search with specific index name.

• [index] Supported to log warning message which record causes posting list overflows.

• [load][dump] Supported Apache Arrow. [GitHub#691]

• [cmake] Supported linking lz4 in embedded static library build. [Original patch by Sergei Golubchik]

• [delete] Supported to cancel.

• [httpd] Updated bundled nginx to 1.13.0

• Exported the following API

  • grn_plugin_proc_get_caller()

• Added index column related function and selector.

  • Added new selector: index_column_df_ratio_between()

  • Added new function: index_column_df_ratio()

Fixes¶

• [delete] Fixed a bug that error isn’t cleared correctly. It affects to following deletions so that it causes unexpected behavior.

• [windows] Fixed a bug that IO version is not detected correctly when the file is opened with O_CREAT flag.

• [vector_slice] Fixed a bug that non 4 bytes vector columns can’t slice. [GitHub#695] [Patch by Naoya Murakami]

• Fixed a bug that non 4 bytes fixed vector column can’t sequential match by specifying index of vector. [GitHub#696] [Patch by Naoya Murakami]

• [logical_select] Fixed a bug that “argument out of range” occurs when setting last day of month to the min. [GitHub#698]

Thanks¶

• Sergei Golubchik

• Naoya Murakami

Improvements¶

• [logical_select] Supported multiple drilldowns[${LABEL}].columns[${NAME}].window.sort_keys and drilldowns[${LABEL}].columns[${NAME}].window.group_keys.

• [windows] Updated bundled LZ4 to 1.7.5.

• [cache] Supported persistent cache feature.

• [log_level] Update English documentation.

• Added the following APIs:

  • grn_set_default_cache_base_path()

  • grn_get_default_cache_base_path()

  • grn_persistent_cache_open()

  • grn_cache_default_open()

• [groonga --cache-base-path] Added a new option to use persistent cache.

• [groonga-httpd] [groonga_cache_base_path] Added new configuration to use persistent cache.

• [windows] Updated bundled msgpack to 2.1.1.

• [object_inspect] Supported not only column inspection, but also index column statistics.

• Supported index search for “.*” regexp pattern. This feature is enabled by default. Set GRN_SCAN_INFO_REGEXP_DOT_ASTERISK_ENABLE=no environment variable to disable this feature.

• [in_records] Added function to use an existing table as condition patterns.

• [Ubuntu] Dropped Ubuntu 12.04 (Precise Pangolin) support because of EOL.

Fixes¶

• [logical_select] Fixed a bug that wrong cache is used. This bug was occurred when dynamic column parameter is used.

• [logical_select] Fixed a bug that dynamic columns aren’t created. It’s occurred when no match case.

• [reindex] Fixed a bug that data is lost by reindex. [GitHub#646]

• [httpd] Fixed a bug that response of quit and shutdown is broken JSON when worker is running as another user. [GitHub ranguba/groonga-client#12]

Improvements¶

• Exported the following API

  • grn_ii_cursor_next_pos()

  • grn_table_apply_expr()

  • grn_obj_is_data_column()

  • grn_obj_is_expr()

  • grn_obj_is_scalar_column()

• [dump] Supported to dump weight reference vector.

• [load] Supported to load array<object> style weight vector column. The example of array<object> style is: [{"key1": weight1}, {"key2": weight2}].

• Supported to search !(XXX OPERATOR VALUE) by index. Supported operator is not only > but also >=, <, <=, == and !=.

• Supported index search for “!(column == CONSTANT)”. The example in this case is: !(column == 29) and so on.

• Supported more “!” optimization in the following patterns.

  • !(column @ "X") && (column @ "Y")

  • (column @ "Y") && !(column @ "X")

  • (column @ "Y") &! !(column @ "X")

• Supported to search XXX || !(column @ "xxx") by index.

• [dump] Changed to use '{"x": 1, "y": 2}' style for not referenced weight vector. This change doesn’t affect to old Groonga because it already supports one.

• [experimental] Supported GRN_ORDER_BY_ESTIMATED_SIZE_ENABLE environment variable. This variable controls whether query optimization which is based on estimated size is applied or not. This feature is disabled by default. Set GRN_ORDER_BY_ESTIMATED_SIZE_ENABLE=yes if you want to try it.

• [select] Added query log for columns, drilldown evaluation.

• [select] Changed query log format for drilldown. This is backward incompatible change, but it only affects users who convert query log by own programs.

• [table_remove] Reduced temporary memory usage. It’s enabled when the number of max threads is 0.

• [select] columns[LABEL](N) is used for query log format instead of columns(N)[LABEL]..

• [Query expansion] Updated example to use vector column because it is recommended way. [Reported by Gurunavi, Inc]

• Supported to detect canceled request while locking. It fixes the problem that request_cancel is ignored unexpectedly while locking.

• [logical_select] Supported initial and filtered stage dynamic columns. The examples are: --columns[LABEL].stage initial or --columns[LABEL].stage filtered.

• [logical_select] Supported match_columns, query and drilldown_filter option.

• [highlight_html] Supported similar search.

• [logical_select] Supported initial and stage dynamic columns in labeled drilldown. The example is: --drilldowns[LABEL].stage initial.

• [logical_select] Supported window function in dynamic column.

• [select] Added documentation about dynamic columns.

• [Window function] Added section about window functions.

• [CentOS] Dropped CentOS 5 support because of EOL.

• [httpd] Updated bundled nginx to 1.11.12

• Supported to disable AND match optimization by environment variable. You can disable this feature by GRN_TABLE_SELECT_AND_MIN_SKIP_ENABLE=no. This feature is enable by default.

• [vector_new] Added a new function to create a new vector.

• [select] Added documentation about drilldown_filter.

Fixes¶

• [lock_clear] Fixed a crash bug against temporary database.

• Fixed a problem that dynamically updated index size was increased for natural language since Grooonga 6.1.4.

• [select] Fixed a bug that “A && B.C @ X” may not return records that should be matched.

• Fixed a conflict with grn_io_flush() and grn_io_expire(). Without this change, if io_flush and load command are executed simultaneously in specific timing, it causes a crash bug by access violation.

• [logical_table_remove] Fixed a crash bug when the max number of threads is 1.

Thanks¶

• Gurunavi, Inc.

Improvements¶

• [in_values] Supported sequential search for reference vector column. [Patch by Naoya Murakami] [GitHub#629]

• [select] Changed to report error instead of ignoring on invalid drilldown[LABEL].sort_keys.

• [select] Removed needless metadata updates on DB. It reduces the case that database lock remains even though select command is executed. [Reported by aomi-n]

• [lock_clear] Changed to clear metadata lock by lock_clear against DB.

• [CentOS] Enabled EPEL by default to install Groonga on Amazon Linux.

• [query] Supported “@X” style in script syntax for prefix(“@^”), suffix(“@$”), regexp(“@^”) search.

• [query] Added documentation about available list of mode. The default mode is MATCH (“@”) mode which executes full text search.

• [rpm][centos] Supported groonga-token-filter-stem package which provides stemming feature by TokenFilterStem token filter on CentOS 7. [GitHub#633] [Reported by Tim Bellefleur]

• [window_record_number] Marked record_number as deprecated. Use window_record_number instead. record_number is still available for backward compatibility.

• [window_sum] Added window_sum window function. It’s similar behavior to window function sum() on PostgreSQL.

• Supported to construct offline indexing with in-memory (temporary) TABLE_DAT_KEY table. [GitHub#623] [Reported by Naoya Murakami]

• [onigmo] Updated bundled Onigmo to 6.1.1.

• Supported columns[LABEL].window.group_keys. It’s used to apply window function for every group.

• [load] Supported to report error on invalid key. It enables you to detect mismatch type of key.

• [load] Supported --output_errors yes option. If you specify “yes”, you can get errors for each load failed record. Note that this feature requires command version 3.

• [load] Improve error message on table key cast failure. Instead of “cast failed”, type of table key and target type of table key are also contained in error message.

• [httpd] Updated bundled nginx to 1.11.9.

Fixes¶

• Fixed a bug that nonexistent sort keys for drilldowns[LABEL] or slices[LABEL] causes invalid JSON parse error. [Patch by Naoya Murakami] [GitHub#627]

• Fixed a bug that access to nonexistent sub records for group causes a crash. For example, This bug affects the case when you use drilldowns[LABEL].sort_keys _sum without specifying calc_types. [Patch by Naoya Murakami] [GitHub#625]

• Fixed a crash bug when tokenizer has an error. It’s caused when tokenizer and token filter are registered and tokenizer has an error.

• [window_record_number] Fixed a bug that arguments for window function is not correctly passed. [GitHub#634][Patch by Naoya Murakami]

Thanks¶

• Naoya Murakami

• aomi-n