BloGroonga

2016-10-29

Groonga 6.1.0 has been released!

Groonga 6.1.0 has been released!

How to install in each environment: Install

In this release, documents about index columns has been added. Using this column type makes reducing memory usage.

Changes

Here are important changes in this release:

  • Supported to sort load table names by key instead of ID to make output stable
  • Added document about index columns

Supported to sort load table names by key instead of ID to make output stable

Before this release, dump command result of tables had been sorted by ID. Thus, when changing schema by adding and removing tables, it sometimes happens that a result of dumping database differs from without changing schema database. (Even if both of final definitions are same.) This is not preferable behavior to compare dumping results.

Then, in this release, dumping results are output to be sorted by table name. When they contain reference type columns, it will be taken into consideration. This is incompatible change but it makes easy to compare result of dump command.

Added document about index columns

In this release, documents added about small/medium index columns. This feature had been implemented in earlier version, but it had been missing its concrete usage document.

Then, in this time, we describe its usage. This method can apply indexing target to be reasonably small within data saved in database. If you make a good use of this feature, it makes reducing memory usage. There are two ways to use it.

  • small index column
  • medium index column

Small index column can use when this column usage will not reach 1 billion records. It will reduce memory consumption by 1/256 compared with default index column's usage. For example, specify COLUMN_INDEX|INDEX_SMALL in flags parameter as follows:

column_create --table Ages --name people_age_small_index --flags COLUMN_INDEX|INDEX_SMALL --type People --source age

It can apply when number of People table records is at most a few million.

When using medium size index column, it can be made number of 1 billion records which contains a fixed size scalar column. It reduces memory consumption by 5/24 compared with default index column's usage. There are some constraints to apply this type of column. In more detail, please refer to the Create a medium index column section.

For example, specify COLUMN_INDEX|INDEX_MEDIUM in flags parameter as follows:

column_create --table Ages --name people_age_medium_index --flags COLUMN_INDEX|INDEX_MEDIUM --type People --source age

Conclusions

Please refer to Release 6.1.0 - 2016-10-29 about detailed changes since 6.0.9.

Then, let's go all out to search by Groonga!