Version 5.3.0

Released on 2023-04-04.

Warning

CrateDB 5.3.x versions up to 5.3.3 (excluding) contain a critical bug which can lead to data corruption/loss when using a column definition with a number data type and disabled index (INDEX OFF). It is not recommended to use those versions, use CrateDB >= 5.3.3 instead.

Note

If you are upgrading a cluster, you must be running CrateDB 4.0.2 or higher before you upgrade to 5.3.0.

We recommend that you upgrade to the latest 5.2 release before moving to 5.3.0.

Before upgrading, you should back up your data.

Warning

Due to a bug in the replication layer, rolling upgrades from 5.2.x to 5.3.0 with ongoing write traffic can lead to corrupted shards and in worse case, data loss. We recommend that you stop all write traffic before upgrading and/or perform a full cluster restart.

Warning

Tables that were created before CrateDB 4.x will not function with 5.x and must be recreated before moving to 5.x.x.

You can recreate tables using COPY TO and COPY FROM or by inserting the data into a new table.

Table of Contents

Breaking Changes

  • Removed support for SET LICENSE. The statement had no effect since CrateDB 4.5 and was only kept for backward compatibility.

Deprecations

None

Changes

SQL Standard And PostgreSQL Schema Compatibility

  • Changed the behavior of SHOW search_path to omit the implicit pg_catalog schema, unless the user set it explicitly. This matches the PostgreSQL behavior.

  • Allowed schema and table names to contain upper case letters. This can be achieved by quoting the names. Unquoted names with upper case letters are converted to lower cases which has been the existing behaviour.

  • Allowed schema and table names to start with _.

  • Added the col_description(integer, integer) scalar function for improved PostgreSQL compatibility. CrateDB does not support comments for columns, so this function always returns NULL.

SQL statements

  • Changed the behavior of INSERT INTO on tables with generated columns which use non-deterministic functions. Previously, if the columns were included in the column target list, the provided values were validated against re-computed values of the generated expression. Given that the functions are not deterministic, this validation always failed and made it impossible to copy data between tables with such columns.

    In the new behavior the values are no longer validated for non-deterministic functions but are accepted as is. If the column is not present in the target list, they are re-computed as before. This lets users do copy operations between such tables and decide if the source value should be used as is, or if they should be re-computed by either adding or removing the columns from the INSERT INTO column target list.

Performance Improvements

  • Optimized the evaluation of CASE expressions to prevent stack overflows for very large expressions.

  • Improved the performance of queries using a correlated sub-query inside the WHERE clause in conjunction with a non-correlated filter clause.

  • Improved performance of statements that create multiple partitions at once, which can occur during COPY FROM or INSERTS with multi-values into partitioned tables.

  • Improved ingestion performance by up to 30%.