Changelog

Raijin Database 2.1.4621

[1835] Prohibit usage of explicit _id field.
[2249] Partition filtering during scans.
[2280] Validate configuration parameter inputs.
[2281] Auto-adjust memory related configuration parameters.
[2330] FLOAT is now a 4-byte-width datatype (previously it was an alias to DOUBLE).
[2331] Add endpoint for changing table partitioning function.
[2475] Partition pruning for ORDER BY [LIMIT/OFFSET] on partitioned column.
[2492] Auto-adjusting of params that are greater than the amount of system memory available.
[2188] Fix: crash when complex aggregator specified in subquery.
[2287] Fix: Raijin fails to start after GRANT with dots in db/table name.
[2385] Fix: crash after getting an exception when audit logging is ON.
[2395] Fix: Superset does not recognize INT, BIGINT data types.
[2422] Fix: Raijin cannot detect diskfull in case of specific inode quota.
[2473] Fix: MATCH AGAINST crash.
[2487] Fix: IMATCH misbehaves with UTF fields.

[2248] Redesigned partitioning.
[2299] Fix: Missing heap.desc files when flushing tables.
[2351] Continue sorting with common algorithm after running out of memory in its optimized version.
[2375] Update Raijin OpenAPI to match the current implementation.

[1237] Fix: Failed check in TargetList on aggregated scan with correlated subquery.
[2115] Use the data directory as the base that contains all the stateful files.
[2183] Implementation of LIKE/ILIKE without using regexps.
[2256] Check locking status in Flush tables with read lock.
[2261] Support Authorization header in API.
[2311] Fix: Metrics endpoint requires authentication when it should not.

[1271] Change behaviour of aggregators with variant data.
[1804] Implement SHOW VERSION command.
[1862] Increase column name length limit.
[1926] Fix: error when stop/restart Raijin with locked tables.
[1938] Fix: bigint and int columns become variant when signed data is ingested.
[2015] Extend partition management endpoints to target partitions by their directory names.
[2063] Fix: swagger ignores https.
[2070] Implement SHOW VIEWS command.
[2092] MATCH predicate now throws an error when used on non-string columns.
[2154] Fix: SELECT FIELDS for views causes internal error.
[2163] Fix: importing large data set leads to "Too many open files" error.
[2191] Implement SHOW USERS command.
[2236] New sorting algorithm for ORDER BY col LIMIT n OFFSET m.
[2238] Unordered selects.
[2240] Fix: crash when selecting COUNT(DISTINCT) on large dataset.
[2313] Fix: Memory limit exceeded during ORDER BY with LIMIT.

[1923] Implement GRANT command.
[1976] Implemented CREATE/DROP USER sql statements.
[1977] Username/password authentication. Fully anonymous access disallowed.
[2047] Optimization: do not merge partition’s data before sorting or aggregating.
[2014] Fixed memory leaks in COPY TO.
[2083] Reduce memory consumption during scans.
[2089] Detect CGroups memory limits.
[2190] Enable anonymous access.
[2225] Optimized ORDER BY with LIMIT memory usage.