Sourced from org.questdb:questdb's releases.

9.2.3

QuestDB 9.2.3 is a patch release bringing the usual swathe of fixes, performance enhancements, and a few new features. You'll see higher QPS, faster GROUP BY queries, and better performance for high-contention materialized views.

There will be one more release before Christmas, bringing the new WINDOW JOIN syntax. Stay tuned!

For any questions or feedback, please join us on Slack or on Discourse.

See also our prettier release notes page.

Highlights

• Materialized Views
  • Faster updates under high contention: views which update frequently (multiple times per second) will now use significantly less resources and have reduced refresh latency.
  • Faster historical refreshes: views with very small SAMPLE BY units will now refresh much more quickly, especially for historical or full refreshes.
• TTL
  • Additional safety checks: TTL will now use min(maxTimestamp, wallClockTime) as the predicate for dropping old partitions, rather than just maxTimestamp. This avoids accidental partition drops when erroneous timestamps are written to the table.
• SQL
  • Faster (de)allocations: Memory deallocations have been sped up significantly, with jemalloc now enabled by default on Linux builds. This particularly helps for reducing query tail latencies.
  • Faster query compilation: Query parsing performance has been improved by around ~4x. This reduces latency for lightweight and fast-to-execute queries, and maximises concurrent QPS when prepared statements are not used.
  • Faster single-key GROUP BY: GROUP BY queries that group by a single key will now execute ~2-3x faster.
    • e.g. SELECT symbol, count() as count FROM trades GROUP BY symbol ORDER BY count DESC LIMIT 10;
  • Faster non-keyed averages: Keyless GROUP BY queries will now execute ~1.3x faster.
    • e.g. SELECT AVG(price) FROM trades;
  • Decimal fills: SAMPLE BY with FILL now properly supports the new DECIMAL type.

Changelist

• fix(core): fix missing partition dir on table reader open when writing identical data with dedup by @​ideoma in questdb/questdb#6479
• fix(sql): crash when using parallel ORDER BY with LIMIT and JIT filter by @​puzpuzpuz in questdb/questdb#6482
• fix(ui): remove line ending declarations from web console assembly descriptor by @​emrberk in questdb/questdb#6490
• feat(core): hints for handling histogram out of bounds error for approx_percentile()/approx_median() by @​peter007-cmd in questdb/questdb#6446
• perf(sql): use unordered maps only in single-column case by @​puzpuzpuz in questdb/questdb#6481
• fix(sql): fix parquet export bug when SQL contains now() function by @​kafka1991 in questdb/questdb#6499
• fix(sql): fix empty result set from limit queries by @​bluestreak01 in questdb/questdb#6504
• fix(sql): fix potential crash when ORDER BY symbol column added via ALTER TABLE ADD COLUMN by @​kafka1991 in questdb/questdb#6505
• fix(sql): improve refresh step estimation for sparse data with small sample by interval by @​kafka1991 in questdb/questdb#6517
• chore(utils): update async-profiler to 4.2.1 by @​jerrinot in questdb/questdb#6473
• fix(core): data corruption after range replace on column top partition by @​puzpuzpuz in questdb/questdb#6519
• fix(http): correct HTTP content length for JSON responses by @​kafka1991 in questdb/questdb#6520
• perf(core): optimize mat view writing by skipping transactions that are fully replaced by future commits or full refresh (2) by @​ideoma in questdb/questdb#6495
• fix(core): fix bad connection state after TLS session initialization failure by @​jerrinot in questdb/questdb#6529
• fix(sql): fix potential unexpected exception when parquet file schema changed in read_parquet() by @​kafka1991 in questdb/questdb#6513
• feat(core): use wall clock for TTL to prevent accidental data loss by @​bluestreak01 in questdb/questdb#6531
• fix(sql): support for decimal types in group by fill and selected cursor by @​RaphDal in questdb/questdb#6532
• fix(sql): not to ignore outer limit if a subquery has a filter and another limit by @​glasstiger in questdb/questdb#6533
• perf(sql): reduce query latency from memory deallocation overhead on Linux x86_64 by @​jerrinot in questdb/questdb#6477
• fix(sql): fix symbol column nullValue flag when adding column to non-empty table by @​kafka1991 in questdb/questdb#6507
• perf(sql): improve latency of parsing SQL expressions by @​bluestreak01 in questdb/questdb#6538
• perf(sql): speed up avg() functions in non-keyed GROUP BY queries by @​puzpuzpuz in questdb/questdb#6542
• fix(core): UPDATE to work with bind variables for TIMESTAMP columns by @​glasstiger in questdb/questdb#6510
• fix(sql): allow aggregate functions inside CASE expressions by @​bluestreak01 in questdb/questdb#6550

... (truncated)

• d655b29 [maven-release-plugin] prepare release 9.2.3
• a800ac5 fix(core): fix dedup enable on already deduplicated table to not suspend the ...
• 2f74209 fix(sql): allow aggregate functions inside CASE expressions (#6550)
• 40367ef fix(core): UPDATE to work with bind variables for TIMESTAMP columns (#6510)
• eee1923 test(http): fix flaky testPeerIsSlowToReadPar() (#6545)
• b7dfaa8 chore(sql): improve error message when large sample by queries exceed page si...
• 8021d57 perf(sql): speed up avg() functions in non-keyed GROUP BY queries (#6542)
• 94a578f perf(sql): improve latency of parsing SQL expressions (#6538)
• 3913811 fix(sql): fix symbol column nullValue flag when adding column to non-empty t...
• 39f4bbb perf(sql): reduce query latency from memory deallocation overhead on Linux x8...
• Additional commits viewable in compare view

You can trigger Dependabot actions by commenting on this PR:

• @dependabot rebase will rebase this PR
• @dependabot recreate will recreate this PR, overwriting any edits that have been made to it
• @dependabot merge will merge this PR after your CI passes on it
• @dependabot squash and merge will squash and merge this PR after your CI passes on it
• @dependabot cancel merge will cancel a previously requested merge and block automerging
• @dependabot reopen will reopen this PR if it is closed
• @dependabot close will close this PR and stop Dependabot recreating it. You can achieve the same result by closing it manually
• @dependabot show <dependency name> ignore conditions will show all of the ignore conditions of the specified dependency
• @dependabot ignore this major version will close this PR and stop Dependabot creating any more for this major version (unless you reopen the PR or upgrade to it yourself)
• @dependabot ignore this minor version will close this PR and stop Dependabot creating any more for this minor version (unless you reopen the PR or upgrade to it yourself)
• @dependabot ignore this dependency will close this PR and stop Dependabot creating any more for this dependency (unless you reopen the PR or upgrade to it yourself)