* upstream/master:
fix left asof join with join_use_nulls (#13291)
Revert "Add QueryTimeMicroseconds, SelectQueryTimeMicroseconds and InsertQuer…"
Add test
Added another check just in case
Fixup
Fix range checks in h3 functions
Fix configs
Update tests
Check ALTERs; update test
Sanity checks for MergeTreeSettings
* upstream/master:
Update tips.md
Wait with retries in fast test (#13284)
update #syntax-expressions in syntax.md (#13269)
fix test
Set custom prefix in server-test.xml too.
Update HashTable.h
Attempt to fix "Arcadia" build
More performance tests
Fix error + one more optimization
Attempt to fix "Arcadia" build
Add test
Continue
Fix hash table with NaN keys
Fix tests
Fix errors
Fix errors
Revert some changes
shellcheck --check-sourced --external-sources --severity style --exclude SC1071 --format diff *.sh ../1_stateful/*.sh | git apply
Fix all info from Shellcheck
fgrep -> grep -F
shellcheck --check-sourced --external-sources --severity info --exclude SC1071 --format diff */*.sh | git apply
Minor changes 2
Minor changes
Fix warnings found by Shellcheck
Refresh ya.make
Better code
Fix assertion in KeyCondition
Fix including Common/config.h in the Arcadia build.
Fix bad code
Add more threads for the 01323_too_many_threads_bug
Exclude threads used for query_log processing 01323_too_many_threads_bug
Use ILIKE over lower() LIKE in 01323_too_many_threads_bug
Update test
Fix style
Simplify CMake for libunwind (suggestion from @azat)
Normalize ya.make files, fix "Arcadia" build
Fix 01323_too_many_threads_bug (use arrayUniq() over length())
Fix logging in MergeTreeDataSelectExecutor for multiple threads (attach to thread group)
Use "Not using primary index on part" over "Not using index on part" (add "primary")
Exclude 01413_rows_events from the fasttest (TRUNCATE query_log)
fasttest: use arrays for skip list (to make it more patch friendly)
Update README.md
Fix warning in libunwind
Fix warning in libunwind
Fix assert in geohashesInBox
Fix assert in parseDateTimeBestEffort
Add test
Fix function if with nullable constexpr as cond that is not literal NULL
Fix assert in *Map aggregate functions
More tests
Fix assert in arrayElement
Fix assert in IN with tuple literals and functions
Use stored size for UInt16 case
Revert "Less indirection in FixedHashTable"
Special case for aggregation by 8bit field
Less indirection in FixedHashTable
Update FixedHashTable.h
Avoid increment of size in the inner loop of aggregation by UInt8/UInt16
Trigger CI
Slightly better
Fix fuzz test
Add test
Function "materialize" should work for NULL as expected
Minor changes in Debian scripts
Fix Shellcheck errors
Fix some spurious perf test failures
Fix test
+x for test
Merge with master
Added and updated tests
More sane
Remove ridiculous trash
Use absolute paths to backup locations (shadow)
Add test for macros usage in certain kafka setting.
Add QueryTimeMicroseconds, SelectQueryTimeMicroseconds and InsertQueryTimeMicroseconds to system.events
try fix sync
Fix various code defects in groupArrayMoving functions #12568
fix
fixes
fixes
implement reneme dictionary query
rename database with dictionaries
fix tests
use dictionary ids
fix build
add uuid to generated config
add StorageID to IDictionaryBase 3
add StorageID to IDictionaryBase 2
add StorageID to IDictionaryBase 1
write uuid for ddl dictionaries
fixes
add implicit uuid macro
fix tests
fix, add test
implement rename database for atomic
fix symlinks, add short syntax of attach database
add uuid to atomic databases
Since rdkafka 1.5 (https://github.com/edenhill/librdkafka/releases/tag/v1.5.0)
don't auto-create topics when subscribing, so test start failing
because it was intentionally time-limited to check if flushing by block size worked.
Since 20.5 it's possible to increase kafka_flush_interval_ms per table, so i've
adjusted the test using that feature to test if data is flushed by block size,
not by time.
