bc699511d0
[Fix](array-function) fix array_distinct null values ( #14544 )
...
in the previous the result is:
```
mysql> select array_distinct([1,1,3,3,null, null, null]);
+-----------------------------------------------------+
| array_distinct(ARRAY(1, 1, 3, 3, NULL, NULL, NULL)) |
+-----------------------------------------------------+
| [1, 3, NULL, NULL, NULL] |
+-----------------------------------------------------+
1 row in set (0.00 sec)
```
after this fix, the result becomes:
```
mysql> select array_distinct([1,1,3,3,null, null, null]);
+-----------------------------------------------------+
| array_distinct(ARRAY(1, 1, 3, 3, NULL, NULL, NULL)) |
+-----------------------------------------------------+
| [1, 3, NULL] |
+-----------------------------------------------------+
1 row in set (0.00 sec)
```
2022-11-24 19:07:28 +08:00
ac46922433
[fix](ut) Fix failures for BE UT macOS ( #14543 )
2022-11-24 17:39:37 +08:00
0c4830600d
test(grouping sets) add regression test case for grouping sets ( #14539 )
...
Co-authored-by: tongyang.hty <hantongyang@douyu.tv >
2022-11-24 17:38:12 +08:00
59b31a03c4
[Improvement](agg function) support group_bit_and/group_bit_or/group_bit_xor functions ( #14386 )
2022-11-24 16:46:42 +08:00
608cb6c4ad
[test](jdbc)add new case for mysql external table ( #14530 )
2022-11-24 16:36:44 +08:00
b4d8ae5204
[test](jdbc)add new pg case from other source ( #14445 )
2022-11-24 16:35:59 +08:00
a04e1b49ec
[feature](Nereids) Implement group by grouping sets, cube and rollup ( #14496 )
...
Issue Number: close #13615
The main work:
implement grouping sets/ cube/ rollup.
fix if function Infinite loop problem.
Support for isNull transitions to legacy optimizers.
2022-11-24 16:34:31 +08:00
0680b3b4d5
[opt](nereids) adjust nereids related regression test cases ( #14439 )
...
1. in dateV2, we adjust the dir structure to avoid creating a tpch-1G database
2. use `drop table XXX` to replace `delete * from XXX where key>0`
3. remove explain cases, because
- the explain string itself is variable, and the case is hard to maintain
- it is original planner explain, not nereids
2022-11-24 16:02:52 +08:00
fde474609e
[feature](Nereids) Add dphyp job ( #14485 )
2022-11-24 15:50:05 +08:00
8afe298a0f
[Fix](function) fix function retention lost ARRAY's element type … ( #14538 )
2022-11-24 15:19:50 +08:00
2389a90cd0
[enhancement](snapshot) add missed version log when make_snapshot in engine clone task ( #14284 )
2022-11-24 14:51:28 +08:00
7f4cc61286
[fix](cast)prevent be from crashing when cast function is not available ( #14540 )
...
* [fix](cast)prevent be from crashing when cast function is not available
* format code
2022-11-24 14:17:49 +08:00
6c7f758ef7
[improvement](hashjoin) support partitioned hash table in hash join ( #14480 )
2022-11-24 14:16:47 +08:00
e656dae3f0
[fix](fe) fix leaks of connect context ( #14529 )
...
Remove ConnectContext which built for internal statistics from threadlocal to avoid memory leaks
2022-11-24 13:26:59 +08:00
6472d5506f
[fix](cache) fix cache overflow problem #14515 ( #14516 )
...
Co-authored-by: wangxiangyu@360shuke.com <wangxiangyu@360shuke.com >
2022-11-24 11:18:46 +08:00
ae4f4b9bf1
[fix](agg)having clause should use column name first then alias ( #14408 )
...
* [fix](agg)having clause should use column name first then alias
* fix fe ut
2022-11-24 10:31:58 +08:00
f6de03eb6c
[chore](github) Add a workflow to check the build for third-party libraries ( #14533 )
...
Currently, we build the third-party libraries and release them automatically (See https://github.com/apache/doris-thirdparty/pull/13 ). We must make sure that the changes for third-party libraries are valid.
2022-11-24 10:07:39 +08:00
70ea07bc4b
[fix](nullable) Fix nullable cache to avoid function returning wrong value ( #14463 )
2022-11-24 09:35:08 +08:00
6ccdaf0aaf
[fix](storage-policy) use Long instead of Date to persiste cooldowntime in storage policy ( #14532 )
...
Previously, we use "Date" type for cooldownTime in StoragePolicy.
But the serialization method of Date type in Gson is different in java8 and java11, which may cause inconsistent meta error.
This PR use Long to save cooldownTime.
And notice that in FE, the cooldownTime is saved in milliseconds, and in BE, it is saved in seconds.
2022-11-24 08:32:21 +08:00
724e57bb87
[feature](docker)Add runtime docker image related files ( #14436 )
2022-11-23 23:58:44 +08:00
496a92b668
[JavaUDF](loader) Fix compatible problem for JAVA 11 ( #14519 )
2022-11-23 23:36:39 +08:00
404cac42f9
[fix](multi catalog)Fix external table partition name and type inconsistent bug. ( #14522 )
...
Origin code using Set to store hms external table partition columns,
which couldn't guarantee the order of the columns.
This could cause the column name and column type doesn't match.
Using List instead of Set to fix the problem.
2022-11-23 21:40:44 +08:00
9e39a04b63
[Doc](flink connector) add flink connector faq ( #14520 )
2022-11-23 21:40:00 +08:00
181f1cf176
[Docs](function) add some missing function docs ( #14510 )
2022-11-23 21:39:17 +08:00
6770bfc7f0
[fix](pipeline) adjust mem limit to 30% ( #14523 )
2022-11-23 20:07:45 +08:00
648fd93dc5
[DOCS](function) add document for grouping and grouping_id ( #14472 )
2022-11-23 18:07:48 +08:00
d14e1d25ff
[Bug](vectorized) Fix wrong column type ( #14387 )
2022-11-23 18:07:33 +08:00
1520e5c88a
[enhancement](agg)use new method to serialize keys in batch if the key is too large ( #14484 )
...
* [enhancement](agg)use new method to serialize keys in batch if the key is too large
* fix compile error
2022-11-23 17:35:39 +08:00
fd3af489a4
[memory](chunkallocator) disable chunkallocator when reserved bytes == 0 ( #14494 )
...
disable chunkallocator when reserved bytes == 0
disable chunkallocator by default
2022-11-23 17:12:53 +08:00
8d5eabb64f
[enhancement](Nereids) reduce CostAndEnforcerJob call times ( #14442 )
...
record pruned plan's cost to avoid optimize same GroupExpression more than once.
2022-11-23 16:57:41 +08:00
388f067300
[chore](workflow) Disable memory tracker by default on BE UT (macOS) ( #14508 )
2022-11-23 16:25:42 +08:00
6fcffd041c
[test](jdbc)add new mysql jdbc case from other source ( #14495 )
2022-11-23 16:23:42 +08:00
09cc385caa
[Docs](fucntion) Add docs for function random, mod, fmod ( #14444 )
2022-11-23 16:22:57 +08:00
45975dd321
[enhancement](Nereids): Change circle detector for better performance ( #14438 )
2022-11-23 14:31:14 +08:00
7a7e714fce
[fix](nereids) width and penalty not derive when do stats derive ( #14474 )
...
a previous pr (#13883 ) refactor stats derive code, but missed width and penalty.
2022-11-23 14:26:51 +08:00
fb385dcf23
[opt](nereids) make fragment id in explain get inline with profile ( #14421 )
...
Nereids assign fragment ID in its own way. The fragment Id in explain is different from the fragment id in profile.
This difference makes trouble to understand profile.
This pr aims to print fragment id in explain the same as that in profile.·
2022-11-23 14:14:20 +08:00
7955e52b3e
[fix](version) fix recover bug for lower version ( #14457 )
2022-11-23 14:05:17 +08:00
672d3b8334
[Document](function) update document of to_base64/from_base64 #14487
2022-11-23 14:04:53 +08:00
79688c34a1
[feature](catalog) add max num of same name meta information in catalog recycle bin ( #14482 )
2022-11-23 14:04:14 +08:00
9d165c6bba
[docs](fe.conf) add configuration for decommission_tablet_check_threshold. ( #14355 )
2022-11-23 12:58:52 +08:00
0074f55f96
[refactor](array-type) Remove encoding info for array type ( #14498 )
...
Array column should not have encoding info because it use its sub columns' encoding info
And this encoding info is never used and easy to make people confused.
We should remove it.
2022-11-23 11:45:47 +08:00
d36b561520
[fix](in)fix in predicate datatype mismatch after union ( #14497 )
2022-11-23 09:57:03 +08:00
3b5f4ad198
[fix](unique-key-merge-on-write) fix that unique key with mow may loss some data in the query result with predicates ( #14455 )
...
When unique key with MOW table has sequence column, the query result may be wrong with predicates. There are two problems:
The sequence column needs to be removed from primary key index when comparing key.
The sequence column needs to be removed from min/max key.
2022-11-23 09:08:07 +08:00
249b688663
[chore](github) Add a workflow to check BE UT on macOS ( #14506 )
2022-11-23 08:38:28 +08:00
9ec66f939a
[fix](docs) fix the docs of command "show proc '/statistic'" and ADMIN CHECK TABLET ( #14282 )
...
Co-authored-by: smallhibiscus <844981280>
2022-11-22 21:53:11 +08:00
6124d0e0b7
[fix](doc) add doc description with multi partition ( #14262 )
2022-11-22 21:51:55 +08:00
2eca51f3ba
[enhancement](broker) broker load support tencent cos ( #12801 )
2022-11-22 21:51:15 +08:00
6eeebd47a9
[improvement](doc) add missing documents ( #14460 )
2022-11-22 21:42:00 +08:00
ab8346560d
[Enhancement](storage) add num_values consistency check when build/load IndexedColumn ( #14447 ) ( #14450 )
2022-11-22 21:37:08 +08:00
18b9db17b3
[fix](test) move cases in query to query_p0 ( #14452 )
2022-11-22 21:35:18 +08:00
