owner stringclasses 14
values | repo stringclasses 14
values | id int64 116k 4.61B | issue_number int32 1 181k | author stringlengths 1 39 | body stringlengths 1 262k | created_at timestamp[us]date 2000-06-06 02:40:44 2026-06-02 20:16:18 | updated_at timestamp[us]date 2000-06-06 02:40:44 2026-06-02 20:22:56 | reactions unknown | author_association stringclasses 5
values |
|---|---|---|---|---|---|---|---|---|---|
ClickHouse | ClickHouse | 769,657,726 | 19,801 | ucasfl | ```
:) select xy
SELECT xy
Query id: 8906030c-292f-4fc9-b5fa-4bb754908c9c
0 rows in set. Elapsed: 0.003 sec.
Received exception from server (version 21.2.1):
Code: 47. DB::Exception: Received from localhost:9000. DB::Exception: Missing columns: 'xy' while processing query: 'SELECT xy', required columns:... | 2021-01-29T08:28:53 | 2021-01-29T08:29:54 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 769,898,704 | 19,717 | akuzm | 20.10 is already closed. | 2021-01-29T16:12:41 | 2021-01-29T16:12:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 769,899,164 | 19,718 | akuzm | ODBC integration test failures are known. | 2021-01-29T16:13:28 | 2021-01-29T16:13:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 770,076,158 | 19,584 | azat | >Functional stateless tests flaky check (address) — fail: 0, passed: 72
Finally, passed.
The problem was in `system.asynchronous_metrics`, and that fact that they was populated only after `asynchronous_metrics_update_period_s` seconds from start of the server | 2021-01-29T22:07:06 | 2021-01-29T22:07:06 | {} | MEMBER |
ClickHouse | ClickHouse | 770,061,352 | 19,834 | azat | Finally, "fatal: not a git repository: /place/sandbox-data/tasks/0/2/882869720/ClickHouse/.git/modules/contrib/AMQP-CPP" has be fixed, I'm suspecting ancient git version in the sandbox (do not confuse with the docker images, git clone is done in sandbox), should be something older then 2.8, suspected bug is - https://g... | 2021-01-29T21:31:50 | 2021-01-30T06:14:40 | {} | MEMBER |
ClickHouse | ClickHouse | 770,170,114 | 10,128 | vmihailenco | I have a table that already had `TTL time + INTERVAL 1 MONTH DELETE` and `ttl_only_drop_parts = 1` set when table was created.
Table is partitioned slightly unusually `PARTITION BY intDiv(bitShiftRight(id, 33), 86400)`, which in the end partitions data the same as `toStartOfDay(time)` (Snowflake-like ids). I was tol... | 2021-01-30T07:08:53 | 2021-01-30T07:08:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 770,296,406 | 18,617 | alexey-milovidov | Not implemented or Bad arguments.
(with appropriate exception message)
The point is that it's not a logical error in your code. | 2021-01-30T23:25:25 | 2021-01-30T23:25:25 | {} | MEMBER |
ClickHouse | ClickHouse | 770,320,870 | 19,207 | alexey-milovidov | https://clickhouse-test-reports.s3.yandex.net/0/5afcc927ac0ab8d8219505d16d7d591fe6596eaf/stress_test_(debug).html#fail1 | 2021-01-31T03:51:54 | 2021-01-31T03:51:54 | {} | MEMBER |
ClickHouse | ClickHouse | 770,338,729 | 19,859 | alexey-milovidov | > AST fuzzer (UBSan) — ../src/Functions/GatherUtils/Algorithms.h:763:47: runtime error: negation of -92233
#19862 | 2021-01-31T07:13:00 | 2021-01-31T07:13:00 | {} | MEMBER |
ClickHouse | ClickHouse | 769,633,709 | 19,702 | alesapin | ```$ zgrep -a 'Fatal' clickhouse-server.log
2021.01.25 14:56:30.252699 [ 270 ] {} <Fatal> Application: Child process was terminated by signal 11.
```
??? | 2021-01-29T07:32:52 | 2021-01-29T07:32:52 | {} | MEMBER |
ClickHouse | ClickHouse | 769,974,145 | 13,501 | alexey-milovidov | ```
SELECT *
FROM system.functions
WHERE (NOT is_aggregate) AND (name NOT IN
(
SELECT arrayJoin(used_functions) AS f
FROM system.query_log
))
Query id: d851abab-0097-4d14-9de8-1c76692642ec
┌─name─────────────────────┬─is_aggregate─┬─case_insensitive─┬─alias_to─────────────────┐
│ dictGetUUIDOrDef... | 2021-01-29T18:31:04 | 2021-01-29T18:31:04 | {} | MEMBER |
ClickHouse | ClickHouse | 769,978,774 | 19,722 | azat | >Integration tests (release) — fail: 2, passed: 1087, error: 0
>Integration tests (thread) — fail: 2, passed: 1081, error: 0
Failed in upstream too | 2021-01-29T18:40:32 | 2021-01-29T18:40:32 | {} | MEMBER |
ClickHouse | ClickHouse | 770,045,171 | 10,411 | Avogar | > @Avogar Perfect! Maybe add a test?
It's a good idea, I will add it | 2021-01-29T20:55:37 | 2021-01-29T20:55:37 | {} | MEMBER |
ClickHouse | ClickHouse | 770,223,161 | 14,822 | vzakaznikov | @qoega, I've added a couple of tests for this new functionality. Will you be able to review the PR so that it can be merged? | 2021-01-30T14:50:09 | 2021-01-30T14:50:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 769,629,740 | 19,794 | alexey-milovidov | On Yandex.Cloud instance with 8 vCPU it reproduces if I use 36 threads or more (in 2.5 minutes) and does not reproduce if I use 34 threads or less (in 13 minutes). | 2021-01-29T07:22:58 | 2021-01-29T07:39:35 | {} | MEMBER |
ClickHouse | ClickHouse | 769,689,328 | 19,800 | stigsb | Maybe include an integration test for this?
```
diff --git a/tests/integration/test_materialize_mysql_database/materialize_with_ddl.py b/tests/integration/test_materialize_mysql_database/materialize_with_ddl.py
index 66db3fc6c2..105ac2ccbb 100644
--- a/tests/integration/test_materialize_mysql_database/materialize... | 2021-01-29T09:29:20 | 2021-01-29T09:29:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 769,750,525 | 19,709 | kuraga | > 2. ORDER BY (ts, zid1, zid2) -- mostly you should never place high-cardinality column (ts) as leading in the Primary-Key index.
> probably it should be ORDER BY (zid1, zid2, ts)
@den-crane , where could I read about it?
1 - Thanks!
4 - Ahh... Sure.
3 - Thanks, will re-read in documentation. | 2021-01-29T11:30:57 | 2021-01-29T11:30:57 | {} | NONE |
ClickHouse | ClickHouse | 769,903,437 | 19,706 | qoega | Also changes from #19821 to have green test for kerberized kafka | 2021-01-29T16:20:56 | 2021-01-29T16:20:56 | {} | MEMBER |
ClickHouse | ClickHouse | 769,907,885 | 14,822 | vzakaznikov | > Do you plan to return to this PR?
Yes, if someone is willing to review it. I only had to add some unit tests to finish it. | 2021-01-29T16:28:31 | 2021-01-29T16:28:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 770,129,093 | 19,846 | excitoon | It can't be backported without new AWS SDK | 2021-01-30T01:14:38 | 2021-01-30T01:14:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 770,185,716 | 18,151 | Doslin | > I will try to fixes it.
解决了吗 | 2021-01-30T09:48:27 | 2021-01-30T09:48:27 | {} | NONE |
ClickHouse | ClickHouse | 770,210,436 | 18,914 | baibaichen | @danlark1
I believe there is an issue in libcxx which cause build failed in native macos with MacSDKs 10.15, see my comments
on https://github.com/llvm/llvm-project/commit/67dfba96296b37f7bac9b4a68572288bc44b63b2, here are comments:
```
// For C++14, usr/include/malloc/_malloc.h declares aligned_alloc()) ... | 2021-01-30T13:10:24 | 2021-01-30T13:10:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 769,631,649 | 18,247 | ghuname | @SaltTan Can you please check my answer above? | 2021-01-29T07:27:33 | 2021-01-29T07:27:33 | {} | NONE |
ClickHouse | ClickHouse | 769,760,833 | 19,600 | abyss7 | > > @zlx19950903 because `hyperscan` is an optional dependency in our code, which means that you have to conditionally exclude your code if `hyperscan` is not available (i.e. checking the related macro). Your tests and code should compile and run fine in other checks.
>
> If i exclude `hyperscan` from my source code... | 2021-01-29T11:53:22 | 2021-01-29T11:53:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 769,762,448 | 19,797 | filimonov | You can just set up default timezone to UTC system-wide, or in clickhouse configuration. | 2021-01-29T11:57:22 | 2021-01-29T11:57:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 769,979,026 | 19,721 | azat | >Integration tests (release) — fail: 2, passed: 1087, error: 0
>Integration tests (thread) — fail: 2, passed: 1081, error: 0
Failed in upstream too
| 2021-01-29T18:41:05 | 2021-01-29T18:41:05 | {} | MEMBER |
ClickHouse | ClickHouse | 770,010,523 | 19,805 | filimonov | Closes #13613 | 2021-01-29T19:43:56 | 2021-01-29T19:43:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 770,062,110 | 19,824 | azat | >AST fuzzer (TSan) — Lost connection to server. See the logs.
```
2021.01.30 00:08:02.812460 [ 82 ] {} <Fatal> Application: Child process was terminated by signal 9 (KILL). If it is not done by 'forcestop' command or manually, the possible cause is OOM Killer (see 'dmesg' and look at the '/var/log/kern.log' for the... | 2021-01-29T21:33:33 | 2021-01-29T21:33:33 | {} | MEMBER |
ClickHouse | ClickHouse | 770,346,636 | 19,584 | azat | >Please look at the GRPC tests. They are green in master.
978f23e3437d0d0eb64727e87ffe1b80289c9fc2 | 2021-01-31T08:31:26 | 2021-01-31T08:31:26 | {} | MEMBER |
ClickHouse | ClickHouse | 769,629,918 | 19,794 | alexey-milovidov | It always gives latency slightly more than one second. | 2021-01-29T07:23:22 | 2021-01-29T07:23:22 | {} | MEMBER |
ClickHouse | ClickHouse | 769,639,832 | 19,702 | alexey-milovidov | @alesapin See the `main.log`. | 2021-01-29T07:47:57 | 2021-01-29T07:47:57 | {} | MEMBER |
ClickHouse | ClickHouse | 769,640,307 | 19,796 | alexey-milovidov | #19795 | 2021-01-29T07:49:00 | 2021-01-29T07:49:00 | {} | MEMBER |
ClickHouse | ClickHouse | 769,681,721 | 19,229 | filimonov | > https://presentations.clickhouse.tech/dataops_2019/#
Also http://www.brendangregg.com/Slides/Velocity2015_LinuxPerfTools.pdf
> Machines has 4 of these installed.
> Data Width: 64 bits
> Speed: 2933 MT/s
That gives max theoretical RAM throughput: 4 (memory modules) * 64 bits (data width) * 2933000000 (spe... | 2021-01-29T09:15:32 | 2021-01-29T11:54:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 770,005,347 | 19,765 | kssenii | @alexey-milovidov, is this fix still needed? | 2021-01-29T19:33:23 | 2021-01-29T19:33:23 | {} | MEMBER |
ClickHouse | ClickHouse | 770,082,717 | 19,841 | alexey-milovidov | Improvement, not a bugfix - should not backport. | 2021-01-29T22:24:30 | 2021-01-29T22:24:30 | {} | MEMBER |
ClickHouse | ClickHouse | 770,322,084 | 19,207 | alexey-milovidov | Maybe it is our exception from `operator new`:
```
catch (const std::bad_alloc &) {
*db = nullptr;
*comp_error = const_cast<hs_compile_error_t *>(&hs_enomem);
return HS_COMPILER_ERROR;
}
catch (...) {
assert(!"Internal error, unexpected exception");
*db = n... | 2021-01-31T04:07:32 | 2021-01-31T04:07:32 | {} | MEMBER |
ClickHouse | ClickHouse | 769,645,333 | 19,799 | CurtizJ | It's because of `arrayFilter`. Filtering data and offsets was inconsistent.
https://github.com/ClickHouse/ClickHouse/blob/10160e5adf77e0ec0182e0c46a6688780abb5824/src/Functions/array/arrayFilter.cpp#L56-L63. | 2021-01-29T08:00:54 | 2021-01-29T08:01:06 | {} | MEMBER |
ClickHouse | ClickHouse | 769,712,577 | 14,735 | yuuch | > There is currently no planned support for bit and json data types.
I have tried to support the BIT type through the String type locally. Is this feasible? If possible, then I will submit a pull request. @zhang2014 | 2021-01-29T10:12:38 | 2021-01-29T10:12:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 769,782,785 | 10,128 | vmihailenco | >When TTL expression is added to the table the old partitions are not being deleted/moved but user would expect this. Currently MATERIALIZE TTL is required. Since it is consistent with ALTER TABLE behaviour in general, it could be an an extension of ALTER TABLE syntax: ALTER TABLE MODIFY TTL [MATERIALIZE] #11042.
So... | 2021-01-29T12:41:42 | 2021-01-29T12:41:42 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 770,203,447 | 19,580 | alesapin | Current state: investigating issues with races on the connection in NuRaft https://github.com/eBay/NuRaft/issues/169. | 2021-01-30T12:16:42 | 2021-01-30T12:26:06 | {} | MEMBER |
ClickHouse | ClickHouse | 770,316,796 | 19,584 | alexey-milovidov | Please look at the GRPC tests. They are green in master. | 2021-01-31T02:58:14 | 2021-01-31T02:58:14 | {} | MEMBER |
ClickHouse | ClickHouse | 769,673,467 | 19,697 | alesapin | Much more clear logs. Thx! | 2021-01-29T08:59:59 | 2021-01-29T08:59:59 | {} | MEMBER |
ClickHouse | ClickHouse | 770,169,960 | 19,570 | alexey-milovidov | We already doing it for system.tables.
```
milovidov-desktop :) SELECT * FROM system.tables WHERE name = 'hits' \G
SELECT *
FROM system.tables
WHERE name = 'hits'
Query id: c81a283c-5965-434e-aed6-0ed098975100
Row 1:
──────
database: test
name: hits
uuid: ... | 2021-01-30T07:07:07 | 2021-01-30T07:08:18 | {} | MEMBER |
ClickHouse | ClickHouse | 770,270,641 | 19,833 | filimonov | 300 MB instead of 5.6 GB. | 2021-01-30T19:46:27 | 2021-01-30T19:46:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 770,272,914 | 19,691 | kitaisreal | Hello, @chenk818, I checked that we did not have support for `reinterpretAsUInt256`, `reinterpretAsInt128`, `reinterpretAsInt256`. So I added it in scope of https://github.com/ClickHouse/ClickHouse/pull/19858.
When it will be merged, this functions can be used for this task.
```
SELECT reinterpretAsUInt256(hash), ... | 2021-01-30T20:03:41 | 2021-01-30T20:03:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 769,644,196 | 19,795 | alexey-milovidov | Please check if the patch https://github.com/ClickHouse/ClickHouse/pull/19800 helps.
The builds will be available in one hour. | 2021-01-29T07:58:30 | 2021-01-29T07:58:53 | {} | MEMBER |
ClickHouse | ClickHouse | 770,039,751 | 19,750 | akuzm | The test failures are not related to the changes. | 2021-01-29T20:44:05 | 2021-01-29T20:44:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 770,081,306 | 18,247 | SaltTan | > SELECT count()
> FROM system.zookeeper
> WHERE path = '/clickhouse/task_queue/ddl/'
>
> Node 1 and 2 (shard1) 124, node 3 and 4 (shard2) 1
I don't understand this.
I would expect all nodes to look at the same znode and see the same number of the tasks.
Can you see your DROP TABLE in the list? | 2021-01-29T22:20:22 | 2021-01-29T22:20:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 770,082,705 | 19,843 | alexey-milovidov | Improvement, not a bugfix - should not backport. | 2021-01-29T22:24:27 | 2021-01-29T22:24:27 | {} | MEMBER |
ClickHouse | ClickHouse | 770,321,477 | 19,766 | ghost | I added documentation. | 2021-01-31T04:00:19 | 2021-01-31T04:00:36 | {} | NONE |
ClickHouse | ClickHouse | 769,623,016 | 19,794 | alexey-milovidov | It because of BackgroundSchedulePool is using Poco::NotificationQueue that TCPServer also using. | 2021-01-29T07:07:43 | 2021-01-29T07:07:43 | {} | MEMBER |
ClickHouse | ClickHouse | 769,665,937 | 19,701 | kitaisreal | Fixed in https://github.com/ClickHouse/ClickHouse/pull/19767. | 2021-01-29T08:45:44 | 2021-01-29T08:45:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 769,973,368 | 13,501 | alexey-milovidov | Superseded by #19371. | 2021-01-29T18:29:24 | 2021-01-29T18:29:24 | {} | MEMBER |
ClickHouse | ClickHouse | 770,130,372 | 18,662 | awesomeleo | > > > MaterializeMySQL underlying engine is ReplacingMergeTree, this case looks not make sense. Can you insert data to ReplacingMergeTree directly and do the test?
> >
> >
> > @BohuTANG Yes. I have done the test. It's an issue with ReplacingMergeTree.
> > ## MaterializeMySQL with final:
> > localhost :) select c... | 2021-01-30T01:22:04 | 2021-01-30T01:24:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 770,295,479 | 18,617 | vzakaznikov | @alexey-milovidov, so what should be the correct error code?
Materialized view returns the same error. So currently LVs and MVs do not support table functions in the inner most query.
@KochetovNicolai, do you think it is possible to support table functions in MVs and LVs?
```
user@user-host:~$ clickhouse-client... | 2021-01-30T23:17:19 | 2021-01-30T23:17:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 770,322,751 | 19,207 | alexey-milovidov | Yes, it is. | 2021-01-31T04:15:32 | 2021-01-31T04:15:32 | {} | MEMBER |
ClickHouse | ClickHouse | 769,790,705 | 19,790 | alexey-milovidov | @tavplubix After some recent changes client also started to show incorrect messages about "Unmatched parenthesis".
Could you please check what changes broke it and add a test with `expect`? | 2021-01-29T12:59:36 | 2021-01-29T12:59:36 | {} | MEMBER |
ClickHouse | ClickHouse | 769,791,746 | 19,797 | lessenko | @filimonov,
Thank you. It can be closed. | 2021-01-29T13:01:41 | 2021-01-29T13:01:41 | {} | NONE |
ClickHouse | ClickHouse | 769,897,805 | 19,423 | akuzm | ODBC integration test failures are known, ANTLR test should be disabled by now. | 2021-01-29T16:11:11 | 2021-01-29T16:11:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 770,148,082 | 19,745 | sliontc | Did you guys mean it's already available or future plan? | 2021-01-30T03:24:42 | 2021-01-30T03:24:42 | {
"+1": 3
} | NONE |
ClickHouse | ClickHouse | 770,156,866 | 19,850 | keen-wolf | l'd like to take this one～ | 2021-01-30T04:47:00 | 2021-01-30T04:47:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 769,787,431 | 19,702 | alesapin | ```
2021.01.29 14:13:25.486254 [ 186 ] {} <Fatal> : Logical error: 'Missing part 7_117_117_0 is not in our queue.'.
2021.01.29 14:13:25.487441 [ 70975 ] {} <Fatal> BaseDaemon: ########################################
2021.01.29 14:13:25.487592 [ 70975 ] {} <Fatal> BaseDaemon: (version 21.2.1.5845, build id: DC45960B... | 2021-01-29T12:52:26 | 2021-01-29T12:52:26 | {} | MEMBER |
ClickHouse | ClickHouse | 769,856,133 | 19,702 | alesapin | Happened after the table drop. | 2021-01-29T15:00:26 | 2021-01-29T15:00:26 | {} | MEMBER |
ClickHouse | ClickHouse | 770,065,835 | 19,375 | Avogar | > If I understand correctly, then `length(thread_ids) > 1` is not correct, because there will be threads at least for:
>
> * TCPHandler
> * PipelineExecutor
> * parts loading - https://github.com/ClickHouse/ClickHouse/blob/f996b8433499cbbabc32febf55ce2291afbc2290/src/Storages/MergeTree/MergeTreeDataSelectExecutor.... | 2021-01-29T21:42:09 | 2021-01-29T21:42:09 | {} | MEMBER |
ClickHouse | ClickHouse | 770,084,161 | 10,411 | Avogar | Added: https://github.com/ClickHouse/ClickHouse/pull/19851 | 2021-01-29T22:29:00 | 2021-01-29T22:29:00 | {} | MEMBER |
ClickHouse | ClickHouse | 770,144,806 | 19,743 | mdianjun | I'm confused why `AST fuzzer (UBSan)` failed. @akuzm Could you give me a few more details, please? | 2021-01-30T02:58:50 | 2021-01-30T02:58:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 770,216,503 | 17,623 | baibaichen | > Any hope for improved Parquet performance?
@alanpaulkwan @filimonov @alexey-milovidov
I did compare the read performance between merge tree and parquet, but I don't know why Parquet is so slow, what's wrong with Parquet? Any more technical info?
| 2021-01-30T13:58:35 | 2021-01-30T13:58:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 769,880,766 | 19,702 | alesapin | Part checker is ideologically broken. It doesn't have any synchronization with queue or storage, the error happens in the following situation:
```
2021.01.29 14:13:24.280226 [ 13135 ] {} <Information> test_kstf0r.alter_table2 (676e6206-9f17-45f2-8083-8aef54f29f8e): DB::Exception: No active replica has part 7_117_117_... | 2021-01-29T15:42:31 | 2021-01-29T15:42:31 | {} | MEMBER |
ClickHouse | ClickHouse | 770,004,904 | 10,128 | filimonov | > I guess it is because of this improvement. I don't find it intuitive or user friendly. Old behavior was at least harmless.
Materialize ttl should not be heavy, and should not take lot of disk space - it just initialize the metadata required for ttl execution. Need more details of what was happening.
| 2021-01-29T19:32:27 | 2021-01-29T19:32:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 770,011,500 | 19,805 | filimonov | Also #13059 and maybe more. | 2021-01-29T19:45:57 | 2021-01-29T19:45:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 770,042,286 | 12,087 | akuzm | Fixed here https://github.com/ClickHouse/ClickHouse/pull/19681 | 2021-01-29T20:49:30 | 2021-01-29T20:49:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 770,082,692 | 19,844 | alexey-milovidov | Improvement, not a bugfix - should not backport. | 2021-01-29T22:24:25 | 2021-01-29T22:24:25 | {} | MEMBER |
ClickHouse | ClickHouse | 770,275,048 | 19,518 | bharatnc | @vdimir Pushed a minor fix for: [clang-11-darwin build](https://clickhouse-builds.s3.yandex.net/19518/5cb99ae699863ea6f896ea16770e7cdeb33921d4/clickhouse_special_build_check/build_log_883465367_1612034896.txt) failure.
It looks like `Offset` is `UInt64` and `IPV4_MAX_TEXT_LENGTH` is `size_t` so there was a failure w... | 2021-01-30T20:20:24 | 2021-01-30T20:25:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 770,318,813 | 17,623 | alexey-milovidov | @baibaichen Could you please post your scenario in separate issue?
There are many details:
- usage of indexes;
- data compression (Parquet does it but is it enabled?);
- ability to read subset of columns (Parquet does it but do we utilize it correctly?);
- possible extra copying;
- inefficient data conversions; | 2021-01-31T03:22:07 | 2021-01-31T03:22:44 | {} | MEMBER |
ClickHouse | ClickHouse | 770,323,684 | 16,724 | alexey-milovidov | @kitaisreal This is somehow related: https://clickhouse-test-reports.s3.yandex.net/19677/3b6395fbb774c132273083f4f84a4f5a34c12866/fuzzer_debug/server.log | 2021-01-31T04:27:56 | 2021-01-31T04:27:56 | {} | MEMBER |
ClickHouse | ClickHouse | 770,324,456 | 19,855 | alexey-milovidov | @amosbird Could you please take a look? | 2021-01-31T04:37:13 | 2021-01-31T04:37:13 | {} | MEMBER |
ClickHouse | ClickHouse | 769,975,917 | 19,799 | alexey-milovidov | Wow, it's passed! | 2021-01-29T18:34:36 | 2021-01-29T18:34:36 | {} | MEMBER |
ClickHouse | ClickHouse | 770,038,781 | 19,793 | akuzm | The test failures are not related to the changes. | 2021-01-29T20:42:02 | 2021-01-29T20:42:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 770,343,846 | 19,860 | alexey-milovidov | > AST fuzzer (UBSan) — ../src/Functions/FunctionDateOrDateTimeAddInterval.h:73:18: runtime error: signed integer overflo
#19859 | 2021-01-31T08:05:28 | 2021-01-31T08:05:28 | {} | MEMBER |
ClickHouse | ClickHouse | 770,344,305 | 19,773 | alexey-milovidov | Fuzzer reveal all other errors in Bloom Filter index: https://github.com/ClickHouse/ClickHouse/issues/19757 | 2021-01-31T08:09:36 | 2021-01-31T08:09:36 | {} | MEMBER |
ClickHouse | ClickHouse | 769,626,047 | 19,789 | hombit | Thank you, looks good for me! | 2021-01-29T07:14:58 | 2021-01-29T07:14:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 769,639,305 | 19,799 | alexey-milovidov | Filtering by values not equal to 0 and 1 was implementation specific.
I don't understand how it is related to ORDER BY multiple columns with Array column. | 2021-01-29T07:46:29 | 2021-01-29T07:46:29 | {} | MEMBER |
ClickHouse | ClickHouse | 769,646,105 | 19,799 | alexey-milovidov | Ok, I see. Let's wait for perf tests - hope it will not introduce performance regressions. Now the logic is better.
Otherwise we can make it consistent without changing the logic of filtering. | 2021-01-29T08:02:38 | 2021-01-29T08:02:38 | {} | MEMBER |
ClickHouse | ClickHouse | 769,659,063 | 19,795 | stigsb | I can confirm that #19800 fixes replication of Unicode >= 0x10000 for MaterializeMySQL databases.
| 2021-01-29T08:31:42 | 2021-01-29T08:31:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 770,156,405 | 15,552 | alexey-milovidov | This is a good alternative: https://monitor.shodan.io/ | 2021-01-30T04:41:27 | 2021-01-30T04:41:27 | {} | MEMBER |
ClickHouse | ClickHouse | 770,063,216 | 19,824 | azat | >AST fuzzer (UBSan) — ../src/Functions/bitAnd.cpp:24:61: runtime error: nan is outside the range of representable values of type 'long'
Looks like `bitAnd` (and similar) should be prohibited for float/double - #19853 | 2021-01-29T21:35:55 | 2021-01-30T05:02:21 | {} | MEMBER |
ClickHouse | ClickHouse | 770,185,334 | 19,702 | alesapin | Finally fixed. | 2021-01-30T09:44:28 | 2021-01-30T09:44:28 | {} | MEMBER |
ClickHouse | ClickHouse | 770,212,741 | 17,623 | baibaichen | Support Azure Blob see https://github.com/ClickHouse/ClickHouse/issues/19307 | 2021-01-30T13:29:02 | 2021-01-30T13:29:02 | {
"+1": 3
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 770,321,920 | 19,207 | alexey-milovidov | ```
clickhouse-server: ../contrib/hyperscan/src/hs.cpp:277: hs_error_t ue2::hs_compile_multi_int(const char *const *, const unsigned int *, const unsigned int *, const hs_expr_ext *const *, unsigned int, unsigned int, const hs_platform_info_t *, hs_database_t **, hs_compile_error_t **, const ue2::Grey &): Assertion `!... | 2021-01-31T04:05:18 | 2021-01-31T04:05:18 | {} | MEMBER |
ClickHouse | ClickHouse | 770,323,761 | 19,856 | alexey-milovidov | Missing optimization (not a bug). | 2021-01-31T04:29:27 | 2021-01-31T04:29:27 | {} | MEMBER |
ClickHouse | ClickHouse | 769,632,380 | 19,748 | excitoon | Tested manually. | 2021-01-29T07:29:16 | 2021-01-29T07:29:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 769,762,871 | 19,800 | alexey-milovidov | Yes, I keep it as draft before integration test is written.
I will include your test! | 2021-01-29T11:58:25 | 2021-01-29T11:58:25 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 770,011,779 | 10,411 | Avogar | @alexey-milovidov @qoega It seems such behaviour already works after https://github.com/ClickHouse/ClickHouse/issues/10683
```
:) CREATE TABLE collaps (id Int32, name String, sign Int8) ENGINE = CollapsingMergeTree(sign) ORDER BY (id);
CREATE TABLE collaps
(
`id` Int32,
`name` String,
`sign` Int8
... | 2021-01-29T19:46:34 | 2021-01-29T19:47:15 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 770,033,739 | 10,411 | alexey-milovidov | @Avogar Perfect! Maybe add a test? | 2021-01-29T20:31:20 | 2021-01-29T20:31:20 | {} | MEMBER |
ClickHouse | ClickHouse | 770,044,816 | 19,749 | den-crane | Do not really understand what is the problem
Just do not use MV? Insert into both tables from a client ? | 2021-01-29T20:54:52 | 2021-01-29T20:54:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 770,271,577 | 19,833 | qoega | I think we can add something like size check for a directory after a test. It is not OK if it is too large.
I already wanted to check no Fatal messages in logs after test teardown | 2021-01-30T19:53:54 | 2021-01-30T19:53:54 | {} | MEMBER |
ClickHouse | ClickHouse | 770,324,490 | 19,252 | alexey-milovidov | This should be fixed: #19855 | 2021-01-31T04:37:45 | 2021-01-31T04:37:45 | {} | MEMBER |
ClickHouse | ClickHouse | 759,354,885 | 15,889 | KochetovNicolai | All tests seem to be working on current master | 2021-01-13T10:25:06 | 2021-01-13T10:25:06 | {} | MEMBER |
ClickHouse | ClickHouse | 759,398,794 | 18,791 | Avogar | LGTM, just one small comment about an exception message. | 2021-01-13T11:50:38 | 2021-01-13T11:50:38 | {} | MEMBER |
ClickHouse | ClickHouse | 759,410,696 | 18,971 | KochetovNicolai | Well, it is documented: `write operations block reads and each other`.
I think we just need to ignore `parallel_distributed_insert_select` for storage Log. However, it would be probably possible to deadlock using several MATERIALIZED VIEW with Log engine. | 2021-01-13T12:15:52 | 2021-01-13T12:15:52 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.