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 | 635,946,088 | 11,287 | taythebot | The server's timezone is `Etc/UTC`. Not sure about the client, I'm using DBeaver and the driver property for `use_time_zone` is empty.
Should I be setting the driver timezone to UTC as well?
Edit: The driver is using the server timezone | 2020-05-29T12:29:49 | 2020-05-29T12:31:38 | {} | NONE |
ClickHouse | ClickHouse | 635,957,638 | 11,277 | qoega | can you reproduce problem with newer version? 19.9 is not supported for a while. You can check 19.14 LTS, 20.3 LTS or 20.1, 20.4 versions. | 2020-05-29T12:56:40 | 2020-05-29T12:56:40 | {} | MEMBER |
ClickHouse | ClickHouse | 635,975,390 | 11,287 | taythebot | Yes, it seems like the issue was with DBeaver as the time is displayed correctly with Tabix and the playground.
Thank you so much for the quick reply. I'll be closing this issue. | 2020-05-29T13:34:17 | 2020-05-29T13:34:17 | {} | NONE |
ClickHouse | ClickHouse | 636,366,037 | 11,208 | alexey-milovidov | Performance test did not finish. | 2020-05-30T18:15:40 | 2020-05-30T18:15:40 | {} | MEMBER |
ClickHouse | ClickHouse | 636,201,847 | 11,254 | devwout | The behavior described is also applicable to `20.3.10.75` and older versions. The blank output in your comment is not what I would expect. As described in the ticket, I would expect `'"{"'`.
```
$ docker run -i yandex/clickhouse-server:20.3.10.75 clickhouse-local --format JSONEachRow -q "select visitParamExtractRaw... | 2020-05-29T21:33:26 | 2020-05-30T20:42:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 635,943,141 | 11,287 | champtar | @defusevenue what is the timezone on your server and client ? | 2020-05-29T12:22:35 | 2020-05-29T12:22:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 636,110,382 | 11,282 | den-crane | >maxState(minute_flow) as daily_flow,
> sum(flow) as minute_flow
it could not work because MV does not read the source table. It reads only data from INSERTs, moreover it reads data in blocks.
For example Insert broke into two blocks
First block: stattime = 10:00:00 flow = 1
Second block: stattime = 10:00:... | 2020-05-29T18:07:19 | 2020-05-29T18:08:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 636,191,004 | 11,302 | alexey-milovidov | @alesapin Could you please "validate" these Dockerfiles? | 2020-05-29T21:02:56 | 2020-05-29T21:02:56 | {} | MEMBER |
ClickHouse | ClickHouse | 636,338,360 | 11,303 | alexey-milovidov | Build check:
> Build failed without build log. It's inner CI problem. Maybe something wrong with docker or host. Just wait for build restart.
Ok. | 2020-05-30T14:27:24 | 2020-05-30T14:27:24 | {} | MEMBER |
ClickHouse | ClickHouse | 636,339,091 | 11,295 | alexey-milovidov | > Performance Pending — Started
The test did not start in 20 hours. Ignore. | 2020-05-30T14:33:13 | 2020-05-30T14:33:13 | {} | MEMBER |
ClickHouse | ClickHouse | 618,393,466 | 10,398 | 4ertus2 | It looks like we have wrong conversion from nullable to not nullable in case when EngineJoin created with `SETTINGS join_use_nulls = 1` and query has no JOIN, but use engine itself. | 2020-04-23T13:09:01 | 2020-05-30T19:11:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 636,474,907 | 11,314 | alexey-milovidov | > test_ttl_move
- flacky test, reported. | 2020-05-31T13:56:34 | 2020-05-31T13:56:34 | {} | MEMBER |
ClickHouse | ClickHouse | 635,772,141 | 10,936 | slimtom95 | closing. migrate to alter detach/attach | 2020-05-29T05:44:08 | 2020-05-29T05:44:08 | {} | NONE |
ClickHouse | ClickHouse | 635,909,618 | 11,166 | alesapin | Actually this conversion is safe if you don't have NULLs in your Nullable column. But we cannot understand this fact without attempt to execute alter. | 2020-05-29T10:55:17 | 2020-05-29T10:55:17 | {} | MEMBER |
ClickHouse | ClickHouse | 635,981,877 | 11,215 | qoega | Have you tried actual versions of ClickHouse? 20.3, 20.4. | 2020-05-29T13:47:49 | 2020-05-29T13:47:49 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 636,246,099 | 11,282 | jencoldeng | > > maxState(minute_flow) as daily_flow,
> > sum(flow) as minute_flow
>
> it could not work because MV does not read the source table. It reads only data from INSERTs, moreover it reads data in blocks.
> For example Insert broke into two blocks
> First block: stattime = 10:00:00 flow = 1
> Second block: stattime... | 2020-05-30T00:25:20 | 2020-05-30T00:25:20 | {} | NONE |
ClickHouse | ClickHouse | 636,333,241 | 7,057 | jason-godden | Would be great to see this in. Was trying to use tokenbf indices today and they work very well and can pinpoint a handful of rows in a 3 billion row table in under 300ms however the upfront delay on a larger dataset of 30 billion rows for the index scan ends up being the bulk of the runtime at around 4 seconds. | 2020-05-30T13:46:36 | 2020-05-30T13:46:36 | {} | NONE |
ClickHouse | ClickHouse | 635,735,865 | 8,505 | jackpgao | - The most recent version still have the problem.
```sql
xxx :) set max_execution_time = 1 ;
SET max_execution_time = 1
Ok.
0 rows in set. Elapsed: 0.001 sec.
xxx :)
xxx :) select sleep(3);
SELECT sleep(3)
┌─sleep(3)─┐
│ 0 │
└──────────┘
1 rows in set. Elapsed: 3.001 sec.
xxx :) ... | 2020-05-29T03:26:17 | 2020-05-29T03:26:17 | {} | NONE |
ClickHouse | ClickHouse | 635,863,866 | 11,283 | filimonov | 2 concerns:
1) optionate filling of those columns?
2) use `Array(LowCardinality(String))` for header names? | 2020-05-29T09:08:12 | 2020-05-29T09:08:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 636,183,065 | 6,228 | alexey-milovidov | For merge tables as well:
```
SELECT count() FROM merge(test, '^hits$') SAMPLE 1/2
``` | 2020-05-29T20:41:58 | 2020-05-29T20:41:58 | {} | MEMBER |
ClickHouse | ClickHouse | 636,457,214 | 11,296 | alexey-milovidov | > Stress test (thread) — Sanitizer assert (in stderr.log)
Actually it's OOM on Sandbox server (misleading message). | 2020-05-31T11:18:25 | 2020-05-31T11:18:25 | {} | MEMBER |
ClickHouse | ClickHouse | 635,877,243 | 10,884 | brefre | Hi,
Thank you for your patience. I did read your message.
I grepped the logs on all my brokers, and I can see no evidence for the __consumer_offsets topic getting out of sync.
This is the method I used:
`grep -i isr /var/log/kafka/console.log* | grep __consumer_offsets | grep 2020 `
I did find evidence tha... | 2020-05-29T09:36:18 | 2020-05-29T09:36:18 | {} | NONE |
ClickHouse | ClickHouse | 635,961,718 | 11,279 | filimonov | The check is intended to prevent accidental loss of some fragments of the table during reconfiguration.
The main question here is why the part for the table `default.btrace` appeared on the disk which is not listed in `moving_from_ssd_to_hdd` storage policy.
Normally that should not happen and may be caused by ... | 2020-05-29T13:05:57 | 2020-05-29T13:05:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 635,966,591 | 11,245 | qoega | Can reproduce on master as well. | 2020-05-29T13:16:14 | 2020-05-29T13:16:14 | {} | MEMBER |
ClickHouse | ClickHouse | 635,969,472 | 11,226 | And42 | There is a number of changes I want to suggest but it's a bit exhausting for me to go through all my scenarios on each commit. If that's needed though I can try to find anything I can when the pull request is ready | 2020-05-29T13:22:11 | 2020-05-29T13:22:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 636,165,903 | 6,417 | alexey-milovidov | ```
0%Received
```
I will also fix bad formatting of output in clickhouse-client. | 2020-05-29T20:01:27 | 2020-05-29T20:01:27 | {} | MEMBER |
ClickHouse | ClickHouse | 636,172,819 | 5,732 | alexey-milovidov | We can simply check execution speed limits on leaf nodes as well. That will solve the issue. | 2020-05-29T20:16:41 | 2020-05-29T20:16:41 | {} | MEMBER |
ClickHouse | ClickHouse | 636,362,765 | 11,291 | KochetovNicolai | Well, performance tests are not fine.
We need to investigate what is going wrong. | 2020-05-30T17:46:47 | 2020-05-30T17:47:48 | {} | MEMBER |
ClickHouse | ClickHouse | 635,943,379 | 11,253 | traceon | I used to use GNU Make in Ubuntu some time ago, and it worked fine. | 2020-05-29T12:23:12 | 2020-05-29T12:23:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 635,985,201 | 11,226 | akuzm | > There is a number of changes I want to suggest but it's a bit exhausting for me to go through all my scenarios on each commit. If that's needed though I can try to find anything I can when the pull request is ready
Sure, I can ping you for review when I'm finished (today or early next week). I still have to fix so... | 2020-05-29T13:54:30 | 2020-05-29T13:54:30 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 636,181,674 | 5,732 | alexey-milovidov | Implemented in https://github.com/ClickHouse/ClickHouse/pull/11296 | 2020-05-29T20:38:21 | 2020-05-29T20:38:21 | {} | MEMBER |
ClickHouse | ClickHouse | 636,359,492 | 11,312 | alexey-milovidov | @alesapin Is there any simpler way to achieve what I want? | 2020-05-30T17:16:27 | 2020-05-30T17:16:27 | {} | MEMBER |
ClickHouse | ClickHouse | 636,367,190 | 11,302 | alexey-milovidov | > Integration tests (release) — failed:0, passed:646, error:1
#11312
> Functional stateless tests (memory) — fail: 1, passed: 1869, skipped: 8
Added to CI fest. | 2020-05-30T18:25:47 | 2020-05-30T18:25:47 | {} | MEMBER |
ClickHouse | ClickHouse | 636,371,483 | 11,174 | alexey-milovidov | @excitoon This PR does not look ready. | 2020-05-30T19:04:08 | 2020-05-30T19:04:08 | {} | MEMBER |
ClickHouse | ClickHouse | 635,720,073 | 11,276 | alexey-milovidov | ```
SET max_block_size = 1;
INSERT INTO insert_small_block_performance (x) SELECT toString(number) FROM numbers(1000000);
```
master: 484.56 rows/s.
new: 180.03 thousand rows/s. | 2020-05-29T02:27:49 | 2020-05-29T02:27:49 | {} | MEMBER |
ClickHouse | ClickHouse | 635,997,512 | 11,208 | alexey-milovidov | > Stress test (thread) — Sanitizer assert (in stderr.log)
Already fixed by @KochetovNicolai
> Performance Pending — Started
Sandbox quota overload, but I cannot wait. | 2020-05-29T14:18:55 | 2020-05-29T14:18:55 | {} | MEMBER |
ClickHouse | ClickHouse | 636,145,043 | 11,191 | alexey-milovidov | @UnamedRus You are right, we need some modification to this scheme. | 2020-05-29T19:24:01 | 2020-05-29T19:24:01 | {} | MEMBER |
ClickHouse | ClickHouse | 636,168,541 | 6,417 | alexey-milovidov | > But I can still reproduce the same effect if I will set it to very low value
And this behaviour is 100% legitimate. | 2020-05-29T20:05:19 | 2020-05-29T20:05:19 | {} | MEMBER |
ClickHouse | ClickHouse | 636,367,299 | 11,312 | alexey-milovidov | > CalledProcessError: Command 'docker -H unix:///place/sandbox-data/tasks/0/0/688347300/docker_root/docker.sock run -i --rm ubuntu:14.04 ping6 github.com -c4' returned non-zero exit status 2.
@alesapin | 2020-05-30T18:26:47 | 2020-05-30T18:26:47 | {} | MEMBER |
ClickHouse | ClickHouse | 635,792,376 | 1,997 | jimmy17x | > Try to run clickhouse-client with `--multiline` option.
can you share how to read more about --multiline ? | 2020-05-29T06:44:47 | 2020-05-29T06:44:47 | {} | NONE |
ClickHouse | ClickHouse | 635,970,460 | 11,278 | alexey-milovidov | > Functional stateless tests (release) — fail: 21, passed: 1853
Already fixed by @alesapin
> Stress test (thread) — Sanitizer assert (in stderr.log)
Already fixed by @KochetovNicolai
> Yandex synchronization check (only for Yandex employees)
> Yandex third-party checks (only for Yandex employees)
... | 2020-05-29T13:24:09 | 2020-05-29T13:24:09 | {} | MEMBER |
ClickHouse | ClickHouse | 635,982,534 | 11,228 | pic-Nick | If dictGetUInt32 replaced with dictGet, error message changes to
`Code: 42, e.displayText() = DB::Exception: Function dictGet... takes 3 or 4 arguments (version 20.4.4.18 (official build))` | 2020-05-29T13:49:09 | 2020-05-29T13:49:09 | {} | NONE |
ClickHouse | ClickHouse | 636,142,588 | 11,127 | 4ertus2 | Related #10481 | 2020-05-29T19:18:01 | 2020-05-29T19:18:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 636,200,410 | 11,261 | alexey-milovidov | @qoega You did not reported test failure:
`00992_system_parts_race_condition_zookeeper` | 2020-05-29T21:29:03 | 2020-05-29T21:29:03 | {} | MEMBER |
ClickHouse | ClickHouse | 636,378,877 | 9,473 | alexey-milovidov | @akuzm We can disconnect from the server, and return to REPL if Ctrl+C was pressed during query execution more than one time. | 2020-05-30T20:11:36 | 2020-05-30T20:11:36 | {} | MEMBER |
ClickHouse | ClickHouse | 636,381,460 | 11,313 | alexey-milovidov | Ok 26100b6. | 2020-05-30T20:30:25 | 2020-05-30T20:30:25 | {} | MEMBER |
ClickHouse | ClickHouse | 636,383,192 | 11,316 | alexey-milovidov | It looks like this:
```
2020.05.30 22:52:19.113685 [ 21276 ] {} <Error> void DB::BackgroundProcessingPool::threadFunction(): Code: 173, e.displayText() = DB::ErrnoException: Allocator: Cannot malloc 1.00 MiB., errno: 12, strerror: Cannot allocate memory, Stack trace (when copying this message, always include the li... | 2020-05-30T20:44:57 | 2020-05-30T20:44:57 | {} | MEMBER |
ClickHouse | ClickHouse | 635,941,346 | 11,281 | qoega | This case is covered in changelog for 20.4 release
https://clickhouse.tech/docs/en/whats-new/changelog/#clickhouse-release-v20-4-2-9-2020-05-12 | 2020-05-29T12:18:04 | 2020-05-29T12:18:29 | {} | MEMBER |
ClickHouse | ClickHouse | 635,955,918 | 11,286 | alexey-milovidov | Broken after https://github.com/ClickHouse/ClickHouse/pull/5965
It is caused by wrong index analysis for empty/notEmpty and Array data type.
By the way, the query works if you write
```
select count() from count_lc_test where arr != []
``` | 2020-05-29T12:52:40 | 2020-05-29T12:53:42 | {} | MEMBER |
ClickHouse | ClickHouse | 635,968,447 | 7,091 | tavplubix | @jackpgao, it is a known issue. Query can be killed or cancelled only in some points between processing of blocks in pipeline, where query checks elapsed time and some flags to determine if it should stop the execution. Query cannot check it when sleeping on row processing. | 2020-05-29T13:20:02 | 2020-05-29T13:20:02 | {} | MEMBER |
ClickHouse | ClickHouse | 635,997,381 | 11,285 | alesapin | Only functional tests are important in this PR. | 2020-05-29T14:18:39 | 2020-05-29T14:18:39 | {} | MEMBER |
ClickHouse | ClickHouse | 624,613,111 | 10,697 | CurtizJ | TODO:
- [x] Support partition commands
- [ ] Tune MergeSelector's algorithm for in-memory parts
- [ ] Add memory limit for those parts in total | 2020-05-06T12:13:31 | 2020-05-29T17:45:47 | {} | MEMBER |
ClickHouse | ClickHouse | 636,182,725 | 6,228 | alexey-milovidov | I checked that it works perfectly:
```
SET max_execution_speed = 1000000, timeout_before_checking_execution_speed = 0.001;
SELECT count()
FROM test.hits
SAMPLE 1 / 2
┌─count()─┐
│ 4392010 │
└─────────┘
1 rows in set. Elapsed: 8.234 sec. Processed 8.23 million rows, 138.90 MB (999.14 thousand rows/s., 1... | 2020-05-29T20:41:05 | 2020-05-29T20:41:29 | {} | MEMBER |
ClickHouse | ClickHouse | 636,287,026 | 11,201 | blinkov | @jackpgao could you publish that configuration? | 2020-05-30T06:39:01 | 2020-05-30T06:39:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 636,372,392 | 10,398 | alexey-milovidov | @4ertus2 It is reproducing in master:
```
milovidov-desktop :) select * from table_with_join_engine
SELECT *
FROM table_with_join_engine
→ Progress: 0.00 rows, 0.00 B (0.00 rows/s., 0.00 B/s.)
Exception on client:
Code: 271. DB::Exception: Data compressed with different methods, given method byte 0x69, pr... | 2020-05-30T19:13:02 | 2020-05-30T19:13:02 | {} | MEMBER |
ClickHouse | ClickHouse | 636,378,411 | 9,473 | alexey-milovidov | @filimonov
> ability to stop printing the big output to console when Ctrl+C is pressed
Usually the issue is in the terminal side.
clickhouse-client already sent it's output to the terminal but it takes long time to display it. | 2020-05-30T20:08:33 | 2020-05-30T20:26:55 | {} | MEMBER |
ClickHouse | ClickHouse | 635,839,539 | 11,262 | alesapin | Unit tests passed. | 2020-05-29T08:16:08 | 2020-05-29T08:16:08 | {} | MEMBER |
ClickHouse | ClickHouse | 635,917,214 | 4,631 | filimonov | Obsolete. If you see some of the problems discussed here on new versions please open new issue. | 2020-05-29T11:15:08 | 2020-05-29T11:15:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 635,942,583 | 11,286 | alexey-milovidov | Reproduced in master. | 2020-05-29T12:21:12 | 2020-05-29T12:21:12 | {} | MEMBER |
ClickHouse | ClickHouse | 635,955,717 | 11,287 | qoega | What is the type of `date` column? You can define its timezone explicilty.
For debug purposes you can format values on server as strings to omit driver with formatDateTime as in ClickHouse Playground example https://play.clickhouse.tech/?file=datetime
Is your question answered? | 2020-05-29T12:52:12 | 2020-05-29T12:52:12 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 635,961,230 | 11,254 | qoega | 20.3 has no problem.
Also not reproducing in 20.5/master
```
│ VERSION_GITHASH │ 91df18a906dcffdbee6816e5389df6c65f86e35f
```
``` sql
SELECT visitParamExtractRaw('{"k":"{"}', 'k')
┌─visitParamExtrac... | 2020-05-29T13:04:54 | 2020-05-29T13:04:54 | {} | MEMBER |
ClickHouse | ClickHouse | 635,983,040 | 10,973 | qoega | You can check if it works with the query from the issue description - https://github.com/ClickHouse/ClickHouse/issues/11215
It is real use case, when we want to use this tool. | 2020-05-29T13:50:15 | 2020-05-29T13:50:15 | {} | MEMBER |
ClickHouse | ClickHouse | 636,082,116 | 11,276 | alexey-milovidov | > Stress test (thread) — Sanitizer assert (in stderr.log)
Already fixed by @KochetovNicolai
> Yandex synchronization check (only for Yandex employees)
> Yandex third-party checks (only for Yandex employees)
Broken by Yandex DevTools.
> Performance Pending — Started
Very long, cannot wait. | 2020-05-29T17:04:49 | 2020-05-29T17:04:49 | {} | MEMBER |
ClickHouse | ClickHouse | 636,171,527 | 11,297 | alexey-milovidov | It works only in presense of `timeout_before_checking_execution_speed`.
Fixed in #11296 | 2020-05-29T20:13:16 | 2020-05-29T20:13:16 | {} | MEMBER |
ClickHouse | ClickHouse | 636,338,653 | 11,304 | alexey-milovidov | > Performance Pending — Started
Did not finish in 17 hour, probably not enough quota in Sandbox. Ignore. | 2020-05-30T14:29:38 | 2020-05-30T14:29:38 | {} | MEMBER |
ClickHouse | ClickHouse | 636,410,202 | 8,505 | jackpgao | > @jackpgao it works as expected. Timeout check happens in between blocks. You need more blocks
>
> ```
> set max_execution_time = 1 ;
> select sleep(1) from numbers(10) settings max_block_size=1;
>
> Code: 159. DB::Exception: Received from localhost:9000. DB::Exception: Timeout exceeded: elapsed 1.999990824 se... | 2020-05-31T02:13:50 | 2020-05-31T02:13:50 | {} | NONE |
ClickHouse | ClickHouse | 635,904,526 | 11,249 | KochetovNicolai | Stress should be fixed in #11264 | 2020-05-29T10:42:03 | 2020-05-29T10:42:03 | {} | MEMBER |
ClickHouse | ClickHouse | 635,726,318 | 8,889 | jackpgao | > > and the version 20.4.3.16 seems works well.
>
> So, should I close this issue as fixed?
`20.4.3.16` has more critical bug, pls: https://github.com/ClickHouse/ClickHouse/issues/11153 | 2020-05-29T02:50:55 | 2020-05-29T02:50:55 | {} | NONE |
ClickHouse | ClickHouse | 635,731,877 | 11,235 | hczhcz | It is probably good to name it as `-Sort` considering we are using different terms in SQL functions and statements (sort vs order by, if vs where/having, uniq vs distinct, ...) | 2020-05-29T03:11:53 | 2020-05-29T03:11:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 635,969,785 | 11,272 | alexey-milovidov | > https://clickhouse-test-reports.s3.yandex.net/11272/e8f52337aacb261f55d72716a09bc85490464286/stress_test_(thread)/stderr.log
Already fixed by @KochetovNicolai | 2020-05-29T13:22:48 | 2020-05-29T13:22:48 | {} | MEMBER |
ClickHouse | ClickHouse | 634,785,011 | 10,577 | filimonov | Today faced the similar - after updating/restarting the server (to version 20.5.1) get back to some old clickhouse-client session, run some queries and when typed wrong query i get an error message had 'Received exception from server (version 20.4.1):' which is misleading.
| 2020-05-27T16:32:30 | 2020-05-29T13:28:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 635,979,816 | 11,228 | qoega | one query version:
``` sql
WITH 'weekend' AS grouping
SELECT sum(col1)
FROM
(
SELECT
1 AS col1,
toDate('2020-02-01') + number AS col2
FROM system.numbers
LIMIT toUInt64(toDate('2020-02-10') - toDate('2020-02-01')) + 1
)
GROUP BY dictGetUInt32(if(grouping = 'weekend', 'weekendsday... | 2020-05-29T13:43:33 | 2020-05-29T13:43:33 | {} | MEMBER |
ClickHouse | ClickHouse | 636,037,800 | 11,287 | den-crane | SELECT toStartOfDay(date) FROM temp ORDER BY date -- toStartOfDay in UTC TZ
JDBC converts 00:00 UTC to your client timezone +09
Try
SELECT toStartOfDay(date, **'Asia/Tokyo'**) FROM temp | 2020-05-29T15:32:01 | 2020-05-29T15:32:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 636,464,310 | 11,296 | alexey-milovidov | Note: this bugfix will be backported.
But the only line that needs to be backported is:
https://github.com/ClickHouse/ClickHouse/pull/11296/files#diff-70721e41703630a99a23b52236158e68R56 | 2020-05-31T12:25:02 | 2020-05-31T12:25:02 | {} | MEMBER |
ClickHouse | ClickHouse | 636,464,532 | 11,296 | alexey-milovidov | Here is the only commit that has to be backported: https://github.com/ClickHouse/ClickHouse/pull/11296/commits/9d39fc6dcbc49b8a6b187a30fb7a5354229efc76 | 2020-05-31T12:26:56 | 2020-05-31T12:26:56 | {} | MEMBER |
ClickHouse | ClickHouse | 635,886,714 | 11,282 | jencoldeng | > CREATE MATERIALIZED VIEW Summary_local
> ENGINE = ReplicatedSummingMergeTree
>
> this might help
Thans, but it dos not work.... :) | 2020-05-29T09:58:31 | 2020-05-29T09:58:31 | {} | NONE |
ClickHouse | ClickHouse | 635,976,759 | 11,270 | alexey-milovidov | > Stress test (thread) — Sanitizer assert (in stderr.log)
Already fixed by @KochetovNicolai
> 01108_restart_replicas_rename_deadlock
Added to CI fest. | 2020-05-29T13:37:06 | 2020-05-29T13:37:06 | {} | MEMBER |
ClickHouse | ClickHouse | 636,106,611 | 11,294 | CurtizJ | Can you try to update to the latest 20.1 patch release? There are several fixes for related issues. | 2020-05-29T17:59:01 | 2020-05-29T17:59:16 | {} | MEMBER |
ClickHouse | ClickHouse | 636,122,663 | 11,191 | UnamedRus | What will happen, if we run query like this:
SELECT * FROM table WHERE x = 124 and y = 'zoo'
Isn't it would scan 2 granules instead of 1? | 2020-05-29T18:31:19 | 2020-05-29T18:32:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 636,133,733 | 11,244 | den-crane | MV is an insert trigger. They see data from an insert only. They don't read a source table.
test_aggregated_hour_mv -- can not see data from the first insert. | 2020-05-29T18:56:42 | 2020-05-29T18:57:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 636,260,350 | 11,238 | den-crane | The only way is to detach partition from the table and attach a part with this row to MergeTree table and remove this row by "alter delete" or edit .sql file and make MergeTree instead of CollapsingMergeTree, then "alter delete" and make it CollapsingMergeTree back. | 2020-05-30T02:07:08 | 2020-05-30T02:07:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 636,274,045 | 11,194 | amitsharma100 | hi @amy108 how did you solve? | 2020-05-30T04:28:14 | 2020-05-30T04:28:14 | {} | NONE |
ClickHouse | ClickHouse | 635,885,552 | 11,282 | BernieJiangDL | CREATE MATERIALIZED VIEW Summary_local
ENGINE = ReplicatedSummingMergeTree
this might help | 2020-05-29T09:55:46 | 2020-05-29T09:56:08 | {} | NONE |
ClickHouse | ClickHouse | 635,898,479 | 11,281 | alesapin | You just need to delete the metadata file and data directory for each log table. In your case, it seems like you didn't delete `clickhouse/metadata/system/query_log.sql`. | 2020-05-29T10:26:23 | 2020-05-29T10:26:23 | {} | MEMBER |
ClickHouse | ClickHouse | 635,923,540 | 11,276 | alexey-milovidov | Further info (ru):
```
1. После параллельного парсинга, клиент отправлял на сервер блоки меньшего размера.
2. Это не должно было повлиять, так как сервер склеивает блоки в более крупные при вставке.
3. Но при вставке, ещё производится вычисление DEFAULT выражений.
4. У клиента DEFAULT выражения очень забористые.
... | 2020-05-29T11:32:15 | 2020-05-29T11:32:15 | {} | MEMBER |
ClickHouse | ClickHouse | 636,021,166 | 8,505 | den-crane | @jackpgao it works as expected. Timeout check happens in between blocks. You need more blocks
```
set max_execution_time = 1 ;
select sleep(1) from numbers(10) settings max_block_size=1;
Code: 159. DB::Exception: Received from localhost:9000. DB::Exception: Timeout exceeded: elapsed 1.999990824 seconds, maximu... | 2020-05-29T14:59:30 | 2020-05-29T14:59:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 636,123,773 | 11,296 | alexey-milovidov | Broken in https://github.com/ClickHouse/ClickHouse/pull/7181
Since 19.17: https://github.com/ClickHouse/ClickHouse/pull/7181/commits/627d48c19a0fc7e7149a2ec82ea54fb650a37ed5 | 2020-05-29T18:33:30 | 2020-05-29T18:34:25 | {} | MEMBER |
ClickHouse | ClickHouse | 636,158,919 | 6,417 | alexey-milovidov | The issue is no longer relevant, because in recent versions (since 19.17) the estimation (prediction) of query execution time is performed only after `timeout_before_checking_execution_speed` has been passed and only if this setting is set.
But I can still reproduce the same effect if I will set it to very low value... | 2020-05-29T19:53:42 | 2020-05-29T19:53:42 | {} | MEMBER |
ClickHouse | ClickHouse | 636,168,730 | 6,417 | alexey-milovidov | I have added a test. | 2020-05-29T20:05:54 | 2020-05-29T20:05:54 | {} | MEMBER |
ClickHouse | ClickHouse | 636,282,865 | 6,638 | azat | >we should have a cluster with both localhost and remote nodes;
test_cluster_two_shards is enough
> we should have LowCardinality on initiator and different String and LowCardinality(String) on remote nodes;
yes
>two level group by should be activated;
not necessary
>distributed_aggregation_memory_ef... | 2020-05-30T06:01:38 | 2020-05-30T06:01:38 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 636,369,575 | 11,291 | alexey-milovidov | > 2020-05-30 14:50:57 01223_dist_on_dist: [ FAIL ] 1.65 sec. - having stderror:
2020-05-30 14:50:57 [914e480291ef] 2020.05.30 14:50:56.663095 [ 12513 ] {01db1757-53e7-4352-9c49-cd9c9856e9c2} <Warning> ConnectionPoolWithFailover: Connection failed at try №1, reason: C... | 2020-05-30T18:47:01 | 2020-05-30T18:47:01 | {} | MEMBER |
ClickHouse | ClickHouse | 635,972,075 | 11,272 | alexey-milovidov | @nikitamikhaylov FYI one potential issue: arguments with side effects may start to evaluate twice. | 2020-05-29T13:27:26 | 2020-05-29T13:27:26 | {} | MEMBER |
ClickHouse | ClickHouse | 636,184,690 | 6,228 | alexey-milovidov | I have added a test in #11296 | 2020-05-29T20:46:19 | 2020-05-29T20:46:19 | {} | MEMBER |
ClickHouse | ClickHouse | 636,186,634 | 11,300 | alexey-milovidov | > Capture unhandled exceptions
Unhandled exception is processed with std::terminate and abort, then it will be processed with signal handler. | 2020-05-29T20:51:16 | 2020-05-29T20:51:16 | {} | MEMBER |
ClickHouse | ClickHouse | 636,298,981 | 11,300 | blinkov | > transparency - store reports sended to be able to review what sended
@filimonov Sentry SDK doesn't seem to have such an option, the closest thing I see is using an HTTP proxy, where you can dump what goes through. We could file an issue in https://github.com/getsentry/sentry-native though, maybe they'll suggest/a... | 2020-05-30T08:30:27 | 2020-05-30T08:30:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 636,371,615 | 6,638 | alexey-milovidov | Ok. I will commit this test as is and close... | 2020-05-30T19:05:29 | 2020-05-30T19:05:29 | {} | MEMBER |
ClickHouse | ClickHouse | 636,372,169 | 4,616 | alexey-milovidov | @filimonov If it's already fixed, why don't close? | 2020-05-30T19:10:40 | 2020-05-30T19:10:40 | {} | MEMBER |
ClickHouse | ClickHouse | 636,410,045 | 11,201 | jackpgao | > @jackpgao could you publish that configuration?
- config of dict
```xml
<?xml version="1.0" ?>
<yandex>
<dictionary>
<name>vip_did</name>
<source>
<mysql>
<port>xxxx</port>
<user>xxxx</user>
<password>xxxx</password>
<replica>
<host>xxxx.com.... | 2020-05-31T02:11:28 | 2020-05-31T02:11:28 | {} | NONE |
ClickHouse | ClickHouse | 636,479,875 | 11,293 | alexey-milovidov | First commit 4e9a326f122de208ba314f7292242e535ee3d16a will go to master. | 2020-05-31T14:37:35 | 2020-05-31T14:37:35 | {} | MEMBER |
ClickHouse | ClickHouse | 636,492,548 | 11,309 | alexey-milovidov | We can simply assign the merge regardless to the already running merges.
This is perfectly legal when we are merging whole partition (so no intersecting parts were created). | 2020-05-31T16:15:50 | 2020-05-31T16:15:50 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.