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 | 722,413,689 | 16,689 | den-crane | @weiqxu the problem that your query `select sum(null) from remote('xxx', 'system', 'one') ;` looks syntactic.
What is the real usecase / query? | 2020-11-05T14:30:05 | 2020-11-05T14:30:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 722,483,682 | 16,683 | nikitamikhaylov | This branch doesn't contain t-tests. | 2020-11-05T16:23:35 | 2020-11-05T16:23:35 | {} | MEMBER |
ClickHouse | ClickHouse | 722,597,668 | 16,718 | den-crane | related to WITH FILL STEP month (expression)
https://github.com/ClickHouse/ClickHouse/issues/16654#issuecomment-721483515
https://github.com/ClickHouse/ClickHouse/issues/7752 | 2020-11-05T19:37:37 | 2020-11-05T19:39:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 722,628,587 | 16,435 | alexey-milovidov | No changes. | 2020-11-05T20:34:38 | 2020-11-05T20:34:38 | {} | MEMBER |
ClickHouse | ClickHouse | 722,814,628 | 5,618 | DipalPrajapati | I still see the same issue. I dont see clickhouse writing to disk :( | 2020-11-06T04:52:31 | 2020-11-06T04:52:31 | {} | NONE |
ClickHouse | ClickHouse | 723,243,567 | 16,721 | nikitamikhaylov | `Read too much text and still can't parse a query. Aborting.` - weird error, ignore. Stress test (undefined) is broken in master. | 2020-11-06T18:53:05 | 2020-11-06T18:53:05 | {} | MEMBER |
ClickHouse | ClickHouse | 722,481,996 | 16,686 | nikitamikhaylov | This branch doesn't contain t-tests. | 2020-11-05T16:20:53 | 2020-11-05T16:20:53 | {} | MEMBER |
ClickHouse | ClickHouse | 722,565,046 | 16,367 | alexey-milovidov | The only unresolved issue: sometimes server can create cut metadata files - only when run with UBSan in stress test in Sandbox and only when the test that quickly creates and drops 10 000 tables is present. The issue does not reproduce with release, ASan, MSan, TSan, and we can safely continue. | 2020-11-05T18:40:08 | 2020-11-05T18:40:50 | {} | MEMBER |
ClickHouse | ClickHouse | 722,720,703 | 16,732 | Fallenyasha | Чуть больше исследовал, посмотрел потребление памяти при выполнении запросов инсерта, вставлял блоками по ~2М записей, получается следующая картина
**185.00 MiB**
**353.01 MiB**
72.26 MiB
82.23 MiB
**452.26 MiB**
80.82 MiB
81.38 MiB
82.51 MiB
**836.26 MiB**
76.76 MiB
83.88 MiB
77.32 MiB
74.88 MiB
77.76 Mi... | 2020-11-06T00:17:34 | 2020-11-06T00:19:19 | {} | NONE |
ClickHouse | ClickHouse | 723,051,112 | 13,252 | CurtizJ | It's not working in current implementation. I had some troubles with mutations in selection of parts to mutate. In general to implement this logic the such action as `DROP PART` is required. It's not easy due to algorithms of selection parts in merges and mutations. Also there is another attempt to implement it #15511... | 2020-11-06T12:17:49 | 2020-11-06T12:17:49 | {} | MEMBER |
ClickHouse | ClickHouse | 723,078,991 | 16,610 | zhang2014 | AST fuzzer — Assertion `V && "PHI node got a null value!"' failed -- looks unrelated
test_https_replication/test.py::test_replication_after_partition -- looks unrelated
test_distributed_ddl_parallel/test.py::test_two_in_parallel_two_queued -- looks unrelated
| 2020-11-06T13:24:33 | 2020-11-06T13:24:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 722,429,487 | 16,689 | alexey-milovidov | How to reproduce:
```
select sum(null) from remote('127.0.0.{1,2}', 'system', 'one')
``` | 2020-11-05T14:55:42 | 2020-11-05T14:55:42 | {} | MEMBER |
ClickHouse | ClickHouse | 722,450,266 | 6,679 | CurtizJ | Implemented in #14742. | 2020-11-05T15:29:29 | 2020-11-05T15:29:29 | {} | MEMBER |
ClickHouse | ClickHouse | 722,483,610 | 16,684 | nikitamikhaylov | This branch doesn't contain t-tests. | 2020-11-05T16:23:29 | 2020-11-05T16:23:29 | {} | MEMBER |
ClickHouse | ClickHouse | 722,555,594 | 13,274 | TheGreatAbyss | Hello
I've encountered what I think is a related issue to the above. I have two tables `table_a` and `table_b` that have materialized columns `column_a` and `column_b` respectively. When I try to write a query that joins the tables and uses those columns:
```
SELECT column_a, column_b
FROM table_a a
AL... | 2020-11-05T18:23:28 | 2020-11-05T18:23:28 | {} | NONE |
ClickHouse | ClickHouse | 722,859,230 | 16,591 | filimonov | > I'm wondering, does it make any sense for queries with File table engine to use the format settings from the current query context? I think not, the format settings are a property of the table, and you just break the table if you change that. OTOH, `file` table function should always use settings from current context... | 2020-11-06T05:54:20 | 2020-11-06T05:54:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 723,036,077 | 16,613 | nikitamikhaylov | Stress test (undefined) broken in master. | 2020-11-06T11:41:01 | 2020-11-06T11:41:01 | {} | MEMBER |
ClickHouse | ClickHouse | 722,406,847 | 16,711 | alexey-milovidov | If you're asking why metadata in ZooKeeper was accidentially deleted, the answer is - we don't know and it looks like an error on your side. | 2020-11-05T14:18:30 | 2020-11-05T14:18:30 | {} | MEMBER |
ClickHouse | ClickHouse | 722,424,835 | 16,705 | Slach | @mikhno-s
I can't successfully reproduce your queries
your bug report contains the following errors:
- `toYYYY` is not exists function
- PARTITION BY + generateRandtom raise "Too Many partitions used" for INSERT INTO SELECT
- `generateRandom('timestamp DateTime64(3), string String()', 1, 32, 2)` can't contains S... | 2020-11-05T14:48:09 | 2020-11-05T14:48:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 722,527,320 | 14,591 | alexey-milovidov | I write
```
Fix function `has` with `LowCardinality` of `Nullable`.
``` | 2020-11-05T17:34:56 | 2020-11-05T17:35:05 | {} | MEMBER |
ClickHouse | ClickHouse | 722,555,908 | 16,718 | alexey-milovidov | Maybe we already have everything that we need?
```
milovidov-desktop :) SELECT today() + INTERVAL number MONTH FROM numbers(20)
SELECT today() + toIntervalMonth(number)
FROM numbers(20)
Query id: 6fd58183-95ac-415e-ae15-29cd67af6272
┌─plus(today(), toIntervalMonth(number))─┐
│ ... | 2020-11-05T18:24:03 | 2020-11-05T18:24:03 | {} | MEMBER |
ClickHouse | ClickHouse | 722,628,370 | 16,637 | azat | @CurtizJ can you please take a look? | 2020-11-05T20:34:12 | 2020-11-05T20:34:12 | {} | MEMBER |
ClickHouse | ClickHouse | 722,647,929 | 16,727 | azat | >Fast test requires explicit linking of configs.
Thanks! By some reason I though that `programs/server/config.d` is always used in tests. | 2020-11-05T21:12:01 | 2020-11-05T21:12:01 | {} | MEMBER |
ClickHouse | ClickHouse | 722,677,638 | 16,719 | vzakaznikov | @alexey-milovidov, Hi Alexey, this PR includes instrumentation to debug `ALTER` fails. TestFlows check is green. This PR is ready for merge. | 2020-11-05T22:16:44 | 2020-11-05T22:16:44 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 722,940,767 | 16,526 | BayoNet | I think we need just close this PR.
Another person will do this task. | 2020-11-06T08:10:38 | 2020-11-06T08:10:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 722,404,012 | 16,695 | den-crane | >cat /etc/clickhouse-server/config.d/delimiter.xml
Wrong folder for user settings.
Should be /etc/clickhouse-server/**users.d**/delimiter.xml
config.d -- server settings (config.xml)
users.d -- users settings (user.xml)
conf.d -- any (config.xml and user.xml) | 2020-11-05T14:14:00 | 2020-11-05T14:14:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 722,759,899 | 16,726 | zhang2014 | This requires us to implement several authentication methods to MySQL(only mysql_native_password is currently supported) | 2020-11-06T01:58:56 | 2020-11-06T01:58:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 722,909,042 | 16,711 | xjxyxgq | I think I found reason.
Maybe I should specify this parameter `database_atomic_delay_before_drop_table_sec`, but I'm not found it from document.
In the early testing process, we may frequently delete and rebuild the same table, it can be misleading.
~~~
void DatabaseCatalog::loadDatabases()
{
drop_dela... | 2020-11-06T06:49:24 | 2020-11-06T06:49:24 | {} | NONE |
ClickHouse | ClickHouse | 722,987,592 | 16,726 | bohutang | To support MariaDB, the main work is not the auth plugin, MariaDB GTID format is incompatible with MySQL, it's lot work Todo. | 2020-11-06T09:53:05 | 2020-11-06T09:53:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 723,079,314 | 16,338 | ucasfl | Yes, I also thought of the first point. In fact, I'd like to think of the query as a chain, if we have UNION DISTINCT, we can flatten all nodes before it and its child nodes. But I donn't understand the second point clearly? | 2020-11-06T13:25:16 | 2020-11-06T13:25:16 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 723,133,447 | 16,338 | akuzm | To rephrase #2, if you have UNION ALL and its child node is UNION DISTINCT, the child nodes of this UNION DISTINCT stay below it, we cannot move them to the parent UNION ALL.
Let's use `D` for `UNION DISTINCT`, `A` for `UNION ALL`, numbers for subselects, and Polish notation.
This query:
`(select 1 union distinct se... | 2020-11-06T15:14:52 | 2020-11-06T15:14:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 723,232,293 | 7,663 | alexey-milovidov | The #13924 is a good alternative. | 2020-11-06T18:27:41 | 2020-11-06T18:27:41 | {} | MEMBER |
ClickHouse | ClickHouse | 722,405,686 | 16,694 | den-crane | duplicate https://github.com/ClickHouse/ClickHouse/issues/11755 | 2020-11-05T14:16:40 | 2020-11-05T14:16:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 722,796,323 | 16,253 | amosbird | > > Actually it can be rewritten into smth like
>
> Yeah. It's possible. Equality is another special operator that can be treated as an always monotonic indicator for deterministic functions. I'll extend it.
Now we can do `where x = 100` and partition pruning will work correctly. @filimonov Could you please check... | 2020-11-06T03:53:46 | 2020-11-06T03:53:46 | {
"hooray": 2
} | COLLABORATOR |
ClickHouse | ClickHouse | 723,037,232 | 13,256 | alexey-milovidov | Скорее всего да. | 2020-11-06T11:43:51 | 2020-11-06T11:43:51 | {} | MEMBER |
ClickHouse | ClickHouse | 722,859,384 | 16,591 | filimonov | BTW `clickhouse-local` creates engine=file and don't construct a settings block for that.
So it will just stop working I guess. | 2020-11-06T05:54:43 | 2020-11-06T13:49:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 723,176,432 | 16,636 | alexey-milovidov | Still does not. | 2020-11-06T16:36:17 | 2020-11-06T16:36:17 | {} | MEMBER |
ClickHouse | ClickHouse | 723,231,321 | 13,065 | alexey-milovidov | Looks like a real issue but we need to reproduce it again - to get the full chain of "value was stored at..." diagnostics. | 2020-11-06T18:25:43 | 2020-11-06T18:25:43 | {} | MEMBER |
ClickHouse | ClickHouse | 722,427,236 | 16,705 | den-crane | @Slach it's not a bug. It's a mistake in the documentation. I am fixing it.
https://github.com/ClickHouse/ClickHouse/pull/16712 | 2020-11-05T14:52:07 | 2020-11-05T15:22:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 722,447,669 | 11,228 | alexey-milovidov | This issue looks related to the changes that allowed to pass table name in dictGet. @amosbird Could you please take a look? | 2020-11-05T15:25:12 | 2020-11-05T15:25:12 | {} | MEMBER |
ClickHouse | ClickHouse | 722,635,033 | 16,620 | akuzm | It still doesn't work -- now the robot doesn't have enough priviliges -- but I think it's good enough to merge. The test failures are not related, because I'm only touching the perf test. | 2020-11-05T20:47:22 | 2020-11-05T20:47:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 723,173,657 | 12,580 | CurtizJ | I didn't reproduce your case, but I investigated Poco redis library and fixed in #13082 the only case, that could lead to this kind of errors. So, mostly likely it is already fixed. | 2020-11-06T16:30:49 | 2020-11-06T16:30:49 | {} | MEMBER |
ClickHouse | ClickHouse | 723,189,828 | 12,615 | filimonov | @npetrusic how many protobuf messages do you put in the single Kafka message?
Can you check if `SELECT * FROM kafka_table LIMIT 1` works?
Can you provide a binary dump of the problematic message example? | 2020-11-06T17:01:07 | 2020-11-06T17:01:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 722,444,152 | 11,294 | alexey-milovidov | This issue has been fixed in the patch release long time ago. | 2020-11-05T15:19:29 | 2020-11-05T15:19:29 | {} | MEMBER |
ClickHouse | ClickHouse | 722,457,713 | 12,975 | alexey-milovidov | 2. Maybe the issue is related to the `input_format_parallel_parsing` setting. The setting is enabled by default. You can check this hypothesis by disabling the setting. | 2020-11-05T15:41:17 | 2020-11-05T15:41:17 | {} | MEMBER |
ClickHouse | ClickHouse | 722,458,164 | 12,975 | alexey-milovidov | 3. It argues about the `row 650`. Could you please copy-paste the contents of this row verbatim? | 2020-11-05T15:42:04 | 2020-11-05T15:42:04 | {} | MEMBER |
ClickHouse | ClickHouse | 722,526,600 | 14,591 | alexey-milovidov | The message for changelog is not actually useful. | 2020-11-05T17:33:50 | 2020-11-05T17:33:50 | {} | MEMBER |
ClickHouse | ClickHouse | 722,917,173 | 16,253 | amosbird | > I guess the next thing to support would be IN / Not In operators.
I'm afraid it's not possible to support either `NOT IN` or `NOT EQUAL`. | 2020-11-06T07:06:38 | 2020-11-06T07:15:46 | {
"+1": 1
} | COLLABORATOR |
ClickHouse | ClickHouse | 722,979,250 | 11,228 | pic-Nick | It seems to be fixed on 20.7.2.30 already. | 2020-11-06T09:35:10 | 2020-11-06T09:56:31 | {} | NONE |
ClickHouse | ClickHouse | 723,193,933 | 3,994 | azat | >Exactly. And i'd like to see a server, executed a query, logged in query_log table, not a host handled SELECT * FROM query_log_distributed request.
>So, this won't solve my problem, unfortunately
You can use [`_shard_num`](https://clickhouse.tech/docs/en/engines/table-engines/special/distributed/#virtual-columns) ... | 2020-11-06T17:08:41 | 2020-11-06T17:08:41 | {} | MEMBER |
ClickHouse | ClickHouse | 723,197,049 | 12,718 | alexey-milovidov | @qoega UBSan reports in 3rd-party libraries don't necessarily indicate a bug.
Most of the 3rd-party C++ libraries have low development culture and don't meant to be built with UBSan.
Here we see misaligned memory access that is an undefined behaviour in C++.
Compiler can exploit the assumption that misaligned memo... | 2020-11-06T17:14:50 | 2020-11-06T17:14:50 | {} | MEMBER |
ClickHouse | ClickHouse | 723,207,331 | 12,629 | alexey-milovidov | @qoega UBSan reports in 3rd-party libraries don't necessarily indicate a bug.
Most of the 3rd-party C++ libraries have low development culture and don't meant to be built with UBSan.
You should first kindly ask the author of this 3rd party library, does they at least familiar of the existence of UBSan. | 2020-11-06T17:35:37 | 2020-11-06T17:35:37 | {} | MEMBER |
ClickHouse | ClickHouse | 723,223,385 | 12,493 | alexey-milovidov | ```
Syntax error: failed at position 19 ('CODEC'):
CREATE TABLE t (c CODEC)
Expected one of: CODEC, REMOVE, ALIAS, DEFAULT, MATERIALIZED, COMMENT
``` | 2020-11-06T18:09:11 | 2020-11-06T18:09:11 | {} | MEMBER |
ClickHouse | ClickHouse | 723,230,533 | 10,004 | azat | @alexey-milovidov I guess can be closed? (no details, no interest and there were some changes anyway since then) | 2020-11-06T18:24:03 | 2020-11-06T18:24:03 | {} | MEMBER |
ClickHouse | ClickHouse | 723,239,346 | 16,252 | alexey-milovidov | We decided to merge this PR without a test and @filimonov will help to write it. | 2020-11-06T18:43:52 | 2020-11-06T18:43:52 | {} | MEMBER |
ClickHouse | ClickHouse | 722,406,148 | 16,711 | alexey-milovidov | If you accidentially delete metadata in ZooKeeper, data in the table will remain safe and you can continue to query it.
It was designed this way to prevent data loss due to accidents and operation errors. | 2020-11-05T14:17:18 | 2020-11-05T14:17:18 | {} | MEMBER |
ClickHouse | ClickHouse | 722,452,245 | 5,486 | CurtizJ | Long time ago query `ALTER TABLE ... MATERIALIZE TTL` was introduced. #8775. | 2020-11-05T15:32:30 | 2020-11-05T15:32:30 | {} | MEMBER |
ClickHouse | ClickHouse | 722,483,531 | 16,685 | nikitamikhaylov | This branch doesn't contain t-tests. | 2020-11-05T16:23:21 | 2020-11-05T16:23:21 | {} | MEMBER |
ClickHouse | ClickHouse | 722,525,110 | 14,697 | alexey-milovidov | Not a bugfix. | 2020-11-05T17:31:27 | 2020-11-05T17:31:27 | {} | MEMBER |
ClickHouse | ClickHouse | 722,649,273 | 11,133 | azat | >distributed engine does not remove already existing empty directories, and this is pretty easy to fix, and is on my todo list, and this will help in your case, but if and only if, those directories are empty.
There are some complexity to do this by TTL, but at server start it was done in #16729 | 2020-11-05T21:14:51 | 2020-11-05T21:14:51 | {} | MEMBER |
ClickHouse | ClickHouse | 722,800,371 | 16,729 | azat | >Functional stateless tests (thread) — fail: 1, passed: 2407, skipped: 14
00840_long_concurrent_select_and_drop_deadlock - flacky | 2020-11-06T04:06:31 | 2020-11-06T04:06:31 | {} | MEMBER |
ClickHouse | ClickHouse | 722,990,585 | 13,256 | draev | > Check that you have configured access_control_path in config.xml
>
> ```
> <!-- Path to folder where users and roles created by SQL commands are stored. -->
> <access_control_path>/var/lib/clickhouse/access/</access_control_path>
> ```
А Сервер нужно перезагружать после изменения конфига ? | 2020-11-06T09:59:27 | 2020-11-06T09:59:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 723,208,322 | 12,719 | alexey-milovidov | @qoega UBSan reports in 3rd-party libraries don't necessarily indicate a bug.
Most of the 3rd-party C++ libraries have low development culture and don't meant to be built with UBSan.
You should first kindly ask the author of this 3rd party library, does they at least familiar of the existence of UBSan. | 2020-11-06T17:37:43 | 2020-11-06T17:37:43 | {} | MEMBER |
ClickHouse | ClickHouse | 722,410,980 | 16,711 | den-crane | >2020.11.05 20:39:13.043298 [ 13301 ] {} <Information> DatabaseCatalog: Will try drop my.test1 (28ed7dac-4a78-4115-adfb-02e78e88de31)
Do you mean that you dropped the table and and data stayed on disk for >10 minutes?
CH version 20.10 ? Database atomic? | 2020-11-05T14:25:31 | 2020-11-05T14:25:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 722,478,500 | 16,714 | alexey-milovidov | CC @abyss7 @azat | 2020-11-05T16:15:08 | 2020-11-05T16:15:08 | {} | MEMBER |
ClickHouse | ClickHouse | 722,562,173 | 16,703 | alexey-milovidov | One test has to be updated. | 2020-11-05T18:35:08 | 2020-11-05T18:35:08 | {} | MEMBER |
ClickHouse | ClickHouse | 722,939,534 | 15,998 | winter7 | When I tried to execute `REVOKE SELECT ON mytable FROM john`, I got
```
Code: 495. DB::Exception: Received from localhost:9000. DB::Exception: Cannot update User `john` in users.xml because this storage is readonly. (version 20.4.4.18 (official build))
```
even when `users.xml` file was in `666` mod.
Other users... | 2020-11-06T08:07:33 | 2020-11-06T08:07:33 | {} | NONE |
ClickHouse | ClickHouse | 722,863,644 | 16,253 | filimonov | > Now we can do where x = 100 and partition pruning will work correctly. @filimonov Could you please check if it covers common cases?
Great progress! I guess the next thing to support would be IN <s>Not In</s> operators. | 2020-11-06T05:59:44 | 2020-11-06T13:42:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 723,249,486 | 16,123 | alexey-milovidov | @ucasFL We forgot about one little thing...
For -State and -Merge functions we need to add -OrNull before, not after...
Need to check how it will work. | 2020-11-06T19:06:22 | 2020-11-06T19:06:22 | {} | MEMBER |
ClickHouse | ClickHouse | 723,256,731 | 12,494 | alexey-milovidov | @qoega
> Unknown numeric column of type: DB::ColumnAggregateFunction
is a different message than
> Unknown numeric column of type: DB::IColumn const*
The original issue was fixed long time ago. | 2020-11-06T19:23:48 | 2020-11-06T19:23:48 | {} | MEMBER |
ClickHouse | ClickHouse | 722,329,852 | 16,559 | abyss7 | > @abyss7 How is it possible that the bugfix release contains any "Improvements"?
> Should check carefully, maybe changelog is incorrect.
I used script from @akuzm - will investigate before trying to generate changelists for other versions. Looks like it wasn't used some time because it was still broken after migra... | 2020-11-05T11:51:18 | 2020-11-05T15:26:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 722,469,826 | 4,061 | alexey-milovidov | @NemoAA `Resource temporarily unavailable` may happen if DNS is not working (misconfigured) on the server. | 2020-11-05T16:01:03 | 2020-11-05T16:01:03 | {} | MEMBER |
ClickHouse | ClickHouse | 722,773,834 | 16,711 | xjxyxgq | > > 2020.11.05 20:39:13.043298 [ 13301 ] {} DatabaseCatalog: Will try drop my.test1 (28ed7dac-4a78-4115-adfb-02e78e88de31)
>
> Do you mean that you dropped the table and and data stayed on disk for >10 minutes?
> CH version 20.10 ? Database atomic?
I'm sorry I didn't describe it in detail enough.
Version is ... | 2020-11-06T02:39:26 | 2020-11-06T06:19:04 | {} | NONE |
ClickHouse | ClickHouse | 723,197,368 | 12,714 | alexey-milovidov | @qoega UBSan reports in 3rd-party libraries don't necessarily indicate a bug.
Most of the 3rd-party C++ libraries have low development culture and don't meant to be built with UBSan.
You should first kindly ask the author of this 3rd party library, does they at least familiar of the existence of UBSan. | 2020-11-06T17:15:27 | 2020-11-06T17:16:02 | {} | MEMBER |
ClickHouse | ClickHouse | 723,211,804 | 12,719 | qoega | The idea is to suppress problem or disable a test in ubsan run and allow CI runs with ubsan. Issue is just a place for information on the problem | 2020-11-06T17:45:11 | 2020-11-06T17:45:11 | {} | MEMBER |
ClickHouse | ClickHouse | 723,257,238 | 12,494 | alexey-milovidov | @qoega You forgot to attach the reproducer for this assertion failure. | 2020-11-06T19:25:03 | 2020-11-06T19:25:03 | {} | MEMBER |
ClickHouse | ClickHouse | 717,231,759 | 16,391 | lisun | https://github.com/lisun/homebrew-clickhouse/
Add homebrew tap:
brew tap lisun/clickhouse
Install ClickHouse :
brew upgrade
brew install clickhouse
start server:
brew services start clickhouse
I fork and update a brew for mac
Welcome to use! | 2020-10-27T13:09:48 | 2020-11-05T14:53:06 | {} | NONE |
ClickHouse | ClickHouse | 722,968,131 | 16,626 | chang0391 | If it can automatically select the appropriate method, hash or partial merge, according to the size of the memory used, it will be better | 2020-11-06T09:12:30 | 2020-11-06T09:12:30 | {} | NONE |
ClickHouse | ClickHouse | 723,085,825 | 16,253 | filimonov | > I'm afraid it's not possible to support either NOT IN or NOT EQUAL.
Yes, of course, you're right. 'Not' can't work | 2020-11-06T13:39:57 | 2020-11-06T13:39:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 723,148,722 | 15,713 | amosbird | It sounds like a reasonable feature request. Perhaps we can somehow extend `LIMIT BY`? https://github.com/ClickHouse/ClickHouse/issues/111
```
limit 5 by page 2 by browser 2 by os
``` | 2020-11-06T15:43:29 | 2020-11-06T15:43:29 | {
"+1": 1
} | COLLABORATOR |
ClickHouse | ClickHouse | 722,464,174 | 13,014 | alexey-milovidov | Still relevant on 20.11. | 2020-11-05T15:52:03 | 2020-11-05T15:52:03 | {} | MEMBER |
ClickHouse | ClickHouse | 722,476,610 | 11,228 | amosbird | I cannot reproduce any of these in 20.11.1.1 | 2020-11-05T16:12:07 | 2020-11-05T16:12:07 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 722,509,020 | 14,893 | ns-gzhang | I'd like to add two use cases from our real scenarios: (1) transfer aggregates from Druid to ClickHouse for migration. The challenge is that the summary data are kept longer than the details and we don't have detailed data anymore and won't be able to get equivalent aggregates in Druid into ClickHouse except using the ... | 2020-11-05T17:04:45 | 2020-11-05T17:04:45 | {} | NONE |
ClickHouse | ClickHouse | 722,774,039 | 16,695 | trollhe | > > cat /etc/clickhouse-server/config.d/delimiter.xml
>
> Wrong folder for user settings.
> Should be /etc/clickhouse-server/**users.d**/delimiter.xml
>
> config.d -- server settings (config.xml)
> users.d -- users settings (user.xml)
> conf.d -- any (config.xml and user.xml)
thanks den-crane,
i'm try to... | 2020-11-06T02:40:08 | 2020-11-06T02:48:45 | {} | NONE |
ClickHouse | ClickHouse | 722,800,252 | 16,727 | azat | >Functional stateless tests (thread) — fail: 1, passed: 2408, skipped: 14
00840_long_concurrent_select_and_drop_deadlock flacky | 2020-11-06T04:06:07 | 2020-11-06T04:06:07 | {} | MEMBER |
ClickHouse | ClickHouse | 723,055,291 | 16,338 | akuzm | I think I've seen your comment before but it's gone now...
Anyway, suppose we have an AST for UNIONs. It is a binary tree, each node is either a plain select or another union. We want to flatten it for execution, into a structure where we have just a list of SELECT and then apply UNION over all of them. Currently, wit... | 2020-11-06T12:28:28 | 2020-11-06T12:28:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 723,089,415 | 15,998 | filimonov | > When I tried to execute `REVOKE SELECT ON mytable FROM john`, I got
>
> ```
> Code: 495. DB::Exception: Received from localhost:9000. DB::Exception: Cannot update User `john` in users.xml because this storage is readonly. (version 20.4.4.18 (official build))
> ```
>
> even when `users.xml` file was in `666` m... | 2020-11-06T13:47:39 | 2020-11-06T13:47:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 722,454,469 | 13,173 | CurtizJ | Now version 20.4 is outdated and out of support. | 2020-11-05T15:36:06 | 2020-11-05T15:36:06 | {} | MEMBER |
ClickHouse | ClickHouse | 722,457,180 | 12,975 | alexey-milovidov | @calebeaires We have the following hypothesis:
1. When setting `format_csv_allow_single_quotes` is enabled, ClickHouse will recognize both single and double quotes. But usually CSV files have only one type of quoting (double quotes by RFC) and parsing single quotes as the beginning of the string literal may break th... | 2020-11-05T15:40:21 | 2020-11-05T15:40:21 | {
"+1": 2
} | MEMBER |
ClickHouse | ClickHouse | 722,579,767 | 14,414 | alexey-milovidov | ```
SELECT
check_name,
test_name,
count() AS c,
min(check_start_time) AS min_time,
max(check_start_time) AS max_time,
round(c * 86400 / (max_time - min_time), 3) AS frequency,
argMax(commit_sha, check_start_time) AS sha
FROM `gh-data`.checks
WHERE
test_status = 'FAIL'
--... | 2020-11-05T19:06:45 | 2020-11-05T19:09:02 | {} | MEMBER |
ClickHouse | ClickHouse | 722,933,053 | 16,726 | maqroll | MariaDB handles **mysql_native_password** just fine. It doesn't provide the **default_authentication_plugin** variable to check beforehand.
Without that check one of two things could happen: either the database provides that authentication method and the connection will establish or not (in that case the connection... | 2020-11-06T07:50:51 | 2020-11-06T07:50:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 723,187,784 | 12,615 | filimonov | Can't reproduce it :| See #16762 | 2020-11-06T16:57:21 | 2020-11-06T16:57:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 723,201,893 | 16,512 | filimonov | Yes. There is known issue in older versions of ClickHouse - when DROP of kafka table can hang.
see #11145 #10910 #15473
Please update to 20.8 or newer.
| 2020-11-06T17:24:26 | 2020-11-06T17:24:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 723,204,166 | 8,262 | filimonov | @alexey-milovidov Kafka itself can't check if target tables are ready.
We should not lose any data despite that message (it can't flush the data, so it will not commit, so will keep retrying).
So it's more a cosmetic issue than a real problem.
We can wait for https://github.com/ClickHouse/ClickHouse/issues/1... | 2020-11-06T17:28:54 | 2020-11-06T17:33:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 723,207,834 | 12,490 | alexey-milovidov | This data race in OpenSSL is bening. | 2020-11-06T17:36:39 | 2020-11-06T17:36:39 | {} | MEMBER |
ClickHouse | ClickHouse | 723,232,564 | 13,924 | alexey-milovidov | That's a good idea :+1:
To "publish" subset of config parameters that we want to publish and that were in effect.
@filimonov
It's much better than simply export all the configuration. | 2020-11-06T18:28:11 | 2020-11-06T18:28:33 | {} | MEMBER |
ClickHouse | ClickHouse | 799,015,980 | 19,907 | den-crane | Is it fixed? When?
```
Code: 246, e.displayText() = DB::Exception: Bad size of marks file '/data/clickhouse/data/default
/.../20210314_6786837_6786931_19/skp_idx_timeidx.mrk3': 72, must be: 96 (version 20.12.8.5 (official build))
```
| 2021-03-15T00:39:28 | 2021-03-15T00:39:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 799,156,227 | 21,660 | shadowDy | @amosbird
Thank you for your reply
In fact, the basic table is a table with tens of billions of data, which can be written in real time, query the data of the past three hours every minute, and aggregate the data by column a + minute granularity. Calculate the mean value, quantile value and other operations.
In m... | 2021-03-15T06:38:36 | 2021-03-15T06:38:36 | {} | NONE |
ClickHouse | ClickHouse | 799,366,616 | 21,268 | kssenii | > @kssenii Hi! Maybe it's better to merge my PR first, and then merge the changes from the second PR into it?
Hi! It seems too complicated to make any more changes to that second PR, it is too broken. Lets wait some more time, please. | 2021-03-15T12:05:30 | 2021-03-15T12:06:49 | {} | MEMBER |
ClickHouse | ClickHouse | 799,635,108 | 9,404 | alexey-milovidov | Fixed. | 2021-03-15T18:10:22 | 2021-03-15T18:10:22 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.