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 | 825,703,979 | 9,514 | nikitamikhaylov | Also this version is out of support, because it too old. But I've added a test with `TTL MOVE TO VOLUME` #23518 | 2021-04-23T14:38:14 | 2021-04-23T14:38:14 | {} | MEMBER |
ClickHouse | ClickHouse | 825,874,518 | 23,517 | azat | >Ok. Now 01675_distributed_bytes_to_delay_insert is flaky.
Yes, that's why I've marked as draft, now 01675_distributed_bytes_to_delay_insert also fixed | 2021-04-23T19:32:14 | 2021-04-23T19:32:14 | {} | MEMBER |
ClickHouse | ClickHouse | 825,984,600 | 23,539 | l1t1 | ```sql
with t as (select number n from system.numbers limit 1000000000),
t1 as (select n%100 x,count(*)c from t group by x),
t2 as (select toUInt8(x/10)x1,groupArray(c)ac from t1 group by x1)
select t2.*,2 from t2
union all
select 0,sumForEach(ac) ac,3 from t2
```
also showed scanned twice sql as fast as scan o... | 2021-04-23T23:03:35 | 2021-04-23T23:03:35 | {} | NONE |
ClickHouse | ClickHouse | 826,005,423 | 23,559 | filimonov | Try to clone repo from the scratch into another folder or (better) do that on the another node. Or just check diffoscope report in the issue. There was much more differences. For ex. From obvious things - order of system.licences etc. | 2021-04-24T00:36:52 | 2021-04-24T01:28:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 826,022,019 | 23,560 | l1t1 | groupArray
Syntax: groupArray(x) or groupArray(max_size)(x)
Creates an array of argument values.
Values can be added to the array in any (indeterminate) order.
The second version (with the max_size parameter) limits the size of the resulting array to max_size elements. For example, groupArray(1)(x) is equivale... | 2021-04-24T02:42:14 | 2021-04-24T02:42:14 | {} | NONE |
ClickHouse | ClickHouse | 826,024,669 | 23,560 | l1t1 | eg.
with t as (select number n from system.numbers limit 100000),
t1 as (select n%20 x,count(*)c from t group by x),
t2 as (select toUInt8(x/4)x1,groupArray(c)ac from t1 group by x1)
--select t2.*,2 from t2
--union all
select x1% 3 x2,groupArray(ac) ac,3 from t2 group by x2
; | 2021-04-24T03:06:26 | 2021-04-24T03:06:26 | {} | NONE |
ClickHouse | ClickHouse | 826,072,066 | 23,508 | tavplubix | Thanks! It helped to find out the reason of the error:
```
:) select version(), toTypeName(toInt64(123) % 100)
┌─version()─┬─toTypeName(modulo(toInt64(123), 100))─┐
│ 21.2.5.5 │ Int8 │
└───────────┴───────────────────────────────────────┘
:) select version(), toTypeName(toInt... | 2021-04-24T10:33:07 | 2021-04-24T10:33:07 | {} | MEMBER |
ClickHouse | ClickHouse | 826,140,155 | 23,517 | mergify[bot] | **Command `update`: success**
> **Branch has been successfully updated**
>
| 2021-04-24T19:11:53 | 2021-04-24T19:11:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 826,163,852 | 23,600 | alexey-milovidov | @filimonov Please check how does it help for #22113 | 2021-04-24T22:55:44 | 2021-04-24T22:55:44 | {
"confused": 1
} | MEMBER |
ClickHouse | ClickHouse | 825,511,864 | 23,455 | kitaisreal | @olgarev Thanks. Updated https://github.com/ClickHouse/ClickHouse/pull/23543. | 2021-04-23T08:58:11 | 2021-04-23T08:58:11 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 825,698,556 | 23,549 | vdimir | We are skipping such columns in `AddedColumns`, but not in `joinTotals`
https://github.com/ClickHouse/ClickHouse/blob/88334e77cdf866d1d9a05f53e154e6dd167c3876/src/Interpreters/HashJoin.cpp#L708-L710 | 2021-04-23T14:29:32 | 2021-04-23T14:29:32 | {} | MEMBER |
ClickHouse | ClickHouse | 825,802,433 | 21,893 | traceon | This is problematic, as a different configuration will be required, with a stored bind user password in plain text, a separate search filter for locating login user DN, and then an ability to manually encode and match passwords to hashes stored in LDAP. This is overly complicated, unsafe, and overall not very viable se... | 2021-04-23T17:20:44 | 2021-04-23T17:20:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 826,005,251 | 23,412 | alexey-milovidov | Need help from @filimonov to fix flaky test. | 2021-04-24T00:35:48 | 2021-04-24T00:35:48 | {} | MEMBER |
ClickHouse | ClickHouse | 826,021,496 | 23,310 | alexey-milovidov | Reference: #16253 | 2021-04-24T02:37:29 | 2021-04-24T02:37:29 | {} | MEMBER |
ClickHouse | ClickHouse | 826,122,150 | 23,517 | azat | >[AST fuzzer (debug) — Fuzzer failed (1). See the logs.](https://clickhouse-test-reports.s3.yandex.net/23517/f1187aeb69109c88f0be978b8083080c7a843820/fuzzer_debug/report.html#fail1)
#23595 | 2021-04-24T17:01:16 | 2021-04-24T17:01:16 | {} | MEMBER |
ClickHouse | ClickHouse | 825,468,392 | 22,477 | akuzm | Postgres disallows both of these frames as well:
```
postgres=# select count(*) over (rows between 0 following and current row);
ERROR: frame starting from following row cannot have preceding rows
LINE 1: ...elect count(*) over (rows between 0 following and current ro...
... | 2021-04-23T07:51:02 | 2021-04-23T07:52:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 825,538,593 | 8,521 | laowangriji | Hello, I have the same problem, is this question resolved? | 2021-04-23T09:44:02 | 2021-04-23T09:44:02 | {} | NONE |
ClickHouse | ClickHouse | 825,745,724 | 20,964 | filimonov | We saw that issue few times in different scenarios recently.
@EvgenyVinogradov can you also try to disable `optimize_on_insert` ? | 2021-04-23T15:42:41 | 2021-04-23T15:42:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 825,880,217 | 21,893 | traceon | > I am unable to change my ldap server to allow for anonymous binding
You don't have to allow anonymous binding (you need to actually forbid it). Instead, you should configure the LDAP server in a way it actually able to authenticate the registered users by allowing them to bind using their credentials. | 2021-04-23T19:44:21 | 2021-04-23T19:44:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 826,005,698 | 23,314 | mergify[bot] | **Command `update`: failure**
> **Pull request can't be updated with latest base branch changes**
> Mergify needs the author permission to update the base branch of the pull request.
> excitoon-favorites needs to [authorize modification on its head branch](https://help.github.com/articles/allowing-changes-to-a-pull-req... | 2021-04-24T00:38:37 | 2021-04-24T00:38:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 826,018,003 | 23,559 | alexey-milovidov | @filimonov
> Or just check diffoscope report in the issue.
Could you please just enumerate the differences that are still needed to be fixed? | 2021-04-24T02:06:54 | 2021-04-24T02:06:54 | {} | MEMBER |
ClickHouse | ClickHouse | 826,158,997 | 23,555 | genzgd | I will capture these values the next time it happens (we can't run our production servers in a state where the replication queue has grown to 150k entries). The system.stack_trace query currently returns more than 1000 entries; if there's a way to limit the query to get what you're looking for that would be helpful.
... | 2021-04-24T22:01:19 | 2021-04-24T22:01:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 826,165,461 | 23,046 | alexey-milovidov | We don't use `pbuilder`. See `docker/packager/deb` instead.
I will remove irrelevant part of `release` script in #23605. | 2021-04-24T23:14:26 | 2021-04-24T23:14:26 | {} | MEMBER |
ClickHouse | ClickHouse | 825,382,510 | 23,508 | PhantomPhreak | > Is there any stacktrace for `Cannot read all data` exception in server logs?
Sorry, forgot to add it. Added it now by collapsible block | 2021-04-23T04:46:49 | 2021-04-23T04:46:49 | {} | NONE |
ClickHouse | ClickHouse | 825,447,517 | 23,540 | akuzm | previous comparison against 20.3 is here https://github.com/ClickHouse/ClickHouse/pull/16919 | 2021-04-23T07:16:18 | 2021-04-23T07:16:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 825,627,965 | 23,541 | tavplubix | Is there any other way to consistently dump all data from MySQL without using `FLUSH TABLES WITH READ LOCK`? | 2021-04-23T12:36:03 | 2021-04-23T12:36:03 | {} | MEMBER |
ClickHouse | ClickHouse | 825,689,415 | 21,164 | madm1keru | @alexey-milovidov looks like it broke CLEAR COLUMN feature too. | 2021-04-23T14:15:36 | 2021-04-23T14:15:36 | {} | NONE |
ClickHouse | ClickHouse | 825,703,134 | 23,517 | alexey-milovidov | Ok. Now `01675_distributed_bytes_to_delay_insert` is flaky. | 2021-04-23T14:36:53 | 2021-04-23T14:36:53 | {} | MEMBER |
ClickHouse | ClickHouse | 826,014,393 | 23,312 | alexey-milovidov | @kitaisreal said that you are planning to implement a different fix.
But this is already good enough to merge. Let's merge it and then change to a different solution. | 2021-04-24T01:40:39 | 2021-04-24T01:40:39 | {} | MEMBER |
ClickHouse | ClickHouse | 826,134,404 | 23,559 | alexey-milovidov | > Otherwise you can not guarantee that the order of files (and order of their processing) will be the same on different machines.
What exactly do you mean? | 2021-04-24T18:26:26 | 2021-04-24T18:26:26 | {} | MEMBER |
ClickHouse | ClickHouse | 826,140,125 | 23,517 | alexey-milovidov | @mergifyio update | 2021-04-24T19:11:41 | 2021-04-24T19:11:41 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 825,349,602 | 23,423 | GeraltXLi | Thank you very much for your reply! Especially this document.
I have created the second mv from the source table of first mv and it worked. | 2021-04-23T02:53:50 | 2021-04-23T02:53:50 | {} | NONE |
ClickHouse | ClickHouse | 825,374,648 | 23,413 | TCeason | Please review this pr.
Thanks. @tavplubix | 2021-04-23T04:21:00 | 2021-04-23T04:21:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 825,426,793 | 23,062 | marising | @tavplubix do you have a time to review this pr?
I saw the test exception of Functional stateful tests and Integration tests has nothing to do with my changes. | 2021-04-23T06:37:54 | 2021-04-23T06:37:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 825,870,501 | 21,893 | rlazimi-dev | You bring up a good point. However, I do think it will be a good feature, as I am unable to change my ldap server to allow for anonymous binding.
Although I mentioned keeping the password in the configuration in the example, I think a more secure way to implement this is to ask for bind credentials alongside the nor... | 2021-04-23T19:24:04 | 2021-04-23T19:41:25 | {} | NONE |
ClickHouse | ClickHouse | 826,147,353 | 22,963 | filimonov | > If we will use CPack it will prevent switching to Bazel.
Why? I think it's possible to use them together. | 2021-04-24T20:12:45 | 2021-04-24T20:12:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 826,166,730 | 23,553 | alexey-milovidov | > test_materialize_mysql_database
#23462 | 2021-04-24T23:30:31 | 2021-04-24T23:30:31 | {} | MEMBER |
ClickHouse | ClickHouse | 825,538,286 | 8,521 | laowangriji | 1 | 2021-04-23T09:43:30 | 2021-04-23T09:43:30 | {} | NONE |
ClickHouse | ClickHouse | 825,675,501 | 23,503 | Algunenano | > Seems like all of mutations are. Multiple mutation commands are joined into one operation and it may produce too deep execution pipeline.
First of all, thanks a lot for the information.
When I look into it in more detail I can see that it's queueing ~ 20k mutations:
```
>>> StorageMergeTree::tryMutatePart()... | 2021-04-23T13:54:51 | 2021-04-23T13:54:51 | {} | MEMBER |
ClickHouse | ClickHouse | 825,996,289 | 23,560 | l1t1 | groupArray() | 2021-04-23T23:51:25 | 2021-04-23T23:51:25 | {} | NONE |
ClickHouse | ClickHouse | 826,005,683 | 23,314 | alexey-milovidov | @Mergifyio update | 2021-04-24T00:38:31 | 2021-04-24T00:38:31 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 826,006,438 | 9,955 | alexey-milovidov | @tavplubix It does not help, see https://github.com/ClickHouse/ClickHouse/pull/23314. | 2021-04-24T00:43:21 | 2021-04-24T00:43:21 | {} | MEMBER |
ClickHouse | ClickHouse | 826,072,626 | 23,001 | tavplubix | #23508 | 2021-04-24T10:38:11 | 2021-04-24T10:38:11 | {} | MEMBER |
ClickHouse | ClickHouse | 826,166,838 | 23,517 | alexey-milovidov | @Mergifyio update | 2021-04-24T23:31:55 | 2021-04-24T23:31:55 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 825,464,463 | 23,504 | filimonov | > Why it's needed in production use cases?
Reason is very simple. When you configure fqdn during ubuntu / debian installation (debian-installer-launcher)it adds a line
```
127.0.1.1 myclickhouse1.host.name.com
```
to /etc/hosts (motivation is explained [here](https://lists.debian.org/debian-devel/2013/08/msg000... | 2021-04-23T07:44:40 | 2021-04-23T10:03:07 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 825,760,803 | 17,966 | nikitamikhaylov | Will be fixed here #23518 | 2021-04-23T16:06:50 | 2021-04-23T16:06:50 | {} | MEMBER |
ClickHouse | ClickHouse | 826,005,512 | 23,366 | alexey-milovidov | @Mergifyio update | 2021-04-24T00:37:26 | 2021-04-24T00:37:26 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 826,017,854 | 23,559 | alexey-milovidov | @filimonov Please post all the differences here. | 2021-04-24T02:06:01 | 2021-04-24T02:06:01 | {} | MEMBER |
ClickHouse | ClickHouse | 826,048,814 | 23,559 | filimonov | > @filimonov
>
> > Or just check diffoscope report in the issue.
>
> Could you please just enumerate the differences that are still needed to be fixed?
I can not. Diffoscope report showed hundreds of changes, analizing them is a task not for few minutes. Just open the html inside diffoscope archive attached to... | 2021-04-24T07:18:31 | 2021-04-24T07:18:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 826,163,823 | 23,600 | alexey-milovidov | > Stress test (thread) — Fatal message in clickhouse-server.log
A consequence of OOM:
```
2021.04.25 00:57:47.238894 [ 100001 ] {} <Fatal> BaseDaemon: ########################################
2021.04.25 00:57:47.239193 [ 100001 ] {} <Fatal> BaseDaemon: (version 21.6.1.6674, build id: 540E540865E50E8E0D37EAEA8980... | 2021-04-24T22:55:18 | 2021-04-24T22:55:18 | {} | MEMBER |
ClickHouse | ClickHouse | 825,346,798 | 23,423 | den-crane | https://den-crane.github.io/Everything_you_should_know_about_materialized_views_commented.pdf (page 18)
You should use:
```sql
FROM `.inner.test_scene_mv`
```
But only if database is not Atomic.
| 2021-04-23T02:44:28 | 2021-04-23T02:44:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 825,771,524 | 23,457 | KochetovNicolai | В целом все ок | 2021-04-23T16:25:04 | 2021-04-23T16:25:04 | {} | MEMBER |
ClickHouse | ClickHouse | 825,918,262 | 23,457 | sevirov | @KochetovNicolai Thank you for the review, i fixed. | 2021-04-23T20:50:20 | 2021-04-23T20:50:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 826,016,858 | 23,301 | alexey-milovidov | So, the block is reinserted and cleanup thread deleted it by mistake. | 2021-04-24T01:58:41 | 2021-04-24T01:58:41 | {} | MEMBER |
ClickHouse | ClickHouse | 826,138,844 | 22,963 | alexey-milovidov | If we will use CPack it will prevent switching to Bazel. | 2021-04-24T19:02:06 | 2021-04-24T19:02:06 | {} | MEMBER |
ClickHouse | ClickHouse | 824,485,055 | 23,412 | godliness | > Lgtm. But check tests
Thanks @filimonov,
Other failed tests are not related to this PR.
Reasons: https://github.com/ClickHouse/ClickHouse/pull/23415#issuecomment-824482273
`test_kafka_engine_put_errors_to_stream_with_random_malformed_json` test failed(timeout>300s)
![image](https://user-images.githubu... | 2021-04-22T02:10:13 | 2021-04-23T14:08:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 826,161,179 | 23,601 | alexey-milovidov | This CMakeLists.txt file was copy-pasted from the original library. But it is not intended to be used as part of another build system. | 2021-04-24T22:23:08 | 2021-04-24T22:23:08 | {} | MEMBER |
ClickHouse | ClickHouse | 825,424,891 | 23,431 | millin | The same issue with 21.3.6.55 when using GLOBAL IN/JOIN | 2021-04-23T06:33:41 | 2021-04-23T06:33:41 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 825,760,443 | 6,213 | lud4ik | https://altinity.com/blog/2020-07-14-joins-in-clickhouse-materialized-views
Here is described that they work
> materialized views trigger off the left-most table of the join. The materialized view will pull values from right-side tables in the join but will not trigger if those tables change
Is it true? I just di... | 2021-04-23T16:06:15 | 2021-04-23T16:06:27 | {
"+1": 4
} | NONE |
ClickHouse | ClickHouse | 825,923,941 | 23,504 | filimonov | > > I guess we can add a setting
>
> Adding a setting is too much.
Any better options if we want to have one behaviour for production usage, and another for tests ? (the setting can also live in remote_servers). | 2021-04-23T21:02:54 | 2021-04-23T21:02:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 825,998,829 | 23,560 | mauidude | I did see that but the first line in the docs says `Values can be added to the array in any (indeterminate) order.` which I need the columns from event D to match in order. Is there away to use Nested types with groupArray? It was kind of confusing to me because from what I understand, Nested types are arrays themselve... | 2021-04-24T00:03:53 | 2021-04-24T00:03:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 826,019,624 | 23,504 | alexey-milovidov | @filimonov ClickHouse works correctly regardless.
We can simply avoid implementing this feature.
And you will not often use the generated FQDN in configuration of your cluster.
For example, I have `127.0.1.1 github-explorer.ru-central1.internal github-explorer` on one of the machines, but the .internal address h... | 2021-04-24T02:21:15 | 2021-04-24T02:21:15 | {} | MEMBER |
ClickHouse | ClickHouse | 825,323,454 | 23,463 | yishanchuan | I found the "final" takes longer to execute. 90.472852839 sec is larger than before.
sql:
```
Select * From db.distributed_table final
Where auditstatus IN ('W','T')
AND finishstatus='F'
AND coalesce(generatesource,0)=0
```
trace log:
```
[SVR33471DE640] 2... | 2021-04-23T01:43:17 | 2021-04-24T08:27:15 | {} | NONE |
ClickHouse | ClickHouse | 825,765,310 | 21,893 | rlazimi-dev | Hi @traceon, I'd like to come back to this issue.
> ... The Bind DN will be constructed for each user by replacing {user_name} with user name. When rlazimi-dev is trying to authenticate, the following BindDN will be constructed: uid=rlazimi-dev,cn=users,cn=accounts,dc=test,dc=com.
My issue is that in my case, `rl... | 2021-04-23T16:14:38 | 2021-04-23T16:14:38 | {} | NONE |
ClickHouse | ClickHouse | 826,005,548 | 23,366 | mergify[bot] | **Command `update`: success**
> **Branch has been successfully updated**
>
| 2021-04-24T00:37:40 | 2021-04-24T00:37:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 825,508,703 | 23,455 | olgarev | @kitaisreal Hi! The info is now repeated twice in Russian?
Данные загружаются в словарь блоками по `initial_array_size` записей (значение по умолчанию — 1024). Начальный размер плоских массивов словарей контролируется параметром initial_array_size (по умолчанию - 1024). | 2021-04-23T08:52:51 | 2021-04-23T08:52:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 825,613,669 | 23,524 | KochetovNicolai | ```
2021.04.23 14:24:34.478656 [ 318595 ] {0a0f164d-6b44-4ca5-9cce-37941626b1c6} <Debug> executeQuery: (from [::1]:47336, using production parser) (comment: '/usr/share/clickhouse-test/queries/0_stateless/00992_syst
em_parts_race_condition_zookeeper_long.sh') INSERT INTO alter_table SELECT rand(1), rand(2), 1 / rand(... | 2021-04-23T12:08:56 | 2021-04-23T12:08:56 | {} | MEMBER |
ClickHouse | ClickHouse | 825,631,177 | 17,718 | vdimir | There are also other bug that columns from right table with `Join` engine not accessible if columns with same name presented in left table.
Like:
```
SELECT dd.pu_num, dd.s_amnt, k.name_rus, k.pu_num
from (
select a.pu_id, a.pu_num, sum(a.amnt) as s_amnt
from table1 a
group by a.pu_id, a.pu_num
... | 2021-04-23T12:41:53 | 2021-04-23T13:57:52 | {} | MEMBER |
ClickHouse | ClickHouse | 825,701,280 | 23,504 | alexey-milovidov | > I guess we can add a setting
Adding a setting is too much. | 2021-04-23T14:33:57 | 2021-04-23T14:33:57 | {} | MEMBER |
ClickHouse | ClickHouse | 826,049,243 | 22,113 | filimonov | > [diffoscope.tar.gz](https://github.com/ClickHouse/ClickHouse/files/6214430/diffoscope.tar.gz)
@alexey-milovidov check that file. Please. :)
> Some reading: https://reproducible-builds.org/_lfs/presentations/2015-08-13-CCCamp15.pdf
Also check tgat, it point all typical pitfalls and possible solutions. Also ht... | 2021-04-24T07:22:43 | 2021-04-24T07:30:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 826,051,880 | 23,463 | filimonov | You read twice more data... Check if the same conditions were pushed to prewhere. | 2021-04-24T07:44:31 | 2021-04-24T07:44:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 826,164,063 | 23,598 | alexey-milovidov | I'm not sure if it's always needed because I'm not an "expert" in CMake (did not even read the manual).
For example, how does it expand the unquoted parameter if variable has spaces:
```
statement(abc${PARAM}def)
```
? | 2021-04-24T22:58:22 | 2021-04-24T22:58:22 | {} | MEMBER |
ClickHouse | ClickHouse | 826,166,997 | 23,517 | mergify[bot] | **Command `update`: success**
> **Branch has been successfully updated**
>
| 2021-04-24T23:33:43 | 2021-04-24T23:33:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 825,430,513 | 23,414 | mjh568412187 | > > the username cannot be a pure number
>
> Yes, looks like, this is a limitation of XML config format.
>
> > do I need to configure all users in ldap one by one in users.xml?
>
> No, you can use LDAP as a user directory, see: https://clickhouse.tech/docs/en/operations/external-authenticators/ldap/#ldap-exter... | 2021-04-23T06:45:26 | 2021-04-23T06:45:26 | {} | NONE |
ClickHouse | ClickHouse | 825,481,998 | 22,351 | akuzm | I'll close this one as a duplicate and will continue in #22121 | 2021-04-23T08:12:04 | 2021-04-23T08:12:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 825,623,447 | 23,547 | tavplubix | It works as expected with `--input_format_with_names_use_header=0`, so it looks like duplicate of #22406, fixed in #23202 | 2021-04-23T12:27:42 | 2021-04-23T12:27:42 | {} | MEMBER |
ClickHouse | ClickHouse | 825,729,454 | 20,957 | nikitamikhaylov | Will be fixed in #23518 by enabling `<allow_to_drop_target_partitions>` to `true` in task configuration. | 2021-04-23T15:17:13 | 2021-04-23T15:17:13 | {} | MEMBER |
ClickHouse | ClickHouse | 825,775,988 | 21,067 | filimonov | Any update there? | 2021-04-23T16:32:58 | 2021-04-23T16:32:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 825,995,517 | 23,559 | alexey-milovidov | ```
milovidov@mtlog-perftest03j:~/ClickHouse/build$ md5sum programs/clickhouse
f55e2081efcffb8e6ac224942d2c4060 programs/clickhouse
milovidov@mtlog-perftest03j:~/ClickHouse/build$ md5sum programs/clickhouse1
f55e2081efcffb8e6ac224942d2c4060 programs/clickhouse1
``` | 2021-04-23T23:47:46 | 2021-04-23T23:47:46 | {} | MEMBER |
ClickHouse | ClickHouse | 825,996,471 | 22,113 | alexey-milovidov | The only offender was RocksDB. | 2021-04-23T23:52:17 | 2021-04-23T23:52:17 | {
"-1": 1
} | MEMBER |
ClickHouse | ClickHouse | 826,015,247 | 21,474 | alexey-milovidov | Ok, but the whole idea of the method "with retries" looks incorrect:
```
/// Network errors and similar. You should reinitialize ZooKeeper session in case of these errors
bool isHardwareError(Error code);
```
Retries never help. It looks like an outdated code.
There are no comments before these methods:
``... | 2021-04-24T01:47:44 | 2021-04-24T01:47:44 | {} | MEMBER |
ClickHouse | ClickHouse | 825,350,027 | 23,425 | GeraltXLi | Thank you very much for your reply!
I will try it later. | 2021-04-23T02:54:52 | 2021-04-23T02:54:52 | {} | NONE |
ClickHouse | ClickHouse | 825,537,250 | 23,516 | filimonov | > When new key will appear, new subcolumn with all defaults will be added to ColumnObject. After all rows will be parsed, types of every subcolumns will be optimized to lowest common for all values.
Please check also whitepapers of the Dremel (it's used by Google BigQuery): https://storage.googleapis.com/pub-tools-p... | 2021-04-23T09:41:27 | 2021-04-23T09:41:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 822,514,409 | 23,199 | taiwhite | @Jokser
There is only 1 part allocated for the `s3` table in both cases. ~~That is correct, `NOT ignore(*)` should simply evaluate to 1 for all rows.~~ See comment below for correct behavior.
| 2021-04-19T14:32:59 | 2021-04-23T14:30:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 825,736,969 | 23,199 | taiwhite | @Jokser
[bug.zip](https://github.com/ClickHouse/ClickHouse/files/6366387/bug.zip)
I ran my tests again and it seems like this is not only a problem with S3 tables, I had very long time with a local table as well. The zip file contains logs for running the query with 10k rows on local and s3 tables, and logs for ru... | 2021-04-23T15:28:41 | 2021-04-23T15:28:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 825,939,554 | 23,555 | filimonov | Please run
```
SET allow_introspection_functions = 1;
WITH arrayMap(x -> demangle(addressToSymbol(x)), trace) AS all SELECT thread_id, query_id, arrayStringConcat(all, '\n') AS res FROM system.stack_trace WHERE res ILIKE '%Pool%';
select name, value from system.settings where name like '%pool%'
select me... | 2021-04-23T21:40:52 | 2021-04-23T21:40:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 826,011,614 | 23,522 | alexey-milovidov | Not relevant for 21.3. | 2021-04-24T01:18:26 | 2021-04-24T01:18:26 | {} | MEMBER |
ClickHouse | ClickHouse | 826,057,015 | 23,463 | yishanchuan | @filimonov thanks very much. you're right. I make a mistake that zookeeper replicate the data to the machine in same shard.
In terms of replication mechanism, it seem that {macro.replica} config is more priority than {shard.replica} config in metrika.xml file. Oh my god... | 2021-04-24T08:24:49 | 2021-04-24T08:24:49 | {} | NONE |
ClickHouse | ClickHouse | 826,133,984 | 23,561 | alexey-milovidov | ALTER query hung in `zkutil::ZooKeeper::waitForDisappear` | 2021-04-24T18:23:39 | 2021-04-24T18:23:39 | {} | MEMBER |
ClickHouse | ClickHouse | 826,005,477 | 23,366 | alexey-milovidov | > test_storage_kafka/test.py::test_kafka_virtual_columns2
Need to investigate the failure of this test. | 2021-04-24T00:37:14 | 2021-04-24T00:37:14 | {} | MEMBER |
ClickHouse | ClickHouse | 826,150,584 | 15,065 | alexey-milovidov | @Slach the work on this task is just started. | 2021-04-24T20:41:48 | 2021-04-24T20:41:48 | {} | MEMBER |
ClickHouse | ClickHouse | 825,762,357 | 23,524 | KochetovNicolai | There is still a bug:
```
==157==ERROR: AddressSanitizer: heap-buffer-overflow on address 0x60604548b8d8 at pc 0x00001e2df247 bp 0x7f6d53fde9d0 sp 0x7f6d53fde9c8
READ of size 8 at 0x60604548b8d8 thread T1959 (TCPHandler)
#0 0x1e2df246 in boost::intrusive_ptr<DB::IColumn const>::operator->() const obj-x86_64-lin... | 2021-04-23T16:09:34 | 2021-04-23T16:09:34 | {} | MEMBER |
ClickHouse | ClickHouse | 826,012,104 | 23,511 | filimonov | Heh, i just had exactly the same request / discussion today :) | 2021-04-24T01:22:32 | 2021-04-24T01:22:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 826,018,242 | 23,559 | alexey-milovidov | > into another folder
Paths are still embedded to the binary and we are not going to fix it in this PR. | 2021-04-24T02:08:40 | 2021-04-24T02:08:40 | {} | MEMBER |
ClickHouse | ClickHouse | 826,021,653 | 23,561 | alexey-milovidov | SQLancer test: wrong report https://github.com/sqlancer/sqlancer/pull/340 | 2021-04-24T02:39:05 | 2021-04-24T02:39:05 | {} | MEMBER |
ClickHouse | ClickHouse | 826,147,879 | 23,561 | alexey-milovidov | Ok. The logic works. But the RSS in tests under TSan can grow quite quickly. | 2021-04-24T20:16:47 | 2021-04-24T20:16:47 | {} | MEMBER |
ClickHouse | ClickHouse | 816,574,952 | 22,927 | alexey-milovidov | This is how it is supposed to work. | 2021-04-09T10:05:54 | 2021-04-09T10:05:54 | {} | MEMBER |
ClickHouse | ClickHouse | 816,613,733 | 22,928 | zvonand | > Could you please add the documentation?
Of course, in a couple of hours | 2021-04-09T11:22:00 | 2021-04-09T11:22:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,709,882 | 17,769 | xjewer | >Ignore if setting was not set is Ok.
it will mask errors, i.e. if setting name has a typo. Though it will be the same behaviour with `IF EXISTS` statement. | 2021-04-09T14:09:13 | 2021-04-09T14:09:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,827,973 | 22,885 | abyss7 | > because cross-compiling for macos is far from reproducing the real environment - you can see that by looking at the trend of the reported issues for mac compilation.
That is unfortunate and unexpected in some way. The most of issues I saw relates to Apple Clang, which we don't support. Can you point out any example ... | 2021-04-09T17:11:56 | 2021-04-09T17:11:56 | {} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.