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 | 462,137,871 | 4,246 | k3box | Dear ClickHouse Team,
I apologise for the long lasting delay with fixing review issues.
I hope I will tackle them in coming days | 2019-02-10T14:35:17 | 2019-02-10T14:35:17 | {} | NONE |
ClickHouse | ClickHouse | 462,169,500 | 4,245 | alexey-milovidov | It doesn't work for generic types:
```
SELECT topKWeighted(toString(number), number) FROM numbers(10)
``` | 2019-02-10T20:34:21 | 2019-02-10T20:34:21 | {} | MEMBER |
ClickHouse | ClickHouse | 462,395,066 | 4,332 | ztlpn | For simplicity (comparison of NULLs requires special logic).
> Nevertheless, if you modify the code for allowing them, everything sems to be working.
Not really (I've removed the exception-throwing code in the `checkKeyExpression` function):
```
:) create table t (x Nullable(UInt32)) engine MergeTree order by x... | 2019-02-11T16:31:57 | 2019-02-11T16:34:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 462,476,692 | 4,350 | mcarbonneaux | ok fine !
i've created table with MySQL engine and created view on it...
there no way to do direct view on mysql function ?
like that:
```
CREATE VIEW clikhdb.clickhview
AS
SELECT
mysqlcolumn,
FROM mysql('<mysqlhost>:<mysqlport>','mymysqldbs', 'mymysqltable', 'mysqluser', 'mysqlpass')
GROUP BY
my... | 2019-02-11T20:12:28 | 2019-02-11T20:13:39 | {} | NONE |
ClickHouse | ClickHouse | 462,764,836 | 4,235 | alexey-milovidov | Ok. Now it is implemented for data insertion.
How hard is to implement `BrotliWriteBuffer` and to use it when `enable_http_compression` is enabled and client said that it accepts `br`? And probably to provide `http_brotli_compression_level` setting.
Do you want to implement it in next pull request? | 2019-02-12T13:46:15 | 2019-02-12T13:46:15 | {} | MEMBER |
ClickHouse | ClickHouse | 462,769,338 | 3,759 | alexey-milovidov | @andyyzh
This PR will be prepared for merging (most of commits will be squashed). Then the feature will be evaluated, PR will be merged to master and the next stable version of ClickHouse will be released. It may take a few weeks. | 2019-02-12T13:58:53 | 2019-02-12T13:58:53 | {
"+1": 2
} | MEMBER |
ClickHouse | ClickHouse | 462,881,699 | 4,110 | 4ertus2 | fixed in master | 2019-02-12T18:40:16 | 2019-02-12T18:40:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 462,994,597 | 3,995 | alexey-milovidov | This is fixed in #4369 | 2019-02-13T00:00:09 | 2019-02-13T00:00:09 | {} | MEMBER |
ClickHouse | ClickHouse | 462,141,334 | 4,246 | alexey-milovidov | It doesn't work at all:
```
SELECT number FROM numbers(10) LIMIT 0 + 1
``` | 2019-02-10T15:21:48 | 2019-02-10T15:21:48 | {} | MEMBER |
ClickHouse | ClickHouse | 462,361,416 | 3,291 | alexey-milovidov | Outdated.
Will be implemented in a separate PR. | 2019-02-11T15:08:00 | 2019-02-11T15:08:00 | {} | MEMBER |
ClickHouse | ClickHouse | 462,458,392 | 3,198 | harris-chiu | Have anyone faced this issue after new release?
Our clickhouse version is at 19.1.6 version.
However, the netlink file descriptors still keeps constantly increasing.
`lsof -p <pid> | wc -l
` | 2019-02-11T19:20:58 | 2019-02-11T19:21:08 | {} | NONE |
ClickHouse | ClickHouse | 462,584,003 | 3,198 | harris-chiu | A bit more info, our setup is with Kafka Engine with kafka_num_consumers = 4
with gdb setting break point at `TaskStatsInfoGetter::TaskStatsInfoGetter()` and `TaskStatsInfoGetter::~TaskStatsInfoGetter()`
We can see that for each materialized view interval, the constructor is called 4 times and no destructors is ca... | 2019-02-12T02:11:12 | 2019-02-12T02:11:12 | {} | NONE |
ClickHouse | ClickHouse | 462,715,846 | 2,922 | arctica | Seems like there must be again build issues because I just spent a good while figuring out why an install via the testing repo wouldn't support data skipping indexes as shown in the documentation. The feature seems to have been merged just a week ago and none of the last 2 -testing releases have debian packages built f... | 2019-02-12T11:01:22 | 2019-02-12T11:01:22 | {} | NONE |
ClickHouse | ClickHouse | 462,772,682 | 4,362 | hatarist | Weird. `toString` does work for me:
```
:) CREATE TABLE issue4362 (uuid UUID, string String) Engine = Memory();
:) INSERT INTO issue4362 SELECT generateUUIDv4() uuid, toString(uuid);
:) INSERT INTO issue4362 SELECT generateUUIDv4() uuid, toString(uuid);
:) INSERT INTO issue4362 SELECT generateUUIDv4() ... | 2019-02-12T14:08:24 | 2019-02-12T14:09:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 462,999,482 | 4,218 | alexey-milovidov | @4ertus2 Still reproduces in master. | 2019-02-13T00:18:33 | 2019-02-13T00:18:33 | {} | MEMBER |
ClickHouse | ClickHouse | 462,294,527 | 4,324 | BayoNet | Пулл оказался неудачным, опять не получается комментировать файлы, которые вставлены вместо символических ссылок. Закрываю. | 2019-02-11T11:23:11 | 2019-02-11T11:23:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 462,370,961 | 3,925 | alexey-milovidov | We have prepared this PR for merge and we have evaluated this feature, but the use case looks very narrow.
At the same time, there are too many additional things to consider about this feature (how does it work in distributed setup; how the client can commit read data; persistency options) and this feature is likely... | 2019-02-11T15:32:45 | 2019-02-11T15:32:45 | {} | MEMBER |
ClickHouse | ClickHouse | 462,437,445 | 3,764 | blinkov | > No response yet, so I'll ask again: what are the consequences of disabling `group_by_two_level`?
Should be mostly performance and scalability. | 2019-02-11T18:23:19 | 2019-02-11T18:23:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 462,490,572 | 4,350 | alexey-milovidov | Ok, I understand - the user should be allowed to do a query if a table function is used indirectly via view. | 2019-02-11T20:55:16 | 2019-02-11T20:55:16 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 462,468,852 | 4,017 | harris-chiu | Hi, we are still experiencing the increase of netlink socket overtime when using materialized view with version 19.1.6. Happy to provide more info if needed. Thanks
Edit:
Behavior still exists | 2019-02-11T19:49:44 | 2019-02-11T22:37:47 | {} | NONE |
ClickHouse | ClickHouse | 463,002,434 | 3,033 | alexey-milovidov | The issue is still relevant. You have to enable some tcp port for server to work. It's recommended to use some firewalled port that's only available from localhost. | 2019-02-13T00:30:23 | 2019-02-13T00:30:23 | {} | MEMBER |
ClickHouse | ClickHouse | 463,178,749 | 3,852 | bgranvea | > It is a very reasonable solution - allow to store states of simple aggregate functions as ordinary columns, just as in SummingMergeTree.
>
> (but it will break backwards compatibility of existing AggregatingMergeTree tables if we will do it transparently)
that's why I propose a new function `SimpleAggregateFunc... | 2019-02-13T12:18:46 | 2019-02-13T12:18:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 462,080,945 | 4,301 | alexey-milovidov | It's Ok. | 2019-02-09T21:28:25 | 2019-02-09T21:28:25 | {} | MEMBER |
ClickHouse | ClickHouse | 462,426,950 | 3,764 | reflection | No response yet, so I'll ask again: what are the consequences of disabling `group_by_two_level`? Since this issue is not fixed, it is the only workaround we have, and searching through ClickHouse documentation hasn't yielded anything. Thanks. | 2019-02-11T17:54:13 | 2019-02-11T17:54:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 460,237,833 | 4,218 | proller | Test with crash:
```
SET join_use_nulls = 1;
DROP TABLE IF EXISTS test.table1;
DROP TABLE IF EXISTS test.table2;
CREATE TABLE test.table1 ( id String ) ENGINE = Log;
CREATE TABLE test.table2 ( parent_id String ) ENGINE = Log;
insert into test.table1 values ('1');
SELECT
test.table2.parent_id = '',
... | 2019-02-04T12:49:35 | 2019-02-13T00:18:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 463,008,521 | 2,353 | alexey-milovidov | @4ertus2 JFYI | 2019-02-13T00:55:43 | 2019-02-13T00:55:43 | {} | MEMBER |
ClickHouse | ClickHouse | 463,206,291 | 3,852 | alexey-milovidov | Do you mean a new data type "SimpleAggregateFunction"? | 2019-02-13T13:51:34 | 2019-02-13T13:51:34 | {} | MEMBER |
ClickHouse | ClickHouse | 462,163,693 | 4,246 | alexey-milovidov | Merged. | 2019-02-10T19:30:35 | 2019-02-10T19:30:35 | {} | MEMBER |
ClickHouse | ClickHouse | 462,193,460 | 4,245 | alexey-milovidov | Merged. | 2019-02-10T23:52:42 | 2019-02-10T23:52:42 | {} | MEMBER |
ClickHouse | ClickHouse | 462,256,738 | 4,301 | alesapin | Performance test failure is expected. | 2019-02-11T09:03:20 | 2019-02-11T09:03:20 | {} | MEMBER |
ClickHouse | ClickHouse | 462,269,866 | 3,920 | alexey-milovidov | @alesapin If you substitute `test.hits` for `hashfile` it will work. | 2019-02-11T09:53:41 | 2019-02-11T09:53:41 | {} | MEMBER |
ClickHouse | ClickHouse | 462,670,338 | 4,357 | sdlcwangsong | > This feature has been added few days ago in #4287 and will be available in next release, stay tuned.
thank you, but those mutations has exist long time ago, Can I delete it directly from zookeeper? | 2019-02-12T08:46:27 | 2019-02-12T08:58:27 | {} | NONE |
ClickHouse | ClickHouse | 462,881,745 | 3,998 | 4ertus2 | fixed in master | 2019-02-12T18:40:23 | 2019-02-12T18:40:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 462,995,696 | 2,416 | simPod | Thanks for update, your work is appreacited! | 2019-02-13T00:04:07 | 2019-02-13T00:04:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 462,327,773 | 4,334 | blinkov | * Там был порядок от свежего к старому, а новые добавлены в конец
* Хорошо бы русскую версию тоже сразу проапдейтить https://github.com/yandex/ClickHouse/blob/master/docs/ru/security_changelog.md | 2019-02-11T13:31:11 | 2019-02-11T13:31:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 462,457,901 | 4,333 | alexey-milovidov | "0000-00-00 00:00:00" is a special value that indicates the global default value of `DateTime` data type.
Aggregate functions in ClickHouse return non-Nullable data type if were called for non-Nullable arguments (this contradicts the SQL standard and we can eventually implement an option to change this behaviour). T... | 2019-02-11T19:19:43 | 2019-02-11T19:20:45 | {
"+1": 1,
"confused": 1
} | MEMBER |
ClickHouse | ClickHouse | 462,686,581 | 4,332 | edonin | Ok,
In may case, 95% of the columns are nullable, and this limit is very restrictive.
Currently, null comparaisons seems to be well managed in the ordering of the select statement, by not in the KeyCondition of the merge tree. I will take a look.
Thanks | 2019-02-12T09:35:58 | 2019-02-12T09:35:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 462,746,791 | 4,353 | alesapin | OK, can be merged | 2019-02-12T12:49:25 | 2019-02-12T12:49:25 | {} | MEMBER |
ClickHouse | ClickHouse | 462,905,039 | 4,359 | alexey-milovidov | @den-crane Это поведение проявляется не только для констант:
`SELECT arrayJoin([x, y, z]), arrayJoin([x, y, z]) FROM (SELECT 1 AS x, materialize(2) AS y, 3 AS z)`
@banyrule Это документированное поведение: https://clickhouse.yandex/docs/ru/query_language/functions/#skleika-odinakovykh-vyrazhenii | 2019-02-12T19:41:54 | 2019-02-12T19:41:54 | {} | MEMBER |
ClickHouse | ClickHouse | 462,994,102 | 2,417 | alexey-milovidov | @dylanzr This issue is fixed in #4369 | 2019-02-12T23:58:28 | 2019-02-12T23:58:28 | {} | MEMBER |
ClickHouse | ClickHouse | 462,996,170 | 3,497 | alexey-milovidov | The "File not found" exception at server startup is fixed in #4369.
The issue may also happen if ClickHouse and you are trying to delete the same directory simultaneously.
It's not recommended to manually manipulate data inside ClickHouse directory.
> Can I directly delete those partition directories(i.e. 201810... | 2019-02-13T00:05:56 | 2019-02-13T00:07:03 | {} | MEMBER |
ClickHouse | ClickHouse | 462,075,784 | 4,309 | alexey-milovidov | Your own test has failed:
```
2019-02-08 17:21:10 -['qwerty','qwerty'] [] [1]
2019-02-08 17:21:10 +['qwerty','qwerty'] [] [1]
```
Because you have accidentially substituted tabs with spaces in test TSV output. | 2019-02-09T20:15:54 | 2019-02-09T20:16:04 | {} | MEMBER |
ClickHouse | ClickHouse | 462,347,713 | 1,201 | vitlibar | Implemented syntax `toStartOfInterval(time_or_data, INTERVAL x unit [, time_zone])` (see https://github.com/yandex/ClickHouse/pull/4304#issue-251152989 for examples). | 2019-02-11T14:31:47 | 2019-02-11T14:31:47 | {} | MEMBER |
ClickHouse | ClickHouse | 462,567,566 | 3,271 | iameugenejo | It's still happening with non replicated tables in version 18.16.1.
Any update on this?
Another table set we are seeing the same problem with has the following statement:
```
CREATE MATERIALIZED VIEW sample.log_stats_20181212 (
timestamp DateTime,
m_type String,
hits UInt64
) ENGINE = SummingM... | 2019-02-12T01:12:14 | 2019-02-12T01:15:21 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 462,711,331 | 4,353 | alesapin | Do not merge please. | 2019-02-12T10:47:38 | 2019-02-12T10:47:38 | {} | MEMBER |
ClickHouse | ClickHouse | 463,128,101 | 4,362 | andatt | Hi thanks!
I tried posting the queries above and I got this:
```
INSERT INTO issue4362 SELECT
generateUUIDv4() AS uuid,
toString(uuid)
Received exception from server (version 18.10.3):
Code: 43. DB::Exception: Received from localhost:9000, 127.0.0.1. DB::Exception: UInt128 cannot be write as a te... | 2019-02-13T09:38:53 | 2019-02-13T09:39:52 | {} | NONE |
ClickHouse | ClickHouse | 463,129,045 | 4,364 | blinkov | в PR приехали какие-то левые изменения | 2019-02-13T09:41:36 | 2019-02-13T09:41:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 408,068,642 | 1,306 | Felixoid | Из обсуждения с @alexey-milovidov в телеграме:
> Felixoid 🚜🇩🇪, [28.09.17 17:52]
> Собираюсь переводить один графит на кликхаус и столкнулся со следующей вещью: http://graphite.readthedocs.io/en/latest/config-carbon.html#storage-aggregation-conf
>
> В исходном хранилище данных графита, виспере, описываются 2 схе... | 2018-07-26T11:39:38 | 2019-02-13T16:27:21 | {} | MEMBER |
ClickHouse | ClickHouse | 462,090,178 | 4,198 | alexey-milovidov | Ok. | 2019-02-09T23:46:45 | 2019-02-09T23:50:46 | {} | MEMBER |
ClickHouse | ClickHouse | 462,168,730 | 4,228 | alexey-milovidov | Nothing magical - it was just incorrect code:
https://github.com/yandex/ClickHouse/pull/4228/files#diff-a03f181ffa68622a640f819e4386cc3cR275
You are trying to parse `a BETWEEN b AND c NOT` instead of `a NOT BETWEEN b AND c`. | 2019-02-10T20:25:55 | 2019-02-10T20:25:55 | {} | MEMBER |
ClickHouse | ClickHouse | 462,424,549 | 4,223 | reflection | Any update on this? Thanks | 2019-02-11T17:47:40 | 2019-02-11T17:47:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 462,668,489 | 4,357 | blinkov | This feature has been added few days ago in #4287 and will be available in next release, stay tuned. | 2019-02-12T08:40:43 | 2019-02-12T08:40:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 462,841,030 | 4,359 | den-crane | это баг и он проявляется только для констант, видимо как оптимизация алиасится где-то внутри
[ arrayJoin([1, 2, 3]) , arrayJoin([1, 2, 3]) ] -> [ a , a as b ] | 2019-02-12T16:55:52 | 2019-02-12T16:57:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 462,997,048 | 3,170 | alexey-milovidov | The issue has been fixed in master. | 2019-02-13T00:09:16 | 2019-02-13T00:09:16 | {} | MEMBER |
ClickHouse | ClickHouse | 463,005,172 | 3,531 | alexey-milovidov | > Question 1> Can I manually delete those directories 20181112_* since the system doesn't have knowledges about them and they are not in the detached directory.
You can manually delete the directory if DETACH was failed with `File not found` error and the partition does not exists in `system.parts`. It is safe becau... | 2019-02-13T00:42:13 | 2019-02-13T00:42:13 | {} | MEMBER |
ClickHouse | ClickHouse | 463,077,496 | 4,359 | makrushin-evgenii | @alexey-milovidov спасибо за ответ.
Тем не менее, такое поведение все еще похоже на баг, один из моих аргументов:
Вычисляемым _один раз_ значением функции `arrayJoin` является преобразование unfold.
> В запросе может быть использовано несколько функций arrayJoin. В этом случае, соответствующее преобразование де... | 2019-02-13T06:34:02 | 2019-02-13T06:34:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 463,159,448 | 3,852 | alexey-milovidov | > To mitigate 2) and 3), wouldn't it be possible to use the existing AggregatingMergeTree with a new function SimpleAggregationFunction so that stored state is the value itself with its real type? This could be used only in a few limited cases like any, anyLast, min, max.
It is a very reasonable solution - allow to ... | 2019-02-13T11:12:34 | 2019-02-13T11:12:34 | {} | MEMBER |
ClickHouse | ClickHouse | 462,258,814 | 3,920 | amosbird | @alesapin It's a data file shared by @alexey-milovidov
It has the following schema and I use it for hash testing
```
DESCRIBE TABLE hashfile
┌─name─────────────┬─type───┬─default_type─┬─default_expression─┬─comment_expression─┬─codec_expression─┐
│ MobilePhoneModel │ String │ │ ... | 2019-02-11T09:11:52 | 2019-02-11T09:11:52 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 462,376,778 | 4,346 | alexey-milovidov | Motivation:
Some users build ClickHouse from .tar.gz without submodules. And `pdqsort` is not available as a separate library in any OS. | 2019-02-11T15:47:08 | 2019-02-11T15:47:08 | {} | MEMBER |
ClickHouse | ClickHouse | 462,465,144 | 3,290 | 4ertus2 | #4351 | 2019-02-11T19:39:31 | 2019-02-11T19:39:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 462,667,696 | 4,357 | sdlcwangsong | Can I delete it directly from zookeeper? | 2019-02-12T08:38:08 | 2019-02-12T08:38:08 | {} | NONE |
ClickHouse | ClickHouse | 463,155,629 | 3,852 | alexey-milovidov | > Can I be sure when blocks are merged that rows with same (ts,id) are taken in their natural insertion order, so that anyLast really selects the last inserted row?
Yes, it is guaranteed for background merges. Data parts are selected for merge in contiguous intervals of insertion order and merge sorting merges iden... | 2019-02-13T10:59:57 | 2019-02-13T10:59:57 | {} | MEMBER |
ClickHouse | ClickHouse | 462,079,548 | 4,309 | alexey-milovidov | The implementation doesn't work at all:
```
SELECT arrayWithConstant(3, number)
FROM numbers(10)
Received exception from server (version 19.2.0):
Code: 9. DB::Exception: Received from localhost:9000, 127.0.0.1. DB::Exception: Size of offsets doesn't match size of column..
``` | 2019-02-09T21:08:21 | 2019-02-09T21:08:21 | {} | MEMBER |
ClickHouse | ClickHouse | 462,174,527 | 4,239 | alexey-milovidov | It should be added for all the factories (table functions, storage engines, etc.) | 2019-02-10T21:21:11 | 2019-02-10T21:21:11 | {} | MEMBER |
ClickHouse | ClickHouse | 462,338,210 | 3,770 | ztlpn | This error was fixed in #3945. The fix is available in 19.1.6. | 2019-02-11T14:03:44 | 2019-02-11T14:03:44 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 462,362,498 | 4,028 | alexey-milovidov | This can be continued. | 2019-02-11T15:10:51 | 2019-02-11T15:10:51 | {} | MEMBER |
ClickHouse | ClickHouse | 462,668,636 | 4,235 | fandyushin | its seems that I'm done | 2019-02-12T08:41:04 | 2019-02-12T08:41:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 462,997,769 | 3,170 | alexey-milovidov | I've added a test from your example. | 2019-02-13T00:12:16 | 2019-02-13T00:12:16 | {} | MEMBER |
ClickHouse | ClickHouse | 462,149,900 | 4,235 | alexey-milovidov | Looks good.
Need to add a test. | 2019-02-10T16:44:16 | 2019-02-10T16:44:16 | {} | MEMBER |
ClickHouse | ClickHouse | 462,373,380 | 4,241 | alexey-milovidov | It is Ok for a prototype (demo) but cannot be merged right now, because we don't see the motivation for this specific function.
PS. This pull request was created during the ClickHouse C++ Sprint event. | 2019-02-11T15:38:49 | 2019-02-11T15:40:45 | {} | MEMBER |
ClickHouse | ClickHouse | 462,569,620 | 3,198 | harris-chiu | The following has been observed, when ever the materialized view is trying to query the consumer(queue) table. It would increment the file descriptor count by 4. When there is no materialized view, there is no increment.
#4017 is in 19.1.6 release. However, this behavior still exists
2019.02.12 01:12:45.289371 [ 44... | 2019-02-12T01:15:02 | 2019-02-12T01:15:25 | {} | NONE |
ClickHouse | ClickHouse | 462,881,803 | 3,290 | 4ertus2 | fixed in master | 2019-02-12T18:40:31 | 2019-02-12T18:40:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 463,102,437 | 3,852 | bgranvea | I have to make more benchmarks but performance with `finalizeAggregation` looks interesting.
I see a couple of drawbacks with this approach however:
1) NULL values seem to be lost and replaced by a "default" value (0 or empty string). Is this a bug/limitation of `finalizeAggregation` that could be improved? or is... | 2019-02-13T08:19:14 | 2019-02-13T08:19:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 462,269,640 | 3,920 | alexey-milovidov | Selected columns from `test.hits`. | 2019-02-11T09:52:55 | 2019-02-11T09:52:55 | {} | MEMBER |
ClickHouse | ClickHouse | 462,447,824 | 3,764 | alexey-milovidov | > what are the consequences of disabling group_by_two_level
Severe performance degradation of GROUP BY with large result (merging will be done single-threaded). | 2019-02-11T18:51:44 | 2019-02-11T18:51:44 | {} | MEMBER |
ClickHouse | ClickHouse | 462,458,078 | 4,110 | 4ertus2 | #4351 | 2019-02-11T19:20:08 | 2019-02-11T19:20:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 462,463,014 | 3,998 | 4ertus2 | #4351 | 2019-02-11T19:33:35 | 2019-02-11T19:33:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 462,993,565 | 2,416 | alexey-milovidov | Still not for the first half of 2019. | 2019-02-12T23:56:44 | 2019-02-12T23:56:44 | {
"eyes": 5
} | MEMBER |
ClickHouse | ClickHouse | 463,136,298 | 4,207 | andyyzh | I seem to done it, Let's take a look at that failed test case. @alexey-milovidov | 2019-02-13T10:01:52 | 2019-02-13T10:01:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 462,074,027 | 4,238 | alexey-milovidov | It doesn't work for mutual entropy:
```
SELECT entropy(number % 2, number % 5), log2(10), log2(uniq(number % 2, number % 5)) FROM numbers(10)
``` | 2019-02-09T19:51:22 | 2019-02-09T19:51:22 | {} | MEMBER |
ClickHouse | ClickHouse | 462,350,681 | 4,312 | alesapin | https://github.com/yandex/ClickHouse/pull/4341 fix
Bug was not so terrible as I thought earlier. Reproduces only on dates more than 2037 year. | 2019-02-11T14:39:54 | 2019-02-11T14:40:15 | {} | MEMBER |
ClickHouse | ClickHouse | 462,797,564 | 2,508 | filimonov | Does https://github.com/yandex/ClickHouse/pull/3396 (and `kafka_max_block_size`) resolve that? | 2019-02-12T15:13:04 | 2019-02-12T15:59:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 463,006,734 | 2,375 | alexey-milovidov | @rpfennin It's per user setting. You can define it in `users.xml` file for a user profile. | 2019-02-13T00:48:54 | 2019-02-13T00:48:54 | {} | MEMBER |
ClickHouse | ClickHouse | 462,083,467 | 4,207 | alexey-milovidov | > 00834_cancel_http_readonly_queries_on_client_close_error_log.txt
It was poorly written test. Just merge with master and the test will be fixed. | 2019-02-09T22:05:05 | 2019-02-09T22:05:05 | {} | MEMBER |
ClickHouse | ClickHouse | 462,344,430 | 4,334 | ghost | > * Там был порядок от свежего к старому, а новые добавлены в конец
Исправлено.
> * Хорошо бы русскую версию тоже сразу проапдейтить
Добавим позже отдельным PR. | 2019-02-11T14:22:23 | 2019-02-11T14:22:23 | {} | NONE |
ClickHouse | ClickHouse | 462,058,078 | 4,302 | alexey-milovidov | Thank you!
Merged. | 2019-02-09T16:26:28 | 2019-02-09T16:26:28 | {} | MEMBER |
ClickHouse | ClickHouse | 462,316,659 | 4,316 | alexey-milovidov | There is fundamental issue with current implementation of table locks.
Nevertheless, we are going to fix it. | 2019-02-11T12:51:43 | 2019-02-11T12:51:43 | {} | MEMBER |
ClickHouse | ClickHouse | 462,730,044 | 4,357 | ztlpn | @sdlcwangsong Yes you can, it is safe. | 2019-02-12T11:49:34 | 2019-02-12T11:49:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 462,770,304 | 4,235 | fandyushin | I've fixed .gitmodules. It changes too often :)
I think its not complicated to implement write buffer. I can do it. | 2019-02-12T14:01:31 | 2019-02-12T14:01:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 463,001,274 | 3,953 | alexey-milovidov | You're right. Mutations are executed asynchronously. Moreover, if there are multiple mutations to be performed, they are coalesced and run together. This is the intended behaviour and should be clearly documented. | 2019-02-13T00:25:30 | 2019-02-13T00:25:30 | {} | MEMBER |
ClickHouse | ClickHouse | 463,003,487 | 2,731 | alexey-milovidov | @podshumok
ClickHouse generates `(a, b) IN ((a1, b1), (a2, b2), ...)` queries for MySQL and OR-AND-chains for generic ODBC sources, because we cannot be sure that `(a, b) IN ((a1, b1), (a2, b2), ...)` is going to work for particular DBMS and whether it will use index.
We need to have a way to determine if a "tup... | 2019-02-13T00:34:52 | 2019-02-13T00:35:15 | {} | MEMBER |
ClickHouse | ClickHouse | 463,006,384 | 2,396 | alexey-milovidov | `MODIFY PRIMARY KEY` was removed. We have added `MODIFY ORDER BY` instead:
https://clickhouse.yandex/docs/en/query_language/alter/#manipulations-with-key-expressions
But you still cannot have Nullable columns in your keys. | 2019-02-13T00:47:23 | 2019-02-13T00:47:23 | {} | MEMBER |
ClickHouse | ClickHouse | 462,047,193 | 4,207 | andyyzh | Two test cases :
00281_compile_sizeof_packed
00834_cancel_http_readonly_queries_on_client_close
reported errors that seemed unrelated to the code I had submitted,
I need your help, thanks. @alexey-milovidov
[00834_cancel_http_readonly_queries_on_client_close_error_log.txt](https://github.com/yandex/ClickHouse/... | 2019-02-09T14:03:52 | 2019-02-09T14:03:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 462,163,788 | 4,326 | alexey-milovidov | There are unrelated test failures. | 2019-02-10T19:31:32 | 2019-02-10T19:31:32 | {} | MEMBER |
ClickHouse | ClickHouse | 462,165,323 | 4,324 | blinkov | Они не **для** логирования, а по внутреннему устройству напоминают лог. По тексту ревью снова не комментируется... | 2019-02-10T19:45:52 | 2019-02-10T19:45:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 462,312,538 | 4,335 | BayoNet | С раздельными коммитами для удаления симлинков и последующим добавлением файлов тоже не получилось, Гитхаб не даёт комментить новые файлы. В общем, в случае замены симлинков на реальные файлы получается надо делать пулл с удаление симлинков и отдельный пулл с добаление новых файлов взамен.
Засим пулл закрываю. | 2019-02-11T12:35:58 | 2019-02-11T12:35:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 462,448,782 | 4,350 | alexey-milovidov | Yes. You can use `MySQL` table engine (`CREATE TABLE ... ENGINE = MySQL(...)`) instead of `mysql` table function.
Table function is just a way to dynamically create a table for single query. | 2019-02-11T18:54:26 | 2019-02-11T18:54:26 | {} | MEMBER |
ClickHouse | ClickHouse | 462,603,958 | 3,759 | andyyzh | When will this feature be released? | 2019-02-12T03:44:59 | 2019-02-12T03:44:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 462,998,606 | 4,061 | alexey-milovidov | You should look for the cause of `Resource temporarily unavailable` error in the server logs.
It may be the limit on number of threads. Does the server process other queries successfully? | 2019-02-13T00:15:22 | 2019-02-13T00:15:22 | {} | MEMBER |
ClickHouse | ClickHouse | 463,000,114 | 2,748 | alexey-milovidov | The issue is not relevant because we have migrated from OS OpenSSL package to LibreSSL from submodules. | 2019-02-13T00:20:57 | 2019-02-13T00:21:07 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.