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 | 495,988,963 | 5,244 | svladykin | Parameter type check works fine. | 2019-05-26T10:54:14 | 2019-05-26T10:54:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 496,174,138 | 5,376 | akuzm | This looks like some race condition when using extended query protocol in postgres, e.g. a Bind message arrives before Parse message. I'd expect UseServerSidePrepare=0 to disable extended protocol, but apparently it's still used. Will try to reproduce. | 2019-05-27T11:03:45 | 2019-05-27T11:03:45 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 496,284,034 | 5,394 | alexey-milovidov | `clickhouse-test-server` script is not used in CI.
It is @proller's own script. | 2019-05-27T18:45:56 | 2019-05-27T18:45:56 | {} | MEMBER |
ClickHouse | ClickHouse | 496,484,421 | 5,085 | abyss7 | The problem is solved in v19.8+ - the current solution only allows to start server with broken Kafka tables (they won't retry to connect to Kafka brokers again after failure on start) and re-create/remove Kafka table if required. | 2019-05-28T11:52:07 | 2019-05-28T11:53:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 496,777,822 | 5,453 | alexey-milovidov | You have to use `--data-binary` instead of `-d` for curl.
Otherwise curl will mangle binary data. | 2019-05-29T04:20:05 | 2019-05-29T04:20:05 | {} | MEMBER |
ClickHouse | ClickHouse | 497,094,502 | 5,472 | alexey-milovidov | You can use TTL expressions for this task. Look at the docs. | 2019-05-29T20:21:34 | 2019-05-29T20:21:34 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 497,246,293 | 5,460 | gcw831 | I mean I do not set password for clickhouse user <default>, but
then I find I can login with the password same as ubuntu shell login password, so I am confused.
 from (
select 'key1' as key, '' as value
union all
select 'key2' as key, toNullable('') as v... | 2019-05-30T13:33:17 | 2019-05-30T13:33:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 497,354,963 | 5,485 | den-crane | ```
SELECT t.1 AS k,t.2 AS v
FROM(SELECT arrayJoin([('a', 1), ('a', 2), ('b', 1), ('a', 3), ('b', 2), ('b', 100)]) AS t)
┌─k─┬───v─┐
│ a │ 1 │
│ a │ 2 │
│ b │ 1 │
│ a │ 3 │
│ b │ 2 │
│ b │ 100 │
└───┴─────┘
SELECT
k,
arrayDifference(groupArray(v))
FROM
(
SELECT
t.1 ... | 2019-05-30T14:43:27 | 2019-05-30T14:43:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 497,542,042 | 4,119 | mcarbonneaux | https://github.com/yandex/ClickHouse/pull/4212
It's the ttl expression? | 2019-05-31T01:27:42 | 2019-05-31T01:27:42 | {} | NONE |
ClickHouse | ClickHouse | 496,233,844 | 5,412 | filimonov | > do one batch insert and about ten updates per second about one hour
`ALTER ... UPDATE` has a special ALTER syntax, to highlight that is not designed to be used frequently,
it's asynchronous and quite an expensive operation, It really needs to reprocess all the data.
Please check what is actually going on b... | 2019-05-27T14:38:40 | 2019-05-27T14:59:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 497,409,876 | 5,488 | ygrek | btw current farmHash64 function actually (luckily) evaluates to Fingerprint64 because contrib/libfarmhash is not compiled with SSE enabled afaics and libfarmhash uses same implementation for Fingerprint64 and Hash64 so this change will be not noticeable in practice, would be nice to set this in stone and use Fingerprin... | 2019-05-30T17:19:31 | 2019-05-30T17:19:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 496,452,989 | 2,416 | gcw831 | @derekperkins Could you please show me the issue or link proving you words about ClickHouse only actually runs it once, if you have identical subqueries? Thank you | 2019-05-28T10:07:34 | 2019-05-28T10:07:34 | {} | NONE |
ClickHouse | ClickHouse | 496,508,954 | 5,441 | den-crane | @filimonov you mean https://github.com/yandex/ClickHouse/issues/2290 | 2019-05-28T13:08:29 | 2019-05-28T13:08:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 496,529,247 | 5,441 | filimonov | Yes, bad copy&paste :) Thanks. | 2019-05-28T14:00:34 | 2019-05-28T14:00:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 496,766,693 | 4,518 | alexey-milovidov | I've created another issue. | 2019-05-29T03:11:24 | 2019-05-29T03:11:24 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 496,886,533 | 5,458 | yingfeng | It seems if `hasany` could support bitmap, then such issue is resolved ? For example:
`select count(*) from orders where hasany((select groupbitmapstate(id) from users),user_id);`
Additionally, I found an extra issue for `groupbitmapstate`:
`select groupbitmapstate(id) from users` will consume memory continuousl... | 2019-05-29T10:50:49 | 2019-05-29T10:50:49 | {} | NONE |
ClickHouse | ClickHouse | 497,224,210 | 5,475 | alexey-milovidov | It depends on the size of the groupBitmap state.
No, the only memory required is for compressed bitmap. | 2019-05-30T06:46:27 | 2019-05-30T06:47:11 | {} | MEMBER |
ClickHouse | ClickHouse | 495,918,240 | 5,299 | alexey-milovidov | Now it's ok to not mention arrays at all. | 2019-05-25T13:23:30 | 2019-05-25T13:23:30 | {} | MEMBER |
ClickHouse | ClickHouse | 495,920,988 | 5,331 | sanych73 | > Missing tests.
Added functional tests. | 2019-05-25T14:05:21 | 2019-05-25T14:05:21 | {} | NONE |
ClickHouse | ClickHouse | 494,700,693 | 4,631 | vezir | Hi, the error occured again.
the logs are below;
```May 22 11:09:34 bigdata22 clickhouse-server[31808]: %5|1558512574.943|REQTMOUT|ClickHouse 19.6.2.11#consumer-15| [thrd:GroupCoordinator]: GroupCoordinator/8: Timed out LeaveGroupRequest in flight (after 5220ms, timeout #0): possibly held back by preceeding block... | 2019-05-22T08:12:30 | 2019-05-28T13:27:12 | {} | NONE |
ClickHouse | ClickHouse | 496,507,713 | 5,441 | filimonov | Similar: #2290 | 2019-05-28T13:04:57 | 2019-05-28T14:00:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 496,673,627 | 4,203 | VitalyNeganov | @den-crane Can somebody explain deduplication mechanism in more details?
Does it mean that Clickhouse provide "at most once" semantics on write in ReplicatedMergeTree?
Scenario:
1. Client sends batch of 100 rows for insert.
2. Batch Insert fails on server due to unexpected error, but 50 rows were inserted success... | 2019-05-28T20:20:01 | 2019-05-28T20:20:01 | {} | NONE |
ClickHouse | ClickHouse | 496,758,922 | 4,518 | jangorecki | @alexey-milovidov any idea if `SYSTEM FLUSH LOGS` now flushes both queues?
The issue has been closed not being referred by any commit/PR, just comment linking new documentation page. On that page I can only find:
> To flush the logs forcibly from the memory buffer into the table, use the SYSTEM FLUSH LOGS query.
T... | 2019-05-29T02:27:03 | 2019-05-29T02:39:52 | {} | NONE |
ClickHouse | ClickHouse | 496,890,834 | 5,394 | alesapin | Temporary we can add this `.xml` files to docker images. After DDL for dictionaries will be merged we can write functional tests for dictionaries in simple way. | 2019-05-29T11:06:05 | 2019-05-29T11:06:05 | {} | MEMBER |
ClickHouse | ClickHouse | 496,958,580 | 5,385 | proller | Will be fixed here -
https://github.com/yandex/ClickHouse/pull/5447/files#diff-c8b08341f4ec6c1c503955b6ec6cc23fR58 | 2019-05-29T14:22:48 | 2019-05-29T14:22:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 497,444,352 | 5,429 | azat | Any thoughts on this? | 2019-05-30T18:55:40 | 2019-05-30T18:55:40 | {} | MEMBER |
ClickHouse | ClickHouse | 496,317,936 | 4,064 | den-crane | MVs never supported select from select | 2019-05-27T23:02:52 | 2019-05-27T23:02:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 497,274,795 | 5,473 | 4ertus2 | There's no crash on current master (debug). Trying to reproduce.
```
Received exception from server (version 19.9.1):
Code: 171. DB::Exception: Received from localhost:9000, 127.0.0.1. DB::Exception: Block structure mismatch in Converting stream: different columns:
key String String(size = 1), value Nullable(String... | 2019-05-30T09:55:35 | 2019-05-30T10:21:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 496,055,071 | 910 | ghost | @Negashev
Did you do force_restore_data to solve the problem, right?
```
sudo -u clickhouse touch /var/lib/clickhouse/flags/force_restore_data
```
Has any data been lost or was it completely recovered? | 2019-05-27T02:11:47 | 2019-05-27T02:11:56 | {} | NONE |
ClickHouse | ClickHouse | 496,607,148 | 5,443 | alexey-milovidov | Notes:
- when TZ env. var is empty, we will still find name by path search;
- in most cases, TZ env var is empty - so this PR doesn't change much;
- in previous versions of tzdata, some files were symlinks too. Example: Europe/Moscow is the same as W-SU. And the timezone is displayed in ClickHouse as W-SU. The behav... | 2019-05-28T17:14:31 | 2019-05-28T17:14:31 | {} | MEMBER |
ClickHouse | ClickHouse | 496,625,216 | 5,443 | abyss7 | > Notes:
>
> * when TZ env. var is empty, we will still find name by path search;
> * in most cases, TZ env var is empty - so this PR doesn't change much;
> * in previous versions of tzdata, some files were symlinks too. Example: Europe/Moscow is the same as W-SU. And the timezone is displayed in ClickHouse as W-S... | 2019-05-28T18:02:45 | 2019-05-28T18:02:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 496,770,006 | 4,935 | zhang2014 | This is the plan I gave him(translated from google translation):
```
For example, allow a delay between 20s for different regions:
1. Configure distributed_directory_monitor_batch_inserts = 1, load_balancing=IN_ORDER (users.xml)
2. Create distributed, corresponding to the China, Japanese, and United States ... | 2019-05-29T03:30:48 | 2019-05-29T03:30:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 497,246,398 | 5,460 | gcw831 | @filimonov as above | 2019-05-30T08:14:45 | 2019-05-30T08:14:45 | {} | NONE |
ClickHouse | ClickHouse | 497,248,433 | 5,208 | ghuname | Thanks for an explanation. I will wait for v19.8 in order to check this subject again. | 2019-05-30T08:22:09 | 2019-05-30T08:22:09 | {} | NONE |
ClickHouse | ClickHouse | 496,002,992 | 5,042 | anrodigina | Падают тесты с названием [, т.е. все ок | 2019-05-26T14:02:36 | 2019-05-26T14:02:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 496,021,678 | 5,352 | alexey-milovidov | > Print error messages to stderr.
It isn't consistent with CI scripts. Will revert. | 2019-05-26T18:34:27 | 2019-05-26T18:34:27 | {} | MEMBER |
ClickHouse | ClickHouse | 496,264,790 | 5,428 | filimonov | We should remember to put it back again when it will be released with the proper name. | 2019-05-27T16:49:03 | 2019-05-27T16:49:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 496,928,545 | 5,286 | inv2004 | Please do not forget _timestamp | 2019-05-29T13:07:13 | 2019-05-29T13:07:13 | {
"+1": 4
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 497,207,984 | 5,475 | yingfeng | @alexey-milovidov
So it means if the result set of subquery is large, in the above scenario, if the subquery is changed to `select groupbitmapstate(id) from trips`, although I use bitmap to store ids which is a kind of compressed format, the practical memory consumption is still high? | 2019-05-30T05:28:16 | 2019-05-30T05:28:16 | {} | NONE |
ClickHouse | ClickHouse | 496,292,502 | 5,431 | alexey-milovidov | This should be mentioned in changelog. | 2019-05-27T19:38:51 | 2019-05-27T19:38:51 | {} | MEMBER |
ClickHouse | ClickHouse | 496,510,829 | 5,435 | filimonov | > I found out that it's currently impossible to change the query setting `allow_simdjson=0` when it's written in the `SETTINGS` part after `SELECT`. So the following doesn't work:
>
> I tried to fix changing query settings with `SETTINGS`, but it seems not to be an easy task and I think it's better to do it in a sep... | 2019-05-28T13:13:42 | 2019-05-28T13:13:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 496,600,885 | 5,446 | alexey-milovidov | Ok.
BTW, did you manage to lower the number of template instantiations?
It was N^2 (input types * output types), should be just N. This will speed up build time. | 2019-05-28T16:56:51 | 2019-05-28T16:56:51 | {} | MEMBER |
ClickHouse | ClickHouse | 496,607,292 | 5,443 | alexey-milovidov | PS. Cannot figure out, what bug was fixed? | 2019-05-28T17:15:01 | 2019-05-28T17:15:01 | {} | MEMBER |
ClickHouse | ClickHouse | 497,187,977 | 4,207 | andyyzh | Currently, it can be converted to array to filter.
> DROP TABLE IF EXISTS bitmap_test;
CREATE TABLE bitmap_test(pickup_date Date, city_id UInt32, uid UInt32)ENGINE = Memory;
INSERT INTO bitmap_test SELECT '2019-01-01', 1, number FROM numbers(1,5);
INSERT INTO bitmap_test SELECT '2019-01-02', 2, number FROM number... | 2019-05-30T03:25:07 | 2019-05-30T03:25:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 497,280,735 | 5,460 | filimonov | Looks like you have some non-empty password for `default` user.
Please examine / show the contents `/etc/clickhouse-server/users.xml` and files in `/etc/clickhouse-server/users.d` subfolder.
OR may be you have some default password in your `clickhouse-client` configuration. | 2019-05-30T10:16:24 | 2019-05-30T10:17:47 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 497,311,662 | 5,370 | ajax16384 | I am not gentoo ebuild maintainer. Stable release creation is not gentoo specific issue. I think any v*-stable.zip archive should contain appropriate version string, so version.cmake commit like version tagging is just another preparation before release (there is no uglification). | 2019-05-30T12:24:45 | 2019-05-30T12:24:45 | {} | NONE |
ClickHouse | ClickHouse | 496,486,509 | 5,435 | vitlibar | > Ok.
>
> But I don't understand how the changes in perf test help.
I found out that it's currently impossible to change the query setting `allow_simdjson=0` when it's written in the `SETTINGS` part after `SELECT`. So the following doesn't work:
```
SELECT JSONExtractUInt('{"a", 5}', 'a') SETTINGS allow_simdj... | 2019-05-28T11:59:14 | 2019-05-28T12:00:10 | {} | MEMBER |
ClickHouse | ClickHouse | 496,766,024 | 4,518 | alexey-milovidov | > any idea if SYSTEM FLUSH LOGS now flushes both queues
No, the issue was not fixed. The race condition still exist. This is to do. | 2019-05-29T03:07:37 | 2019-05-29T03:07:37 | {} | MEMBER |
ClickHouse | ClickHouse | 457,770,715 | 4,137 | healiseu | @alexey-milovidov the table (2,780,185 x 41) was loaded on TriaClick (i.e. ClickHouse Associative Semiotic Hypergraph engine) in approximately **63sec**, including optimizations, almost the same time like qlikview ! The size of **TriaClick table engines on the disk was 507MB** ( that's about twice the size of qlikview... | 2019-01-25T23:31:59 | 2019-05-29T11:50:16 | {} | NONE |
ClickHouse | ClickHouse | 497,368,957 | 3,885 | den-crane | @abyss7
clickhouse-client --enable_optimize_predicate_expression=1
```
select count()
from (
select a, runningDifference(a) rd
from
( select *
from (
select 1 a
union all
select 2 a
)
order by a desc
)
Where rd=-1
)
0 rows in set.
```
clickhouse-... | 2019-05-30T15:21:34 | 2019-05-30T15:21:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 495,917,268 | 5,285 | alexey-milovidov | Questionable. Still Ok to merge because you know better. | 2019-05-25T13:09:21 | 2019-05-25T13:09:21 | {} | MEMBER |
ClickHouse | ClickHouse | 495,986,038 | 5,279 | svladykin | @alexey-milovidov @andyyzh
I've fixed the array element type related issues in tests, fixed the code style.
I can not reproduce the crash now:
```
$ clickhouse-dev client -n --testmode < 00829_bitmap_function.sql > 00829_bitmap_function.result
$ diff 00829_bitmap_function.result 00829_bitmap_function.reference
... | 2019-05-26T10:10:06 | 2019-05-26T10:10:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 496,632,208 | 1,148 | 4ertus2 | As workaround you can use hex(murmurHash3_128(reinterpretAsString(number))) | 2019-05-28T18:22:01 | 2019-05-28T18:22:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 496,923,829 | 5,453 | georgantasp | My mistake. I thought `-d` was the proper alias. Thanks for the fast response! | 2019-05-29T12:54:07 | 2019-05-29T12:54:07 | {} | NONE |
ClickHouse | ClickHouse | 497,066,303 | 5,090 | proller | https://github.com/yandex/ClickHouse/pull/5471 | 2019-05-29T18:56:33 | 2019-05-29T18:56:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 497,418,030 | 5,485 | den-crane | OK. The only problem where is no function which support (windows/partition by).
runningAccumulate calculates totalSum without resetting on a next group.
if need this then your runningLag is equal runningDifference
SELECT k, runningDifference(v)
FROM (SELECT intDiv(number, 2) AS k, number v FROM numbers(10) OR... | 2019-05-30T17:43:04 | 2019-05-30T17:43:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 496,175,867 | 5,376 | tarabanton | > This looks like some race condition when using extended query protocol in postgres, e.g. a Bind message arrives before Parse message. I'd expect UseServerSidePrepare=0 to disable extended protocol, but apparently it's still used. Will try to reproduce.
FYI: In single-node setup bug is hard to reproduce. When we te... | 2019-05-27T11:10:31 | 2019-05-27T11:10:31 | {} | NONE |
ClickHouse | ClickHouse | 496,239,159 | 5,386 | filimonov | > Version 19.8 release date?
Soon. Stay tuned! :)
Before release - you can try master build on non-prod environments.
https://github.com/yandex/ClickHouse/commits/master
-> find last commit with greenmark
-> click the greenmark
-> click the Detail link next to ClickHouse build check
-> pick debs from first... | 2019-05-27T14:57:43 | 2019-05-27T14:57:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 496,242,789 | 4,064 | FedorKononov | Stumbled upon the same issue. | 2019-05-27T15:10:31 | 2019-05-27T15:10:31 | {} | NONE |
ClickHouse | ClickHouse | 496,656,331 | 5,452 | 4ertus2 | In continuation of #5446 | 2019-05-28T19:31:08 | 2019-05-28T19:34:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 496,764,991 | 4,935 | zhang2014 | > Do you really need this command? In my opinion, it's much better to use synchronous distributed inserts (the setting insert_distributed_sync = 1)
@alexey-milovidov
#### Motivation(from my friend):
Currently their clickhouse cluster is deployed in japan, us and china, and he needs clickhouse to provide a regul... | 2019-05-29T03:01:53 | 2019-05-29T03:01:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 497,282,282 | 5,473 | 4ertus2 | Reproduced on release build
```
2019.05.30 13:19:18.424934 [ 48 ] {} <Error> BaseDaemon: ########################################
2019.05.30 13:19:18.425005 [ 48 ] {} <Error> BaseDaemon: (version 19.9.1.1) (from thread 44) Received signal Segmentation fault (11).
2019.05.30 13:19:18.425030 [ 48 ] {} <Error> BaseDae... | 2019-05-30T10:21:46 | 2019-05-30T10:21:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 497,287,209 | 5,460 | gcw831 | @filimonov | 2019-05-30T10:40:11 | 2019-05-30T10:40:11 | {} | NONE |
ClickHouse | ClickHouse | 497,325,257 | 3,929 | zhang2014 | @vitlibar I may have to modify these suggestions later, because I'm preparing other [pr](https://github.com/zhang2014/ClickHouse/tree/feature/bloom_filter) | 2019-05-30T13:16:17 | 2019-05-30T13:16:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 495,964,954 | 5,417 | alexey-milovidov | That's awesome! | 2019-05-26T03:18:19 | 2019-05-26T03:18:19 | {} | MEMBER |
ClickHouse | ClickHouse | 496,237,849 | 5,400 | filimonov | If you are sure that it's not a something external killed clickhouse, then please prepare reproducible example. Some sample data & clickhouse version & create statements & exact commands you use to import data.
W/o that we are not able to help you. | 2019-05-27T14:52:49 | 2019-05-27T15:11:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 496,601,136 | 5,446 | alexey-milovidov | Integration test - looks unrelated. | 2019-05-28T16:57:31 | 2019-05-28T16:57:31 | {} | MEMBER |
ClickHouse | ClickHouse | 496,978,072 | 5,432 | filimonov | Fixed by #5437 (in master) | 2019-05-29T15:08:28 | 2019-05-29T15:08:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 497,316,222 | 5,483 | SergioSafonov | Sorry, i'm just trying to multiply Decimal and Float values
but i was confused about error report:
> Illegal types Decimal(18, 3) and Float64 of arguments of function multiply (version 19.5.3.8 (official build))
and also i read about this in documentation about support of multiply operations only for all Int and ... | 2019-05-30T12:43:11 | 2019-05-30T12:43:11 | {} | NONE |
ClickHouse | ClickHouse | 497,322,642 | 5,483 | 4ertus2 | It's expected behaviour. If you do need strict values you use Decimal. If you do not need strict values you use Floats. If you use both Float and Decimal it's not clear what result do you want: strict or not. You should explicitly cast one of arguments to say it. | 2019-05-30T13:06:54 | 2019-05-30T13:07:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 495,934,603 | 5,116 | alexey-milovidov | Now it is Ok.
PS. Let's not forget to mention it as backward incompatible change in changelog. | 2019-05-25T17:21:15 | 2019-05-25T17:21:15 | {} | MEMBER |
ClickHouse | ClickHouse | 496,022,064 | 5,042 | alexey-milovidov | > Падают тесты с названием [, т.е. все ок
It isn't. | 2019-05-26T18:40:43 | 2019-05-26T18:40:43 | {} | MEMBER |
ClickHouse | ClickHouse | 495,501,410 | 5,400 | WeiGangqiang | > oomkiller
Thanks for your comment, but i cannot find any killed process log in demsg like this:
dmesg | egrep -i 'killed process', | 2019-05-24T07:21:13 | 2019-05-27T14:48:03 | {} | NONE |
ClickHouse | ClickHouse | 496,290,037 | 5,174 | ohoshio | Reporting back to confirm that after updating to ClickHouse server version 19.6.2 revision 54418 the bug is gone. | 2019-05-27T19:22:45 | 2019-05-27T19:22:45 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 496,506,921 | 5,441 | den-crane | password_sha256_hex is not supported against **remote_servers**.
**remote_servers** uses these passwords to connect to other instances of CH.
So there is only way is to set open password. | 2019-05-28T13:02:38 | 2019-05-28T13:02:38 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 496,629,104 | 1,148 | 4ertus2 | These queries do work
```
CREATE TABLE default.fs2M
ENGINE = Memory AS
SELECT
toString(murmurHash3_128(reinterpretAsString(number))) AS id,
rand(1) / (rand(2) + 1) AS value1,
toFloat64(rand(3) / (rand(4) + 1)) AS value2,
rand(5) AS value3,
rand64(6) AS value4,
base64Encode(concat(reint... | 2019-05-28T18:13:41 | 2019-05-28T18:14:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 496,756,959 | 5,441 | gle4er | > password_sha256_hex is not supported against **remote_servers**.
>
> **remote_servers** uses these passwords to connect to other instances of CH.
> So there is only way is to set open password.
Okay, I got it. | 2019-05-29T02:16:00 | 2019-05-29T02:16:00 | {} | NONE |
ClickHouse | ClickHouse | 496,770,985 | 4,207 | yingfeng | Could bitmap be used as a filter in the WHERE clause? | 2019-05-29T03:36:25 | 2019-05-29T03:36:25 | {} | NONE |
ClickHouse | ClickHouse | 496,961,487 | 5,462 | filimonov | There is already such a compound IP type, you can store both of them as IPv6.
Just because any IPv4 has it's unique IPv6 mapping:
IPv4 IP address: 191.239.213.197
IPv4-mapped IPv6 address: ::ffff:191.239.213.197
All you have to do is add ::ffff: in front of your IPv4 address.
See also: https://en.wikipedia.o... | 2019-05-29T14:29:48 | 2019-05-29T14:36:20 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 497,286,077 | 5,478 | alesapin | Actually it's not real bugfix. Just reduce amount of logs in MySQL handler. | 2019-05-30T10:35:41 | 2019-05-30T10:35:41 | {} | MEMBER |
ClickHouse | ClickHouse | 495,921,153 | 5,200 | alexey-milovidov | Ok.
Formulas are not completely generalized; there is no static_assert/enable_if in functions for higher order moments. But still good to merge. | 2019-05-25T14:07:43 | 2019-05-25T14:07:43 | {} | MEMBER |
ClickHouse | ClickHouse | 496,073,279 | 5,372 | amosbird | FYI https://github.com/yandex/ClickHouse/blob/96e3574/dbms/src/Interpreters/Aggregator.cpp#L1772
for order-preserving parallel processing. | 2019-05-27T04:27:25 | 2019-05-27T04:27:25 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 496,286,912 | 5,245 | filimonov | #5431 | 2019-05-27T19:03:18 | 2019-05-27T19:03:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 496,349,372 | 5,435 | alexey-milovidov | Ok.
But I don't understand how the changes in perf test help. | 2019-05-28T03:11:17 | 2019-05-28T03:11:17 | {} | MEMBER |
ClickHouse | ClickHouse | 496,697,816 | 4,203 | den-crane | 100 rows if:
1. it is the same shard.
2. it is the exactly the same insert batch (100 rows, the same order) (batch produces the same hashsum).
3. less than 100 insert happened after fail
4. less than week while after fail
5. original insert covered one partition
6...21 and so on. I am tired to type | 2019-05-28T21:31:25 | 2019-05-28T21:31:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 496,766,593 | 4,518 | jangorecki | Thanks for prompt reply. Could we re-open this issue then? | 2019-05-29T03:10:54 | 2019-05-29T03:13:20 | {} | NONE |
ClickHouse | ClickHouse | 497,305,792 | 5,483 | 4ertus2 | It's possible to use Decimals in dictionaries in master #5343
What do you mean saying you cannot use Decimals in operators? | 2019-05-30T12:00:17 | 2019-05-30T12:00:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 495,924,892 | 5,384 | manish53 | @vitlibar #5387,#5388 are created with patch files which shows where i've made changes as compared to this PR where all code are put in one go. | 2019-05-25T15:01:53 | 2019-05-25T15:01:53 | {} | NONE |
ClickHouse | ClickHouse | 496,351,823 | 5,409 | WeiGangqiang | > @WeiGangqiang [#5125 (comment)](https://github.com/yandex/ClickHouse/issues/5125#issuecomment-487599751)
Thank you, this is helpful | 2019-05-28T03:26:46 | 2019-05-28T03:26:46 | {} | NONE |
ClickHouse | ClickHouse | 496,701,447 | 3,925 | vzakaznikov | Currently working on this in my private fork at https://github.com/vzakaznikov/ClickHouse/tree/liveview and once the first draft is ready, I will create a new pull request. | 2019-05-28T21:43:32 | 2019-05-28T21:43:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 496,962,780 | 5,460 | filimonov | Sorry, it's very hard to understand the question due to poor translation to English.
If you can connect only from localhost most probably you need to adjust `<listen>` directive in your `/etc/clickhouse-server/config.xml`.
| 2019-05-29T14:32:51 | 2019-05-29T14:32:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 497,298,452 | 5,370 | alesapin | Fourth number in version doesn't mean anything in releases. It's uniquely determined by third number.
It's hard to fix in our CI, because we update `version.cmake` for release on previous release. We can make additional commit to `version.cmake` before tagging, but it seems ugly. Maybe you can made workaround inside ... | 2019-05-30T11:27:20 | 2019-05-30T11:36:55 | {} | MEMBER |
ClickHouse | ClickHouse | 495,917,424 | 5,246 | alexey-milovidov | I'll merge and you can do the remaining later. | 2019-05-25T13:11:18 | 2019-05-25T13:11:18 | {} | MEMBER |
ClickHouse | ClickHouse | 495,925,194 | 5,384 | manish53 | In mongo atlas for connection purpose atlas provide a connection string which consist of db's username,password,cluster info etc. In order to connect through this connection string(uri) to atlas, clickhouse needed to support such url form. This support is added here with ssl/non-ssl support. | 2019-05-25T15:06:08 | 2019-05-25T15:07:47 | {} | NONE |
ClickHouse | ClickHouse | 497,193,946 | 5,474 | alexey-milovidov | This type of subqueries (independent subqueries that return a single value) are named "scalar subqueries". They are executed only once. Yes, it is guaranteed. | 2019-05-30T04:02:24 | 2019-05-30T04:02:39 | {} | MEMBER |
ClickHouse | ClickHouse | 497,206,417 | 5,475 | alexey-milovidov | Yes, there is a detail that negatively affects performance of this particular query.
The result of scalar subquery is substituted to the parent query in AST level. When you have large array, it will create large temporary value, that will be placed in AST and interpreted.
It is going to be fixed in https://github... | 2019-05-30T05:19:40 | 2019-05-30T05:19:40 | {} | MEMBER |
ClickHouse | ClickHouse | 497,408,070 | 5,485 | svladykin | @den-crane
Thanks, but this does not answer any of my questions :)
Of course it is possible to `groupArray`/`arrayJoin` back and forth, but it is not always convenient and probably it will not be very efficient for a huge number of rows..
| 2019-05-30T17:13:56 | 2019-05-30T17:13:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 493,643,133 | 3,834 | ketbra | Perfect. Thank you! | 2019-05-18T02:59:34 | 2019-05-18T02:59:34 | {} | NONE |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.