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 | 645,201,254 | 11,722 | filimonov | > specify parameter to not-do merge across partitions
You can just add partition key to your ORDER BY | 2020-06-17T07:24:10 | 2020-06-17T07:24:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 645,694,601 | 6,576 | alexey-milovidov | I found that the info was moved here:
https://clickhouse.tech/docs/en/sql-reference/statements/select/group-by/#select-group-by-in-external-memory | 2020-06-18T00:21:22 | 2020-06-18T00:21:22 | {} | MEMBER |
ClickHouse | ClickHouse | 645,744,453 | 6,462 | guoshuai2016 | we also have the same issue. what's more, we find that optimize table final after detach and attach the overdue partition, the ttl works. more background: the ttl is added when table is created.
BTW, we use the version v19.14.7.15, and does upgrade help? | 2020-06-18T03:12:26 | 2020-06-18T06:28:07 | {} | NONE |
ClickHouse | ClickHouse | 645,939,096 | 11,488 | 4ertus2 | There's not special feature here. The table should be present in query itself or via an alias. In this issue we need just a fail on second `xyz.*`. | 2020-06-18T10:47:28 | 2020-06-18T10:47:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 644,723,538 | 11,590 | akuzm | tcmalloc uses twice less memory than jemalloc ([in this perf test run](https://clickhouse-test-reports.s3.yandex.net/11590/cd902e7d2b0c0c1183507560a6edab8a9d5b1b7a/performance_comparison/report.html#fail1)):
![tcmalloc-MemoryResident](https://user-images.githubusercontent.com/36882414/84772242-56633280-afe3-11ea-947... | 2020-06-16T12:11:30 | 2020-06-16T12:12:49 | {
"+1": 3
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 644,807,151 | 11,629 | filimonov | Stack trace with line numbers (decoded from addresses):
```
DB::MergeTreeData::alterDataPart(DB::NamesAndTypesList const&, std::__1::vector<std::__1::shared_ptr<DB::ASTIndexDeclaration>, std::__1::allocator<std::__1::shared_ptr<DB::ASTIndexDeclaration> > > const&, bool, std::__1::unique_ptr<DB::MergeTreeData::AlterDa... | 2020-06-16T14:37:53 | 2020-06-16T14:38:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 645,028,201 | 11,718 | alexey-milovidov | @abyss7 it looks like that computed gotos in Hyperscan jumps into the code of different functions and back. And without saving frame pointer, unwinding cannot work. | 2020-06-16T21:45:22 | 2020-06-16T21:45:22 | {} | MEMBER |
ClickHouse | ClickHouse | 645,286,275 | 11,722 | siradjev | > You can just add partition key to your ORDER BY
True. Could be achieved this way too at the expense of minor increase of primary key.
However, it would still worth to have it in merge-tree setting level, as some people may expect this behavior to be default (consistent behavior and result for background merges... | 2020-06-17T10:16:26 | 2020-06-17T10:16:26 | {} | NONE |
ClickHouse | ClickHouse | 645,777,564 | 11,743 | siradjev | @den-crane you are right.
It's actually a design https://clickhouse.tech/docs/ru/sql-reference/aggregate-functions/
I'm closing this ticket. | 2020-06-18T05:14:01 | 2020-06-18T05:14:01 | {} | NONE |
ClickHouse | ClickHouse | 644,974,841 | 11,498 | filimonov | And definitely need an opinion of @vitlibar | 2020-06-16T19:44:14 | 2020-06-16T19:44:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 645,009,448 | 11,643 | filimonov | does it work across blocks / threads? | 2020-06-16T21:01:13 | 2020-06-16T21:01:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 645,074,291 | 11,710 | alexey-milovidov | BTW, we support COLLATE modifier for ORDER BY.
It's also nice to implement it for LIKE, but the performance of proper unicode collation algorithm will be much worse. | 2020-06-17T00:14:54 | 2020-06-17T00:14:54 | {} | MEMBER |
ClickHouse | ClickHouse | 645,117,740 | 11,723 | den-crane | Yes. And try some newer CH version. For example 19.14.11.16 or other stable https://repo.yandex.ru/clickhouse/deb/stable/main/ or lts https://repo.yandex.ru/clickhouse/deb/lts/main/
19.9.5.36 is out of support | 2020-06-17T02:59:23 | 2020-06-17T02:59:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 645,166,287 | 11,643 | ghost | `EarlyWindowFunction` collects only local data not for the distributed data.
When Data is distributed by `user_id` among nodes, the partition key of `EarlyWindowFunction` must be the `user_id`. | 2020-06-17T05:49:48 | 2020-06-17T05:50:11 | {} | NONE |
ClickHouse | ClickHouse | 645,178,416 | 11,716 | alexey-milovidov | It did not help, will try to revert changes for Hyperscan. | 2020-06-17T06:26:59 | 2020-06-17T06:26:59 | {} | MEMBER |
ClickHouse | ClickHouse | 645,234,066 | 11,629 | jackpgao | May be the reason is that the table is corrupted:
```sql
Code: 49. DB::Exception: Received from 127.0.0.1:9000. DB::Exception: Could not find a column of minimum size in MergeTree, part /data1/clickhouse/data/xxx/xxx/20190827_428332_429264_30/.
```
```shell
root@xxx:/data1/clickhouse/data/x/xxx/20190827_428332... | 2020-06-17T08:30:21 | 2020-06-17T11:26:41 | {} | NONE |
ClickHouse | ClickHouse | 645,602,001 | 11,479 | millin | On versions 19.14-19.17 such kind of queries worked fine.
`tbl_distributed` is Distributed table with 3 shard * 2 replicas.
```
SELECT toTime(toStartOfHour(dttm)) AS dt
, name
, avg(distance) AS avg_distance
FROM tbl_distributed
GLOBAL ANY
LEFT JOIN
(
SELECT id
... | 2020-06-17T20:18:17 | 2020-06-17T20:18:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 644,718,054 | 11,568 | bookeezhou | > Is there additional text in stderr?
> assertion is raised from /usr/lib64/libc-2.17.so. Can you install debug symbols for libc and find what is in 0x35228 address or just rerun server?
when edit config.xml file , delete **openSSL** item
then clickhouse-server has down. | 2020-06-16T11:59:06 | 2020-06-16T11:59:06 | {} | NONE |
ClickHouse | ClickHouse | 644,808,586 | 11,629 | filimonov | @jackpgao can you provide more details? Does it reproduces always? Can you try to create a reproducible test case? May be smth was happening between you start ALTER and get the segfault? | 2020-06-16T14:40:21 | 2020-06-16T16:09:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 645,034,616 | 11,639 | alexey-milovidov | Build check:
> Build failed without build log. It's inner CI problem. Maybe something wrong with docker or host. Just wait for build restart.
It started to be annoying. | 2020-06-16T22:02:23 | 2020-06-16T22:02:23 | {} | MEMBER |
ClickHouse | ClickHouse | 645,116,347 | 11,723 | byx313 | > Check this #6462
"TTL works only for new parts after alter."
Our TTL time is 30 days,is it mean that only 30 days after my 'alter' action data will be drop? | 2020-06-17T02:53:43 | 2020-06-17T02:53:43 | {} | NONE |
ClickHouse | ClickHouse | 645,200,153 | 11,288 | alexey-milovidov | I have removed `- [ ] Illegal column Const(UInt32) of argument of function IPv4CIDRToRange #11245` from the checklist. | 2020-06-17T07:21:48 | 2020-06-17T07:21:48 | {} | MEMBER |
ClickHouse | ClickHouse | 645,215,252 | 11,508 | nikitamikhaylov | > To remove create ref by your you can run the following:
>
> ```
> git push upstream :pr/11508
> ```
Done
| 2020-06-17T07:53:25 | 2020-06-17T07:53:25 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 645,437,111 | 11,564 | alexey-milovidov | According to log, the time is spent on analysis of minmax (partition) index: `toDate(toDateTime(-antiTimestamp))`, not the primary key.
But it's unclear why it takes so long. | 2020-06-17T15:14:01 | 2020-06-17T15:14:01 | {} | MEMBER |
ClickHouse | ClickHouse | 645,754,924 | 9,577 | alexey-milovidov | @excitoon Any update on this? | 2020-06-18T03:51:02 | 2020-06-18T03:51:02 | {} | MEMBER |
ClickHouse | ClickHouse | 645,794,529 | 8,282 | filimonov | Close in favor to #9259 | 2020-06-18T05:57:44 | 2020-06-18T05:57:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 645,292,020 | 1,261 | kymtwyf | Hi @blinkov @ludv1x may I ask is resharding supported in latest versions ? if yes, how can I use it?
Thanks! | 2020-06-17T10:29:06 | 2020-06-17T10:29:06 | {} | NONE |
ClickHouse | ClickHouse | 645,590,562 | 11,743 | den-crane | tuple() is a WA.
```
SELECT groupArrayInsertAt(NULL, 5)(tuple(toNullable(num)), num).1
FROM
(
SELECT number AS num
FROM system.numbers
LIMIT 3
)
GROUP BY tuple()
┌─tupleElement(groupArrayInsertAt(NULL, 5)(tuple(toNullable(num)), num), 1)─┐
│ [0,1,2,NULL,NULL] ... | 2020-06-17T19:56:34 | 2020-06-17T19:56:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 645,695,523 | 10,047 | alexey-milovidov | https://github.com/ClickHouse/ClickHouse/issues/11741 | 2020-06-18T00:23:30 | 2020-06-18T00:23:30 | {} | MEMBER |
ClickHouse | ClickHouse | 645,085,697 | 11,710 | siradjev | It could be hardcoded on implementation level to replicate ILIKE for a start.
Essentially, ILIKE is syntactic sugar, and adding it to syntax for sake of optimizing `lower(col) like '%searchtxt%'` sounds strange. It sounds more logical to fix it on the code level, to have "lower(xxx) like" and "upper (xxx) like" optim... | 2020-06-17T00:59:08 | 2020-06-17T00:59:08 | {} | NONE |
ClickHouse | ClickHouse | 645,534,659 | 11,739 | azat | Without the patch test fails:
```
> node.query('ATTACH TABLE data')
test_replicated_parse_zk_metadata/test.py:48:
...
E Code: 342. DB::Exception: Received from 172.19.0.5:9000. DB::Exception: Existing table metadata in ZooKeeper differs in skip indexes. Stored in ZooKeeper: key_idx0 key + 0 T... | 2020-06-17T18:08:01 | 2020-06-17T18:08:01 | {} | MEMBER |
ClickHouse | ClickHouse | 645,813,924 | 6,462 | alexey-milovidov | > BTW, we use the version v19.14.7.15, and does upgrade help?
Calculation of TTL for old data after MODIFY TTL will be available in version 20.5 (that is currently in testing). | 2020-06-18T06:44:05 | 2020-06-18T06:44:05 | {} | MEMBER |
ClickHouse | ClickHouse | 645,886,353 | 10,397 | maison2710 | Hi, any update on this? | 2020-06-18T09:05:19 | 2020-06-18T09:05:19 | {} | NONE |
ClickHouse | ClickHouse | 645,209,153 | 11,603 | ildus | Thanks! | 2020-06-17T07:40:45 | 2020-06-17T07:40:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 645,215,079 | 11,508 | azat | Or better using qualified path:
```
git push upstream :refs/heads/pr/11508
``` | 2020-06-17T07:53:05 | 2020-06-17T07:53:05 | {} | MEMBER |
ClickHouse | ClickHouse | 644,927,420 | 11,498 | azat | @abyss7 did you have a chance to look? (I'm curious about your thoughts on the idea in general) | 2020-06-16T18:15:03 | 2020-06-16T18:15:03 | {} | MEMBER |
ClickHouse | ClickHouse | 645,488,828 | 11,736 | alexey-milovidov | CC @Enmk | 2020-06-17T16:45:07 | 2020-06-17T16:45:07 | {} | MEMBER |
ClickHouse | ClickHouse | 645,718,670 | 11,194 | amy108 | @amitsharma100 sorry， i overlooked my email，what @tetafro said is the perfect solution. | 2020-06-18T01:44:53 | 2020-06-18T01:44:53 | {} | NONE |
ClickHouse | ClickHouse | 645,737,811 | 11,488 | hexiaoting | Does the feature of select xyz.* already implemented?
Why I try below command failed?
Need I implement this feature first?
```
dell124 :) show tables;
SHOW TABLES
┌─name──────┐
│ hits_v1 │
│ visits_v1 │
└───────────┘
2 rows in set. Elapsed: 0.003 sec.
dell124 :) select hits_v1.*
SELECT hits_v1... | 2020-06-18T02:49:49 | 2020-06-18T02:49:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 645,807,051 | 7,589 | john8628 | how to solve the problem ,? | 2020-06-18T06:27:26 | 2020-06-18T06:27:26 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 645,822,181 | 11,747 | filimonov | Actually you can't fit with Time into 2 bytes (65536 seconds = 18h), and if you will use 4bytes for that on storage level - you get to DateTime (rounded to some 'zero' day), that is implemented via [toTime](https://clickhouse.tech/docs/en/sql-reference/functions/date-time-functions/#totime) function.
Sidenote: it i... | 2020-06-18T07:02:39 | 2020-06-18T07:03:42 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 645,861,203 | 11,746 | azat | >How it happened to succeed in CI?
This is interesting, I don't know for sure, will take a look later
>Alternative solution: as aggregate function count is used not only through the factory and we have direct dependency, we can move its implementation to dbms library out of clickhouse_aggregate_functions.
But ... | 2020-06-18T08:20:00 | 2020-06-18T08:20:00 | {} | MEMBER |
ClickHouse | ClickHouse | 644,848,409 | 11,506 | filimonov | ILIKE part go to a separate task: #11710 | 2020-06-16T15:46:46 | 2020-06-16T15:46:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 644,890,334 | 4,565 | den-crane | @alesapin
>ALTER query concurrently from a few processes
session1: ALTER TABLE "test" ADD COLUMN IF NOT EXISTS "foo1" String
session2: ALTER TABLE "test" ADD COLUMN IF NOT EXISTS "foo1" String
```
t1. session1 checks | session2 checks
t2. session1 adds col |
t3 | session2 fails... | 2020-06-16T16:59:39 | 2020-06-16T17:00:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 644,925,605 | 11,615 | azat | >@azat Stress test found a datarace https://gist.githubusercontent.com/nikitamikhaylov/acf45aa112f3941494ca508eca35d4c8/raw/25c38d24b47801ada2c95eb09a0cd3ae93e6e7f1/datarace.md
Thanks, fixed!
>It is inconvenient that such things are shown only in the private network when the test is timeout
Indeed, I was waiti... | 2020-06-16T18:11:19 | 2020-06-16T18:11:19 | {} | MEMBER |
ClickHouse | ClickHouse | 645,422,279 | 11,194 | tetafro | @amitsharma100 check your `log_query` setting.
```
SELECT name, value FROM system.settings WHERE name LIKE 'log_queries'
┌─name────────┬─value─┐
│ log_queries │ 1 │
└─────────────┴───────┘
```
It can be set in users.xml
```
<yandex>
<profiles>
<default>
<log_queries>1</log_q... | 2020-06-17T14:49:45 | 2020-06-17T14:49:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 645,485,368 | 11,071 | r-zenine | Hi @Akazz,
I hope you are doing well, everything should be in order.
Thanks again for the careful reviews and for the feedback :)
Have a good day. | 2020-06-17T16:39:08 | 2020-06-17T16:39:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 645,021,133 | 11,718 | alexey-milovidov | And I successfully reproduced it on gcc build from CI on master!
(gcc build, single binary)
Maybe it's a bug in gcc (generating CFI for inline asm). | 2020-06-16T21:27:45 | 2020-06-16T21:28:55 | {} | MEMBER |
ClickHouse | ClickHouse | 645,373,273 | 11,711 | tavplubix | #11732 | 2020-06-17T13:27:42 | 2020-06-17T13:27:42 | {} | MEMBER |
ClickHouse | ClickHouse | 645,007,149 | 10,220 | siradjev | As a side comment we would recommend to store expression of each tuple element and have mapping of each tuple element to dependent columns, rather than overall expression, to cover proper partition elimination/pruning for multi-level partitioning with hash-partitioning cases like below:
`
PARTITION BY (toYYYYMM(date... | 2020-06-16T20:56:04 | 2020-06-16T20:56:04 | {} | NONE |
ClickHouse | ClickHouse | 645,046,854 | 11,508 | azat | @nikitamikhaylov looks like you pushed to the `refs/heads/pr/11508` remote branch in the upstream repository, and it conflicts with other PR's refs, will it be rename/delete eventually? Or should I change the local ref folder?
```
$ git fetch upstream
fatal: Cannot fetch both refs/heads/pr/11508 and refs/pull/115... | 2020-06-16T22:37:31 | 2020-06-16T22:37:31 | {} | MEMBER |
ClickHouse | ClickHouse | 645,119,423 | 11,723 | den-crane | >Our environment is centos6,so 19.9.5.36 is our newest version
Are you sure? I think newer versions should work as well https://repo.yandex.ru/clickhouse/rpm/stable/x86_64/ | 2020-06-17T03:05:52 | 2020-06-17T03:05:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 645,569,219 | 6,576 | jangorecki | @alexey-milovidov your link does not seems to work. Searching for "external memory" on a website it points now is not finding anything. Related issue #5618 has been closed, but it is not clear if it is resolved already. @filimonov wrote
> We have to copy the relevant information
was that information copied at all... | 2020-06-17T19:14:01 | 2020-06-17T19:18:37 | {} | NONE |
ClickHouse | ClickHouse | 645,751,011 | 11,746 | alexey-milovidov | Alternative solution: as aggregate function `count` is used not only through the factory and we have direct dependency, we can move its implementation to `dbms` library out of `clickhouse_aggregate_functions`. | 2020-06-18T03:35:33 | 2020-06-18T03:35:33 | {} | MEMBER |
ClickHouse | ClickHouse | 644,724,512 | 11,479 | 4ertus2 | Could you print your query and schema here. I've tryed to reproduce the issue and it seems to be not supported at all. I've got LOGICAL_ERROR on 20.3 for this case
```
CREATE TABLE source_table1 (a Int32, b String) ENGINE = Memory;
CREATE TABLE distributed_table1 AS source_table1
ENGINE = Distributed('test_shard_... | 2020-06-16T12:13:44 | 2020-06-16T12:14:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 644,850,695 | 8,863 | vitlibar | Fixed in https://github.com/ClickHouse/ClickHouse/pull/8926 and https://github.com/ClickHouse/ClickHouse/pull/9182/commits/5849dd22364b1f8187e134c5a73e02523e778376: now it applies filters defined for both initial user and shard's user, combining them with `AND`. | 2020-06-16T15:50:35 | 2020-06-16T15:52:21 | {} | MEMBER |
ClickHouse | ClickHouse | 644,849,282 | 11,708 | filimonov | Check: https://github.com/ClickHouse/ClickHouse/issues/5747
So for now it is not possible (it is planned but not concrete dates).
Since 20.4 you can use ssl-encrypted connection with zookeeper:
https://github.com/xzkostyan/ClickHouse/blob/a7c5f622ea4f5fdb6776994492533590b4789ff0/tests/integration/test_zookeeper... | 2020-06-16T15:48:12 | 2020-06-16T16:18:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 644,903,116 | 11,606 | akuzm | Still too fast...
Some are even faster than 0.02 s -- you can see this in the _Test times_ table, the column _Ignored short queries_ shows the number of short queries. | 2020-06-16T17:25:33 | 2020-06-16T17:26:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 644,977,817 | 11,706 | alexey-milovidov | The test has failed only due to the change of query formatting in master.
Just regenerate test result... | 2020-06-16T19:51:02 | 2020-06-16T19:51:02 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 645,075,346 | 11,690 | alexey-milovidov | > ClickHouse build check — 19/20 builds are OK
No idea what has happened. It looks like another Sandbox failure. | 2020-06-17T00:18:23 | 2020-06-17T00:18:23 | {} | MEMBER |
ClickHouse | ClickHouse | 645,080,515 | 2,858 | alexey-milovidov | Implemented and available in recent releases. | 2020-06-17T00:37:10 | 2020-06-17T00:37:10 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 645,196,998 | 11,508 | nikitamikhaylov | @azat Yes, I pushed empty commit here and I don’t know how to return it back :) Maybe we can simply close this PR and corresponding branch will be deleted? | 2020-06-17T07:14:25 | 2020-06-17T07:14:25 | {} | MEMBER |
ClickHouse | ClickHouse | 645,212,100 | 11,508 | azat | >Maybe we can simply close this PR and corresponding branch will be deleted?
This will not helps, since you did not push the ref of PR but create new one:
- PR ref - `refs/pull/11508/head`
- You pushed to ref - `refs/heads/pr/11508`
To remove create ref by your you can run the following:
```
git push upstre... | 2020-06-17T07:46:40 | 2020-06-17T07:46:40 | {} | MEMBER |
ClickHouse | ClickHouse | 645,316,030 | 11,726 | alexey-milovidov | > ClickHouse build check — 19/20 builds are OK
Sandbox failure. | 2020-06-17T11:25:16 | 2020-06-17T11:25:16 | {} | MEMBER |
ClickHouse | ClickHouse | 645,692,781 | 11,746 | alexey-milovidov | How it happened to succeed in CI? | 2020-06-18T00:16:36 | 2020-06-18T00:16:36 | {} | MEMBER |
ClickHouse | ClickHouse | 645,852,860 | 11,751 | dependabot-preview[bot] | Looks like requests is up-to-date now, so this is no longer needed. | 2020-06-18T08:04:02 | 2020-06-18T08:04:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 645,877,753 | 11,740 | alesapin | Hi. To investigate the issue, please, provide server logs, table schema, and data samples if possible. Also, I recommend you to update to the latest 20.3 version, we had several fixes for mutations since 20.3.5. | 2020-06-18T08:50:48 | 2020-06-18T08:50:48 | {} | MEMBER |
ClickHouse | ClickHouse | 645,025,712 | 11,718 | alexey-milovidov | @abyss7 I don't see any complex inline assembly in Hyperscan.
But there is a huge switch/case with computed gotos (in `roseRunProgram` function). | 2020-06-16T21:38:57 | 2020-06-16T21:38:57 | {} | MEMBER |
ClickHouse | ClickHouse | 645,052,337 | 11,710 | siradjev | It may be better to use ANSI SQL way using COLLATE settings of db/table/column/session/statement, even if in limited way in the beginning supporting only _CI suffix in statement level itself.
https://oracle-base.com/articles/12c/column-level-collation-and-case-insensitive-database-12cr2
https://stackoverflow.com/qu... | 2020-06-16T22:56:18 | 2020-06-16T22:56:18 | {} | NONE |
ClickHouse | ClickHouse | 645,199,987 | 11,245 | alexey-milovidov | Reproduced on
master
20.4
20.3
20.1
19.14
On 19.1 there is no such function. | 2020-06-17T07:21:27 | 2020-06-17T07:21:27 | {} | MEMBER |
ClickHouse | ClickHouse | 645,232,210 | 8,339 | jackpgao | +1 | 2020-06-17T08:26:33 | 2020-06-17T08:26:33 | {} | NONE |
ClickHouse | ClickHouse | 645,273,983 | 11,711 | tavplubix | Looks like something was broken in #9740 while tests was disabled. Need to fix it before merging this PR. | 2020-06-17T09:50:29 | 2020-06-17T09:50:29 | {} | MEMBER |
ClickHouse | ClickHouse | 645,589,741 | 11,743 | siradjev | Seems like generic bug with groupArray
```
SELECT groupArray(toNullable(num))
FROM
(
SELECT toNullable(1) AS num
UNION ALL
SELECT NULL AS num
)
GROUP BY tuple()
┌─groupArray(toNullable(num))─┐
│ [1] │
└─────────────────────────────┘
```
| 2020-06-17T19:54:52 | 2020-06-17T19:54:52 | {} | NONE |
ClickHouse | ClickHouse | 645,485,960 | 11,599 | filimonov | The current state of formats supported by engine=Kafka
| name | checked | working |
|:-|-:|-:|
| Arrow | 1 | 0 |
| ArrowStream | 1 | 0 |
| Avro | 1 | 0 |
| AvroConfluent | 0 | 0 |
| CSV | 1 | 1 |
| CSVWithNames | 1 | 1 |
| CapnProto | 1 | 1 |
| CustomSeparated | 0 | ? |
| CustomSeparatedIgnoreSpaces | 0 ... | 2020-06-17T16:40:08 | 2020-06-18T05:46:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 645,852,670 | 910 | smedoji | @shuuky @Negashev
I am facing the below error. Tried forcing restore but clickhouse server kept restarting trying to restore but never actually recovered. Any help?
`<Error> Application: DB::Exception: The local set of parts of table map_test._test doesn't look like the set of parts in ZooKeeper: 196.39 million ... | 2020-06-18T08:03:40 | 2020-06-18T08:03:40 | {} | NONE |
ClickHouse | ClickHouse | 645,882,376 | 11,746 | azat | >How it happened to succeed in CI?
Here are some differences:
- gold linker for one error
- lld-8 for another (while there is lld-10 on CI) | 2020-06-18T08:58:58 | 2020-06-18T08:58:58 | {} | MEMBER |
ClickHouse | ClickHouse | 644,763,900 | 11,706 | alexey-milovidov | Maybe also check the behaviour in case when we are adding a column and the same column with IF NOT EXISTS clause in the same ALTER.
(both throwing exception and ignoring the second add is Ok) | 2020-06-16T13:26:18 | 2020-06-16T13:26:18 | {} | MEMBER |
ClickHouse | ClickHouse | 644,764,737 | 11,639 | alexey-milovidov | Build check:
> Build failed without build log. It's inner CI problem. Maybe something wrong with docker or host. Just wait for build restart. | 2020-06-16T13:27:22 | 2020-06-16T13:27:22 | {} | MEMBER |
ClickHouse | ClickHouse | 644,899,158 | 10,973 | akuzm | The tests don't pass and the title says [wip], so I marked this PR as draft for the ease of filtering. | 2020-06-16T17:17:33 | 2020-06-16T17:17:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 644,934,768 | 11,681 | alesapin | Performance not affected. | 2020-06-16T18:26:07 | 2020-06-16T18:26:07 | {} | MEMBER |
ClickHouse | ClickHouse | 644,971,315 | 11,498 | filimonov | 1) does it work for users defined via users.xml?
2) `ALTER USER initial_user ADD PROXY default` - i don't like the syntax - sounds like that proxy will be used always. Maybe smth like `ALLOW PROXYING VIA xxx` / `GRANT IMPERSONATION` or similar. Not clear where checked (on initiator or on target) and what if the user h... | 2020-06-16T19:36:19 | 2020-06-16T19:49:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 645,058,056 | 11,661 | alexey-milovidov | This PR is also notable because we will augment all factories with documentation in similar way
\- to make documentation available within running server. | 2020-06-16T23:15:05 | 2020-06-16T23:15:17 | {} | MEMBER |
ClickHouse | ClickHouse | 645,114,290 | 11,723 | den-crane | Check this https://github.com/ClickHouse/ClickHouse/issues/6462 | 2020-06-17T02:45:36 | 2020-06-17T02:45:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 645,795,871 | 11,277 | filimonov | The issue is most probably caused by
https://github.com/ClickHouse/ClickHouse/issues/6725
and additionally
https://github.com/ClickHouse/ClickHouse/issues/9259
You need to update. | 2020-06-18T06:00:37 | 2020-06-18T06:01:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 644,761,868 | 11,682 | alexey-milovidov | @KochetovNicolai But it fails in all test runs? | 2020-06-16T13:23:19 | 2020-06-16T13:23:19 | {} | MEMBER |
ClickHouse | ClickHouse | 645,009,878 | 11,683 | akuzm | The error in performance tests is expected, because we're adding a new function that is not present in the old master. | 2020-06-16T21:02:05 | 2020-06-16T21:02:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 645,079,864 | 11,701 | gubinjie | > CH 19.16.17.1
Which version is recommended? v20.3 LTS? | 2020-06-17T00:34:25 | 2020-06-17T00:34:25 | {} | NONE |
ClickHouse | ClickHouse | 645,118,659 | 11,723 | byx313 | > Yes. And try some newer CH version. For example 19.14.11.16 or other stable https://repo.yandex.ru/clickhouse/deb/stable/main/ or lts https://repo.yandex.ru/clickhouse/deb/lts/main/
>
> 19.9.5.36 is out of support
Our environment is centos6,so 19.9.5.36 is our newest version
Thanks again for reply | 2020-06-17T03:02:49 | 2020-06-17T03:02:49 | {} | NONE |
ClickHouse | ClickHouse | 645,200,502 | 11,245 | alexey-milovidov | The bug was introduced here: #5095 | 2020-06-17T07:22:32 | 2020-06-17T07:22:32 | {} | MEMBER |
ClickHouse | ClickHouse | 645,460,287 | 6,312 | iameugenejo | @alexey-milovidov which version, commit or PR fixes this issue? I have `19.14.10.16` running and just encountered this issue, and I'd like to upgrade to the one that fixes this issue. | 2020-06-17T15:53:28 | 2020-06-17T15:53:28 | {} | NONE |
ClickHouse | ClickHouse | 644,720,855 | 11,672 | filimonov | 20.1 - affected.
19.17 affected (checked 19.17.2 and 19.17.9, forces enabling any_join_distinct_right_table_keys=1)
19.16 works (after enabling any_join_distinct_right_table_keys=1)
It looks like 19.16 moves another condition to PREWHERE:
```
ANALYZE SELECT count()
FROM foo
ANY INNER JOIN bar USING (dimensio... | 2020-06-16T12:05:20 | 2020-06-16T12:05:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 645,023,217 | 11,718 | alexey-milovidov | ```
#0 0x0000000011a9f97c in libunwind::DwarfInstructions<libunwind::LocalAddressSpace, libunwind::Registers_x86_64>::getSavedRegister (savedReg=..., cfa=140736606982880, registers=...,
addressSpace=...) at ../contrib/libc-headers/x86_64-linux-gnu/bits/string_fortified.h:34
#1 libunwind::DwarfInstructions<lib... | 2020-06-16T21:32:35 | 2020-06-16T21:32:47 | {} | MEMBER |
ClickHouse | ClickHouse | 645,116,071 | 11,643 | ghost | @filimonov
It seems to work fine across blocks as the below example.
```SQL
CREATE TABLE default.wl1
(
`d` Date,
`dt` DateTime,
`id` String,
`v` Int32
)
ENGINE = MergeTree()
PARTITION BY d
ORDER BY id
SETTINGS index_granularity = 8192;
```
```SQL
SELECT *
FROM wl1
FORMAT PrettySpace
... | 2020-06-17T02:52:34 | 2020-06-17T02:52:34 | {} | NONE |
ClickHouse | ClickHouse | 645,112,896 | 11,724 | den-crane | You can change MergeTree a parameter `old-parts-lifetime` in config.xml
```
<yandex>
<merge_tree>
<old_parts_lifetime>10</old_parts_lifetime>
<temporary_directories_lifetime>10</temporary_directories_lifetime>
</merge_tree>
</yandex>
```
---
## old \ _parts \ _lifetime {# o... | 2020-06-17T02:40:49 | 2020-06-17T03:00:32 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 645,203,921 | 11,710 | filimonov | > sake of optimizing lower(col) like '%searchtxt%' sounds strange
When you type lot of queries by hand it is really useful. Also it should give some performance gain as two operations can be executed in single round without copying.
> Most ORMs/smaller BI tools also generate code in this dialect
Because ILIKE ... | 2020-06-17T07:29:32 | 2020-06-17T07:31:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 645,406,344 | 11,053 | 4ertus2 | Closed in favor of #11733 | 2020-06-17T14:24:15 | 2020-06-17T14:24:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 645,570,349 | 11,740 | kanadaj | Restarting the server resulted in even more errors; had to resort to setting force_restore_data but I imagine deletes normally shouldn't have to do it. | 2020-06-17T19:16:23 | 2020-06-17T19:16:23 | {} | NONE |
ClickHouse | ClickHouse | 645,606,969 | 11,744 | alesapin | Just found that we lock structure during replicated fetches: https://github.com/clickhouse/ClickHouse/blob/master/src/Storages/StorageReplicatedMergeTree.cpp#L3086. I'll remove almost all locks soon, but this one is required in 20.3 and 20.4. | 2020-06-17T20:28:38 | 2020-06-17T20:28:58 | {} | MEMBER |
ClickHouse | ClickHouse | 626,256,252 | 10,578 | alexey-milovidov | And in stress test: https://clickhouse-test-reports.s3.yandex.net/10578/0618c806b77bc0682d1c43066d6d6a9c93d7faed/stress_test_(address)/stderr.log | 2020-05-10T00:59:34 | 2020-05-10T00:59:34 | {} | MEMBER |
ClickHouse | ClickHouse | 626,256,502 | 9,577 | alexey-milovidov | Should we continue? | 2020-05-10T01:02:17 | 2020-05-10T01:02:17 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.