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 | 537,873,460 | 7,180 | jpiper | I'll close this out as this seems to be new expected behaviour | 2019-10-03T09:50:09 | 2019-10-03T09:50:09 | {} | NONE |
ClickHouse | ClickHouse | 538,360,035 | 7,113 | akuzm | Here it is:
```
create table p (x Nullable(String)) engine MergeTree order by ifNull(x, 'order-null') partition by ifNull(x, 'partition-null');
insert into p values (null);
```
The segfault occurs when we try to serialize min/max values of partitioning columns, one such value unexpectedly turns out to be NUL... | 2019-10-04T11:31:17 | 2019-10-04T11:31:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 538,814,506 | 7,204 | dsmith47 | I would be interested in working on this. Test/structure wise, would it be fair to say that `arrayConcat` is a good model for adding functions? | 2019-10-07T01:57:45 | 2019-10-07T01:57:45 | {} | NONE |
ClickHouse | ClickHouse | 538,867,328 | 7,136 | filimonov | Same on 19.11.
Stacktrace:
```
2019.10.07 07:04:59.874706 [ 42 ] {fe56a8d4-6445-4b84-b5bd-f34ed4183da4} <Debug> executeQuery: (from 127.0.0.1:40772) SELECT ccc FROM (SELECT 1 AS ccc UNION ALL SELECT * FROM (SELECT 2 AS ccc) ANY INNER JOIN (SELECT 2 AS ccc) USING (ccc)) WHERE ccc > 1
2019.10.07 07:04:59.876837 [ 42... | 2019-10-07T07:08:47 | 2019-10-07T07:08:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 537,197,285 | 7,165 | filimonov | It's an aggregate function. Probably you need `arrayReduce`. | 2019-10-01T19:43:22 | 2019-10-01T19:43:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 537,381,816 | 7,172 | filimonov | Before there was 2 variants.
For replicated (7 zeros)
https://github.com/ClickHouse/ClickHouse/blob/0e47bbca4d67ccda1628fcec3b65a80f38a176fd/dbms/src/Storages/StorageReplicatedMergeTree.cpp#L967
For non-replicated (6 zeros)
https://github.com/ClickHouse/ClickHouse/blob/f0c5ee5877e2021b1c134ed8502aa6784933833c/d... | 2019-10-02T07:58:45 | 2019-10-02T08:10:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 537,588,971 | 7,149 | proller | All changes near ENABLE_.... should be reverted. | 2019-10-02T17:06:34 | 2019-10-02T17:06:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 538,351,883 | 7,113 | akuzm | Looks like it's already fixed here: https://github.com/ClickHouse/ClickHouse/pull/7135
Now that I know the cause, I'll try to create a simple test case. | 2019-10-04T11:01:04 | 2019-10-04T11:01:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 536,980,371 | 7,072 | stavrolia | Conflict with 19.11, 19.13 and 19.14 | 2019-10-01T10:47:59 | 2019-10-01T11:24:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 537,565,155 | 7,173 | ifoo | Upstream Debian package in LXC container on Debian host. | 2019-10-02T16:08:10 | 2019-10-02T16:08:10 | {} | NONE |
ClickHouse | ClickHouse | 537,706,390 | 5,939 | stavrolia | https://clickhouse.yandex/docs/en/query_language/agg_functions/parametric_functions/#histogram | 2019-10-02T22:21:49 | 2019-10-02T22:21:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 536,937,682 | 6,912 | ildus | Yeah, `mapFill` would help but function has third argument `max_key` which is used to extend both arrays. So in the end `mapFill` would require this `max_key` argument and final expression would look like this (10 is `max_key`):
```
SELECT arrayFill([1, 2, 6], 10), arrayCumSum(mapFill([1, 2, 6], [10, 20, 30], 10)[2... | 2019-10-01T08:52:15 | 2019-10-03T08:09:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 538,644,281 | 7,194 | alesapin | Fixed here https://github.com/ClickHouse/ClickHouse/pull/7155/files#diff-bcf50b560562d3f0464f75823443402eR718 | 2019-10-05T12:07:56 | 2019-10-05T12:07:56 | {} | MEMBER |
ClickHouse | ClickHouse | 538,680,430 | 7,203 | alexey-milovidov | To do this task, you have to merge clickhouse-client and clickhouse-local code. | 2019-10-05T19:15:51 | 2019-10-05T19:15:51 | {} | MEMBER |
ClickHouse | ClickHouse | 538,796,989 | 7,176 | alexey-milovidov | > Interestingly enough, when engine is Memory, everything works.
Because Memory engine doesn't have index. | 2019-10-06T23:01:54 | 2019-10-06T23:01:54 | {} | MEMBER |
ClickHouse | ClickHouse | 536,530,497 | 7,151 | edonin | Super, it is working.
Thanks | 2019-09-30T12:02:44 | 2019-09-30T12:02:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 536,693,910 | 7,157 | macdonsp | Might just be a documentation issue - https://clickhouse.yandex/docs/en/query_language/agg_functions/
```
The sum function interprets NULL as 0. In particular, this means that if the function receives input of a selection where all the values are NULL, then the result will be 0, not NULL.
``` | 2019-09-30T18:36:37 | 2019-09-30T18:36:37 | {} | NONE |
ClickHouse | ClickHouse | 536,992,426 | 7,140 | stavrolia | Conflict with 19.11, 19.13 and 19.14 | 2019-10-01T11:25:38 | 2019-10-01T11:25:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 537,863,441 | 7,180 | jpiper | the contents of the .sql file for this table is (column names renamed)
```
ATTACH TABLE table_shard
(
`date` Date,
`country` String,
`county` String,
`city` String,
`value` Float32
)
ENGINE = ReplicatedMergeTree('/clickhouse/tables/database/table_shard/{shard}/', '{replica}')
ORDER ... | 2019-10-03T09:21:38 | 2019-10-03T09:21:38 | {} | NONE |
ClickHouse | ClickHouse | 538,319,090 | 7,192 | alex-krash | @qiaochaojie ,
ALTER TABLE DROP PARTITION
1. Synchronous request
2. Deals only with the data, matching condition predicate (e.g. you have 1 year of data, partitioned by day, and want to drop a single day - only this day will be affected, but not others). No read is performed, just deletion.
ALTER TABLE ... DELET... | 2019-10-04T09:14:02 | 2019-10-04T09:14:02 | {
"+1": 4
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 538,351,139 | 7,192 | qiaochaojie | @alex-krash I use clickhouse-client too. Below is the version information
ClickHouse client version 19.7.3.1.
Connecting to 127.0.0.1:9000 as user default.
Connected to ClickHouse server version 19.7.3 revision 54419 | 2019-10-04T10:58:16 | 2019-10-04T10:58:16 | {} | NONE |
ClickHouse | ClickHouse | 538,686,987 | 7,200 | azat | It will be great to have a shared+unbundled build in your CI, but this will require all libraries in the image
P.S. I cannot find the build matrix in the repo, so I guess this is not part of the repository | 2019-10-05T20:43:41 | 2019-10-05T20:43:55 | {} | MEMBER |
ClickHouse | ClickHouse | 536,577,156 | 7,147 | den-crane | select arrayConcat([arr[1]], arrayFilter(x,y -> x = y, arraySlice(arr, 2), arraySlice(arr, 1, -1))) from ( select [1, 2, 2, 3, 3, 2, 2] arr )
SELECT arrayConcat([arr[1]], arrayFilter(x,y -> y!=0, arr ,arrayDifference(arr))) from (select [1, 2, 2, 3, 3, 2] arr)
| 2019-09-30T14:05:52 | 2019-09-30T14:05:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 537,249,809 | 7,156 | alexey-milovidov | > today(), yesterday(), randConstant()
I don't think I have listed all the relevant functions. | 2019-10-01T21:55:40 | 2019-10-01T21:55:40 | {} | MEMBER |
ClickHouse | ClickHouse | 538,450,332 | 7,195 | Kiennh | Thank you, it's working.
Ps: I just use Log engine for reproduce this issue. | 2019-10-04T15:38:05 | 2019-10-04T15:38:05 | {} | NONE |
ClickHouse | ClickHouse | 538,648,310 | 7,174 | filimonov | Fixed in #6781. Thanks @tavplubix! | 2019-10-05T13:08:14 | 2019-10-05T13:08:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 538,678,794 | 7,200 | alexey-milovidov | Ok. PS. Don't understand why it doesn't fail in our CI. | 2019-10-05T18:54:11 | 2019-10-05T18:54:11 | {} | MEMBER |
ClickHouse | ClickHouse | 535,979,707 | 6,541 | dimarub2000 | Currently working with yandex environment. | 2019-09-27T15:08:14 | 2019-09-30T18:17:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 536,725,319 | 7,141 | filimonov | > > AFAIK that story is in the backlog, but I'm not sure about terms and schedule.
> > @alexey-milovidov can you clarify that?
>
> Full featured distributed on distributed tables are not in our backlog, but it isn't hard thing to do.
Hm. AFAIR @alex-zaitsev was mentioning that. There are few use cases when it ca... | 2019-09-30T19:55:37 | 2019-09-30T19:55:37 | {
"rocket": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 536,729,576 | 7,124 | filimonov | Hard to follow.
> and when i executed sql this,and all data has been deleted
> `alter table test_01 delete where id in (select id from test_01)`
I would also say it should delete everything :)
Can you please provide some simple/synthetic example? We need simple table(s) definition, some data and queries whic... | 2019-09-30T20:07:21 | 2019-09-30T20:07:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 537,247,977 | 7,140 | alexey-milovidov | @stavrolia It isn't a bug fix, so don't worry. | 2019-10-01T21:50:38 | 2019-10-01T21:50:38 | {} | MEMBER |
ClickHouse | ClickHouse | 537,467,232 | 7,174 | filimonov | Same for http:
```
echo 'SELECT {foo:String}, {answer:UInt8};' | curl 'http://localhost:8123/?param_foo=bar¶m_answer=42' --data-binary @-
bar 42
echo '({foo:String}, {answer:UInt8})' | curl 'http://localhost:8123/?query=INSERT%20INTO%20i%20VALUES¶m_foo=bar¶m_answer=42' --data-binary @-
Code: 456, e.d... | 2019-10-02T12:21:56 | 2019-10-02T12:29:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 538,073,454 | 7,173 | alesapin | Fixed here https://github.com/ClickHouse/ClickHouse/pull/7185 | 2019-10-03T18:41:14 | 2019-10-03T18:41:14 | {} | MEMBER |
ClickHouse | ClickHouse | 538,165,607 | 7,187 | lorenz | It makes the first query that was already fast around 30% faster, but doesn't do anything for the already slow one. | 2019-10-03T23:19:49 | 2019-10-03T23:19:49 | {} | NONE |
ClickHouse | ClickHouse | 536,627,687 | 6,817 | alesapin | @amosbird, thank you https://github.com/ClickHouse/ClickHouse/pull/7134! | 2019-09-30T15:54:54 | 2019-09-30T15:54:54 | {} | MEMBER |
ClickHouse | ClickHouse | 536,627,718 | 7,113 | akuzm | Thanks for the data, I could reproduce it locally, will investigate. | 2019-09-30T15:54:57 | 2019-09-30T15:54:57 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 537,440,264 | 6,781 | abyss7 | Reviewed all changes and discussed all the questions in person. Awaiting for some of fixes:
- Add tests to cover all gaps in the code, when template-parsing was enabled by default.
- Fix behavior for case like `VALUES (+1) (+1) (+1) …`.
- Move couple of methods to the base classes `IFunction`, `IFunctionBuilder`, et... | 2019-10-02T10:52:52 | 2019-10-02T10:52:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 537,865,887 | 7,180 | jpiper | Looking at this, it appears to me the DDL could be incomplete? There is no `PARTITION BY` cause, nor is the `Date` column passed as a third argument to `ReplicatedMergeTree`. Could this startup issue be because an older version of clickhouse has allowed a user to create a bad table? | 2019-10-03T09:28:00 | 2019-10-03T09:28:00 | {} | NONE |
ClickHouse | ClickHouse | 537,985,570 | 6,781 | abyss7 | Can be merged after performance tests are OK. | 2019-10-03T15:03:47 | 2019-10-03T15:03:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 538,076,735 | 7,184 | feribg | Thanks @den-crane . And I assume the operator is always `table_1.asof_column >= table2.asof_column` gt or equal, you can't enforce strict inequality or the direction of inequality? | 2019-10-03T18:50:01 | 2019-10-03T18:50:34 | {} | NONE |
ClickHouse | ClickHouse | 538,418,554 | 7,197 | anilxoxo | > Please, show the output
> apt-get -y remove clickhouse*
Not working.unable to remove file inside the **preprocessed** | 2019-10-04T14:23:07 | 2019-10-04T14:23:07 | {} | NONE |
ClickHouse | ClickHouse | 538,651,510 | 7,192 | den-crane | @qiaochaojie
> I have a question. when I use the command like 'alter table xx on cluter drop partition ', it just returned immediately and did not report any error. after a few minutes, I find the partition still exists. It seems like not synchronous
Such behavior is observed when one makes a mistake in partition... | 2019-10-05T13:49:46 | 2019-10-05T13:52:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 538,679,677 | 7,200 | azat | >Otherwise unbundled build fail with:
Egh, sorry, s/unbundled/`SPLIT_SHARED_LIBRARIES`/
>Ok. PS. Don't understand why it doesn't fail in our CI.
Do you test `SPLIT_SHARED_LIBRARIES` build on your CI? | 2019-10-05T19:06:19 | 2019-10-05T19:06:19 | {} | MEMBER |
ClickHouse | ClickHouse | 536,704,138 | 7,156 | alexey-milovidov | We have more than one such function. What about today(), yesterday(), randConstant()... | 2019-09-30T19:01:30 | 2019-09-30T19:01:30 | {} | MEMBER |
ClickHouse | ClickHouse | 538,423,171 | 7,196 | filimonov | 1) 2019.10.04 14:37:02.051591 we get the POST request, and log that here:
https://github.com/ClickHouse/ClickHouse/blob/86bee0b8b5d9044a5c1070bb7155c70fe9e9ffac/dbms/programs/server/HTTPHandler.cpp#L223 we start measuring query execution time.
2) we don't know yet what kind of query is that. We also don't know if fu... | 2019-10-04T14:34:23 | 2019-10-04T15:39:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 538,682,528 | 7,196 | inv2004 | @filimonov
It is logical how it CH works, but, I suppose, it would be more useful if query_duration_ms (or another one) would measure real time of query execution which period is from step 5 to 6 in your list.
It would help to find if query slow or fast on CH side without any dependencies on client-side or connect... | 2019-10-05T19:44:08 | 2019-10-05T20:16:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 536,734,576 | 7,117 | filimonov | Please start debugging with `isql` command line tool (provided as a part of unixODBC): https://www.easysoft.com/blog/isql-connection-string.html
About mssql: not all authentication options work properly from linux - check that:
* https://stackoverflow.com/a/57732870
* https://askubuntu.com/q/167491
* https://docs... | 2019-09-30T20:20:39 | 2019-09-30T20:20:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 536,977,867 | 7,156 | KochetovNicolai | Still doesn't fix :( | 2019-10-01T10:40:10 | 2019-10-01T10:40:10 | {} | MEMBER |
ClickHouse | ClickHouse | 537,239,822 | 7,167 | alexey-milovidov | It isn't the changelog for 19.15. | 2019-10-01T21:32:21 | 2019-10-01T21:32:21 | {} | MEMBER |
ClickHouse | ClickHouse | 537,587,452 | 7,176 | Dlougach | The workaround, by the way, is to explicitly make `toInt32(0)` in `SELECT`. | 2019-10-02T17:02:51 | 2019-10-02T17:02:51 | {} | NONE |
ClickHouse | ClickHouse | 537,870,152 | 7,180 | alexey-milovidov | BTW, The error `The local set of parts of table mydatabase.my_table_shard doesn't look like the set of parts in ZooKeeper: 573.73 thousand rows of 573.73 thousand total rows in filesystem are suspicious. There are 1 unexpected parts with 573733 rows (1 of them is not just-written with 573733 rows), 0 missing parts (wit... | 2019-10-03T09:40:07 | 2019-10-03T09:40:07 | {} | MEMBER |
ClickHouse | ClickHouse | 538,183,670 | 7,187 | lorenz | Thanks! Your query is even a bit faster than my first one, but I have to add the language constraint to the outer query otherwise I get wrong results sometimes.
As for the compounding factors:
* I'm pretty sure that DoubleDelta does in fact compress the data (446MiB column size for 3.5B rows -> 0.1 byte per row), but... | 2019-10-04T00:47:46 | 2019-10-04T00:47:46 | {} | NONE |
ClickHouse | ClickHouse | 538,195,409 | 6,984 | ghost | @alesapin @alexey-milovidov
I am using ClickHouse in 30 machines of cluster and now considering for expanding to 90 machines for the new features.
I need the bloomfilter for Array for the new features.
Let me know what I should do for merging this PR.
Thank you. | 2019-10-04T01:49:39 | 2019-10-04T01:51:22 | {} | NONE |
ClickHouse | ClickHouse | 538,324,992 | 7,192 | qiaochaojie | @alex-krash Thanks for your answer! I have a question. when I use the command like 'alter table xx on cluter drop partition ', it just returned immediately and did not report any error. after a few minutes, I find the partition still exists. It seems like not synchronous | 2019-10-04T09:31:16 | 2019-10-04T09:31:16 | {} | NONE |
ClickHouse | ClickHouse | 538,393,822 | 7,197 | den-crane | Please, show the output
apt-get -y remove clickhouse* | 2019-10-04T13:20:33 | 2019-10-04T13:20:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 538,681,443 | 7,200 | alexey-milovidov | Yes, it's second from the bottom: https://clickhouse-builds.s3.yandex.net/7200/4b9d711f0765b7f49f461204b71542671aefc56a/report.html | 2019-10-05T19:30:26 | 2019-10-05T19:30:26 | {} | MEMBER |
ClickHouse | ClickHouse | 536,498,360 | 7,152 | alexey-milovidov | Otherwise mutation cannot proceed. | 2019-09-30T10:16:39 | 2019-09-30T10:16:39 | {} | MEMBER |
ClickHouse | ClickHouse | 537,868,284 | 7,180 | jpiper | Ah, I have found that on the old version of ClickHouse the table is not attached after startup (and running `ATTACH table XXX.XXX` gives the same error, but ClickHouse still starts up. So it appears it is a behaviour change.
`19.11` -> fail to attach table, carry on booting without this table
`19.13+` -> fail to at... | 2019-10-03T09:34:52 | 2019-10-03T09:34:52 | {} | NONE |
ClickHouse | ClickHouse | 537,870,710 | 7,180 | jpiper | Gotcha, so it somehow the zookeeper data for this replicated table has been deleted even though the data persist on the disk. I think I was just thrown off by the fact this panic was only happening on newer versions, so was assuming it was some sort of bug | 2019-10-03T09:41:54 | 2019-10-03T09:41:54 | {} | NONE |
ClickHouse | ClickHouse | 538,682,358 | 6,448 | kreuzerkrieg | > Any ideas whats about Darwin build? Looks like its failed to start
>
> It will fix after you merge with master.
Already fixed | 2019-10-05T19:41:33 | 2019-10-05T19:41:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 538,686,091 | 7,200 | azat | So, first of all, #7063 fixed this already, and now `${Poco_XML_LIBRARY}` linked twice, so this PR should be reverted.
Secondly, that build uses bundled poco, and I guess that somewhere there is XML linked inside bundled poco, since I tried unbundled+shared (and installing libpoco-dev before) and it fails in yandex/cl... | 2019-10-05T20:30:00 | 2019-10-05T20:30:00 | {} | MEMBER |
ClickHouse | ClickHouse | 536,493,428 | 7,141 | alexey-milovidov | > AFAIK that story is in the backlog, but I'm not sure about terms and schedule.
@alexey-milovidov can you clarify that?
Full featured distributed on distributed tables are not in our backlog, but it isn't hard thing to do. | 2019-09-30T10:01:44 | 2019-09-30T10:01:44 | {} | MEMBER |
ClickHouse | ClickHouse | 537,411,967 | 6,742 | akuzm | We should add an error message about unsupported delimiters, and maybe mention CustomSeparated.
It's just so annoying when a simple thing breaks for unknown reasons with a cryptic message, and the user has no idea what's wrong and how it can be fixed. | 2019-10-02T09:24:01 | 2019-10-02T09:24:01 | {
"heart": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 537,413,625 | 6,526 | MrBuddyCasino | Virtual column handling was changed in those commits, can't spot an obvious error though:
https://github.com/ClickHouse/ClickHouse/commit/cf39c4cc473487650c5848004f2ab41aaacd5fc7#diff-e07f6e9213f722fdd54782f5934373aa
https://github.com/ClickHouse/ClickHouse/commit/7b6c78402d058544905225e2c23c496344d8baf2#diff-e07f6e9... | 2019-10-02T09:28:46 | 2019-10-02T09:28:46 | {} | NONE |
ClickHouse | ClickHouse | 537,866,425 | 7,180 | alexey-milovidov | It is protection from configuration errors.
Could you please tell how exactly you are performing version upgrade?
Do you change configuration file? (For example, if you change `shard` macro, the replica will be attached to wrong shard).
> Looking at this, it appears to me the DDL could be incomplete? There is no... | 2019-10-03T09:29:40 | 2019-10-03T09:29:40 | {} | MEMBER |
ClickHouse | ClickHouse | 538,067,204 | 7,184 | den-crane | ASOF JOIN was implemented to solve this particular problem.
https://clickhouse.yandex/docs/en/query_language/select/
ASOF JOIN is useful when you need to join records that have no exact match. | 2019-10-03T18:24:58 | 2019-10-03T18:24:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 538,169,516 | 7,187 | lorenz | Sure:
```
[green] 2019.10.03 23:37:53.986197 {cd510939-ded6-48f7-9f22-8198d3a6193c} [ 70 ] <Debug> executeQuery: (from 127.0.0.1:37212) SELECT repository_id, location, lines FROM files PREWHERE language = 'Plain Text' ORDER BY lines DESC LIMIT 1
↑ Progress: 0.00 rows, 0.00 B (0.00 rows/s., 0.00 B/s.) [green] 2019.10... | 2019-10-03T23:39:06 | 2019-10-03T23:39:06 | {} | NONE |
ClickHouse | ClickHouse | 538,346,464 | 7,192 | alex-krash | @qiaochaojie , I've just tried dropping on cluster via ```clickhouse-client```, and it was synchronous.
What connector are you using? | 2019-10-04T10:41:11 | 2019-10-04T10:41:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 538,796,709 | 7,177 | alexey-milovidov | Yes, ClickHouse is using shared-nothing architecture.
We have a plan to add support for shared storage as an option without changing overall architecture.
Shared storage can provide benefits in scalability and lower maintenance costs only if it is tightly integrated in some cloud service (BigQuery, Snowflake). But ... | 2019-10-06T22:58:54 | 2019-10-06T22:59:17 | {} | MEMBER |
ClickHouse | ClickHouse | 537,551,572 | 6,661 | abyss7 | The reason behind this behavior is trivial. When we insert into Kafka table, the MV pushes new rows to corresponding tables (standard CH feature), but then the same MV reads these messages from Kafka server and pushes them second time.
Is it good enough for you to have separate Kafka tables? Does it break your scena... | 2019-10-02T15:37:55 | 2019-10-02T15:37:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 537,815,195 | 6,661 | arruzk | @abyss7 during my tests Kafka topic didn't contain the same data as I inserted manually. It's not a big problem for me, I just used temporary table (insert in a temporary table and move data manually) in my scenario, just wanted to avoid one intermediate step. | 2019-10-03T07:00:48 | 2019-10-03T07:00:48 | {} | NONE |
ClickHouse | ClickHouse | 537,822,681 | 6,897 | slimtom95 | Thanks for the hints on `'\n'` delimitter, and it's a good idea to mention it in the docs.
I'll retest these settings with `'\n'` when I get back to work, let's just keep this thread open a few more days.. | 2019-10-03T07:25:44 | 2019-10-03T07:25:44 | {} | NONE |
ClickHouse | ClickHouse | 538,164,407 | 7,187 | den-crane | try to replace `where` with `prewhere`
```
SELECT
repository_id,
location,
lines
FROM files
PREWHERE language = 'Plain Text'
ORDER BY lines DESC
LIMIT 1
``` | 2019-10-03T23:14:32 | 2019-10-03T23:14:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 538,397,320 | 7,195 | den-crane | Workaround:
```
`.inner.a2`
CREATE TABLE a1(label String) engine=Log ;
CREATE MATERIALIZED VIEW a2 engine=Log as select * from a1;
CREATE MATERIALIZED VIEW a3 engine=Log as select * from `.inner.a2`;
CREATE MATERIALIZED VIEW a4 engine=Log as select * from `.inner.a3`;
```
PS: don't use `engine=Log` as ... | 2019-10-04T13:30:31 | 2019-10-04T13:44:56 | {
"+1": 3
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 537,468,223 | 7,174 | filimonov | #5331 #5630 | 2019-10-02T12:24:48 | 2019-10-02T12:24:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 537,490,213 | 7,173 | den-crane | How you run CH? Docker ? Windows subsystem linux? | 2019-10-02T13:24:28 | 2019-10-02T13:24:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 537,672,726 | 6,448 | kreuzerkrieg | Any ideas whats about Darwin build? Looks like its failed to start
| 2019-10-02T20:44:56 | 2019-10-02T20:44:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 537,689,074 | 3,822 | josepowera | Closing since this is problem on too old version. We haven't seen this problem in 2019. | 2019-10-02T21:28:29 | 2019-10-02T21:28:29 | {} | NONE |
ClickHouse | ClickHouse | 537,862,933 | 5,438 | alexey-milovidov | > But I don't know how to make them.
Integration tests are located in `dbms/tests/integration`.
You can look for similar tests. | 2019-10-03T09:20:09 | 2019-10-03T09:20:09 | {} | MEMBER |
ClickHouse | ClickHouse | 538,424,972 | 7,197 | den-crane | apt-get remove -y --force-yes clickhouse* | 2019-10-04T14:39:00 | 2019-10-04T14:39:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 536,787,963 | 7,141 | alex-zaitsev | The rationale is actually the different. If there are 100-1000 nodes in the cluster, it could be quite time consuming to update the configuration on all nodes if there are any changes (adding/replacing a node and so on). So having distributed tables pointing to sub-clusters, instead of individual nodes, would made it e... | 2019-09-30T23:06:31 | 2019-09-30T23:06:31 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 537,034,706 | 6,742 | tavplubix | It can be done with `CustomSeparated` format:
```
$ echo "1 2" | clickhouse-local -S "a int, b int" --input-format=CustomSeparated --format_custom_escaping_rule=CSV --format_custom_field_delimiter=' ' -q "select * from table"
1 2
``` | 2019-10-01T13:25:20 | 2019-10-01T13:25:20 | {
"+1": 3
} | MEMBER |
ClickHouse | ClickHouse | 537,249,139 | 7,156 | alexey-milovidov | Missing test. | 2019-10-01T21:53:28 | 2019-10-01T21:53:28 | {} | MEMBER |
ClickHouse | ClickHouse | 538,318,225 | 7,192 | qiaochaojie | the ch version is 19.7.3 | 2019-10-04T09:11:25 | 2019-10-04T09:11:25 | {} | NONE |
ClickHouse | ClickHouse | 538,419,346 | 5,576 | SaltTan | A workaround for a distributed query is to put the values into a temporary table (for distributed query processing, temporary tables used in a query are passed to remote servers)
```
:) create temporary table tt as select today() as d;
:) select count() from system.query_log where event_date in tt settings force_p... | 2019-10-04T14:25:06 | 2019-10-04T14:25:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 538,503,136 | 7,156 | KochetovNicolai | Now actions from header are reused. Should fix for all functions. | 2019-10-04T18:05:15 | 2019-10-04T18:05:15 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 538,792,282 | 7,197 | alexey-milovidov | When you uninstall database server, the data (including preprocessed files) remains. This is normal. | 2019-10-06T21:59:47 | 2019-10-06T21:59:47 | {} | MEMBER |
ClickHouse | ClickHouse | 538,856,149 | 7,147 | Tasselmi | > select arrayConcat([arr[1]], arrayFilter(x,y -> x = y, arraySlice(arr, 2), arraySlice(arr, 1, -1))) from ( select [1, 2, 2, 3, 3, 2, 2] arr )
>
> SELECT arrayConcat([arr[1]], arrayFilter(x,y -> y!=0, arr ,arrayDifference(arr))) from (select [1, 2, 2, 3, 3, 2] arr)
Thanks for your help ~ | 2019-10-07T06:23:20 | 2019-10-07T06:23:20 | {} | NONE |
ClickHouse | ClickHouse | 536,985,152 | 7,156 | KochetovNicolai | Now won't fail.
Can't solve it better for now. | 2019-10-01T11:02:45 | 2019-10-01T11:02:45 | {} | MEMBER |
ClickHouse | ClickHouse | 537,388,063 | 7,165 | ifoo | @filimonov thanks. i'll renamw the issue to missing documentation. | 2019-10-02T08:17:28 | 2019-10-02T08:31:21 | {} | NONE |
ClickHouse | ClickHouse | 538,166,782 | 7,187 | den-crane | can you show the output of slow query with debug info
set send_logs_level='trace';
...SELECT ... | 2019-10-03T23:25:35 | 2019-10-03T23:25:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 538,682,200 | 6,448 | alexey-milovidov | > Any ideas whats about Darwin build? Looks like its failed to start
It will fix after you merge with master. | 2019-10-05T19:39:39 | 2019-10-05T19:39:39 | {} | MEMBER |
ClickHouse | ClickHouse | 538,819,732 | 7,204 | alexey-milovidov | You can also look at `FunctionArrayMapped.h` and `arrayFilter.cpp`.
You can make this function as a higher order function similar to `arrayMap` and `arrayFiler`. | 2019-10-07T02:37:06 | 2019-10-07T02:37:06 | {} | MEMBER |
ClickHouse | ClickHouse | 536,690,582 | 7,157 | den-crane | postgresql:
```
> select sum(x) from (select cast(Null as int) x) as t;
sum
-----
(1 row)
> select sum(x) from (
select cast(Null as int) x
union all select 0) as t;
sum
-----
0
(1 row)
```
mysql
```
mysql> select sum(x) from (select Null as x) t;
+--------+
| sum(x) |
+------... | 2019-09-30T18:28:47 | 2019-09-30T18:35:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 537,018,247 | 6,897 | tavplubix | Skipping errors in `JSONEachRow` works only if rows are delimited by `\n`, because parser skips all characters until the new line (or until EOF if there is no `\n`) to find the beginning of next row. For queries like
```
insert into default.test_input_ignore_error format JSONEachRow {} {2} {} {} {}
```
parser igno... | 2019-10-01T12:42:41 | 2019-10-01T12:42:41 | {} | MEMBER |
ClickHouse | ClickHouse | 537,066,217 | 7,155 | alesapin | Failed build doesn't connected with changes | 2019-10-01T14:35:39 | 2019-10-01T14:35:39 | {} | MEMBER |
ClickHouse | ClickHouse | 537,684,340 | 7,172 | alexey-milovidov | Perfect. | 2019-10-02T21:15:12 | 2019-10-02T21:15:12 | {} | MEMBER |
ClickHouse | ClickHouse | 538,174,102 | 7,187 | den-crane | > `repository_id` UInt32 CODEC(DoubleDelta),
This is probably should by CODEC(DoubleDelta, LZ4))
Because DoubleDelta does not do compression.
It seems the reason is slow decompression of ZSTD(1). And CH decompresses repository_id,location,lines before order by limit
Try this for a test (to check my guess) :
... | 2019-10-04T00:01:44 | 2019-10-04T00:01:44 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 538,680,392 | 7,203 | alexey-milovidov | PS. Here is an almost exhaustive list of similar tools: https://news.ycombinator.com/item?id=20848581 | 2019-10-05T19:15:17 | 2019-10-05T19:15:17 | {} | MEMBER |
ClickHouse | ClickHouse | 539,735,429 | 7,221 | alexey-milovidov | > replacing hashtable with i.e. sparse_hash_set
Not an option because it will become extremely slow.
> replacing grower to fixed will use maximum memory even for small number of elements
Not an option because it will defeat the whole point of combined data structure.
(It is needed for the typical case of GROU... | 2019-10-08T22:46:42 | 2019-10-08T22:47:19 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.