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
522,723,394
6,552
alexey-milovidov
Also you can change ``` using MYSQL = st_mysql; ``` to ``` using MYSQL = struct st_mysql; ``` and so on.
2019-08-19T19:36:59
2019-08-19T19:37:07
{}
MEMBER
ClickHouse
ClickHouse
522,808,842
6,514
alexey-milovidov
Fixed in 99aa5cce3895d6aa30990da54ca2b1f46e448782
2019-08-20T00:55:07
2019-08-20T00:55:07
{}
MEMBER
ClickHouse
ClickHouse
522,896,461
6,425
KochetovNicolai
@vzakaznikov I've just merged branch with master. Let's wait for tests again and I will merge the pr if they pass.
2019-08-20T07:44:56
2019-08-20T07:44:56
{}
MEMBER
ClickHouse
ClickHouse
522,975,225
6,561
filimonov
> @BayoNet do you have a ticket for this? Actually @nikvas0 already have added some documentation for that. :) (See link above).
2019-08-20T11:41:03
2019-08-20T11:41:03
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,997,745
6,504
filimonov
Client library doesn't support that particular checksums? https://github.com/apache/hawq/blob/4156a8760787894ef351de73df3c38287bbba41d/depends/libhdfs3/src/client/RemoteBlockReader.cpp#L148
2019-08-20T12:46:50
2019-08-20T12:53:01
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,729,850
6,552
rafaeldtinoco
Alright, let me propose a merge then. Thx for the advices.
2019-08-19T19:56:50
2019-08-19T19:56:50
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,846,278
6,535
Sundarcj
which one will be efficient dictGet or Join. i don't have knowledge about dict in CH. Is dictGet can be used in join Engine?
2019-08-20T04:18:32
2019-08-20T04:18:32
{}
NONE
ClickHouse
ClickHouse
522,956,529
6,563
akuzm
It tries to print the error: `There is no supertype for types Array(UInt64), Nullable(Nothing) because some of them are Array and some of them are not (version 19.13.1.1)`, but the error printing code is broken so it fails. The errors means that it cannot deduce the return type for multiIf, because its third argument ...
2019-08-20T10:36:43
2019-08-20T10:36:43
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,998,632
6,560
rafaeldtinoco
This has also been opened by clickhouse Debian maintainer: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91500
2019-08-20T12:49:15
2019-08-20T12:49:15
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,578,734
6,535
filimonov
That is happening when you do `ALTER TABLE ... UPDATE colunmname=... WHERE ....` There is also `ALTER TABLE tablename CLEAR COLUMN colunmname IN PARTITION ... `
2019-08-19T13:36:32
2019-08-19T13:36:32
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,597,607
6,528
alex-krash
@filimonov , got it, we are going to implement some kind of OVER () section from window functions.
2019-08-19T14:22:38
2019-08-19T14:22:38
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,658,482
6,468
alexey-milovidov
Will be added to docs.
2019-08-19T16:43:12
2019-08-19T16:43:12
{}
MEMBER
ClickHouse
ClickHouse
522,792,263
6,523
alexey-milovidov
Also one thought: why don't we clone `queue` and `parts` from source replica, totally replacing our `queue` and `parts`? Maybe because it's more difficult to do atomically?
2019-08-19T23:27:43
2019-08-19T23:27:43
{}
MEMBER
ClickHouse
ClickHouse
522,957,145
6,563
yingfeng
Got it, thank you~
2019-08-20T10:38:58
2019-08-20T10:38:58
{}
NONE
ClickHouse
ClickHouse
522,994,465
6,506
filimonov
Hm. Even if master is dead and need to be recovered? As a workaround: i think that if you will forbid inserts to the table (by using readonly profile for users), close the access to 9009 port from outside, and set `replicated_can_become_leader` to 0 on that replica, you will get something very similar to what you n...
2019-08-20T12:37:31
2019-08-20T12:38:32
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,479,176
6,517
4ertus2
I don't think it's a bug. It's easy to remove nullability from column, ex. with ```assumeNotNull(url)``` that should solve your problem. There're also ifNull() and coalesce() functions.
2019-08-19T09:00:27
2019-08-19T09:00:27
{ "+1": 1 }
CONTRIBUTOR
ClickHouse
ClickHouse
522,722,958
6,552
alexey-milovidov
> Not sure if this would be the best way to address this, any recommendation? If it works both for libmysqlclient and libmariadbclient, it's fine.
2019-08-19T19:35:42
2019-08-19T19:35:42
{}
MEMBER
ClickHouse
ClickHouse
522,724,911
6,554
alexey-milovidov
Approved. But the implementation is incorrect. `numbers_mt` will not scale uniformly due to lack of work stealing. It needs rewrite.
2019-08-19T19:41:40
2019-08-19T19:41:40
{}
MEMBER
ClickHouse
ClickHouse
522,745,693
5,846
josepowera
Running 19.13.2.19 looks OK.
2019-08-19T20:42:20
2019-08-19T20:42:20
{}
NONE
ClickHouse
ClickHouse
522,888,954
6,561
vmarkovtsev
Thanks! This magic command is worth mentioning somewhere near `ADD INDEX` docs :smile:
2019-08-20T07:22:09
2019-08-20T07:22:19
{ "+1": 4 }
NONE
ClickHouse
ClickHouse
522,587,039
6,538
alexey-milovidov
Ok.
2019-08-19T13:56:43
2019-08-19T13:56:43
{}
MEMBER
ClickHouse
ClickHouse
522,647,744
6,532
kreuzerkrieg
@alexey-milovidov is there any chance the system starts to back-off in background pool once number of threads starts to increase? with 32 threads parts fetch starts ~5 minutes after server starts, with 64 threads after ~10 minutes. Though I didnt have a chance to debug it properly. Ideas?
2019-08-19T16:15:02
2019-08-19T16:15:02
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,812,525
6,514
alexey-milovidov
Deadlock: ``` ┌───────elapsed─┬─query───────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┬─query_id─────────────────────────────┐ │ 789.408521682 │ OPTIMIZE TABLE alter_table_3 FINAL ...
2019-08-20T01:15:41
2019-08-20T01:15:41
{}
MEMBER
ClickHouse
ClickHouse
522,827,880
6,560
rafaeldtinoco
I just opened this issue to track it for the future (for you). I'll probably include a mitigation now for our build servers.
2019-08-20T02:35:07
2019-08-20T02:35:07
{ "+1": 1 }
CONTRIBUTOR
ClickHouse
ClickHouse
522,879,376
5,237
filimonov
It may be related to https://github.com/yandex/ClickHouse/issues/6349
2019-08-20T06:49:27
2019-08-20T06:49:27
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,591,875
6,299
MahmoudGoda0
@alexey-milovidov I really appreciate if you answer my above 2 questions.
2019-08-19T14:08:42
2019-08-19T14:08:42
{}
NONE
ClickHouse
ClickHouse
522,607,443
6,425
KochetovNicolai
@vzakaznikov ok. I will wait till you fix tests. I've already rewritten four, but hope it's not a problem :)
2019-08-19T14:44:57
2019-08-19T14:44:57
{}
MEMBER
ClickHouse
ClickHouse
522,704,564
6,473
Akazz
ClickHouse returns `HTTP/1.1 200 OK` in case the insertion is successful. In all of the test setups that I've tried your example query on the insertion has been successful - data gets inserted every time. As to the `HTTP/1.1 400 Bad Request` message, I'd suggest that this is due to misformed data payload: your http ...
2019-08-19T18:43:10
2019-08-19T18:43:39
{ "+1": 1 }
CONTRIBUTOR
ClickHouse
ClickHouse
522,827,607
6,560
rafaeldtinoco
It could be related to the architecture intrinsic as I'm currently using an old AMD CPU: model name : AMD FX(tm)-8350 Eight-Core Processor stepping : 0 microcode : 0x6000852 But I'll let upstream to investigate as I can simply mitigate this using: ``` [PATCH] G++ 9 crashes returning except...
2019-08-20T02:33:49
2019-08-20T02:34:19
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,831,170
6,552
rafaeldtinoco
Im testing our Ubuntu Builds and will propose a patch tomorrow (basically a #ifndef _mysql_h before the declarations, seem to have worked pretty well with minimum changes).
2019-08-20T02:51:57
2019-08-20T02:51:57
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,876,359
6,104
filimonov
> We can replace currentDatabase() call with actual value in AddDefaultDatabaseVisitor. For distributed DDL - it looks like the best option. Also it looks like it should resolve most of the related issues. Some extra thoughts: 1) original DDL statement will not be preserved (I don't think it's very important) ...
2019-08-20T06:38:33
2019-08-20T06:39:54
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,954,870
6,461
4ertus2
It's limited by query memory limit. So you are able to use as much data there as you can allocate in your query context. ``` create table test.t (x Int64) engine = Set; insert into test.t select number from system.numbers; ``` ``` ↘ Progress: 33.55 million rows, 268.44 MB (2.48 million rows/s., 19.82 MB/s.) R...
2019-08-20T10:31:12
2019-08-20T10:33:14
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,957,659
215
philodavies
What was the status of this issue? It got closed, but I don't seem to be able to insert into a table with the int8 value of an enum.
2019-08-20T10:40:48
2019-08-20T10:40:48
{}
NONE
ClickHouse
ClickHouse
522,965,470
6,488
filimonov
Let's continue in #6506 to avoid duplicates.
2019-08-20T11:08:33
2019-08-20T11:08:33
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,511,234
5,680
filimonov
@spyros87 can you check 19.11.7?
2019-08-19T10:32:27
2019-08-19T10:32:27
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,656,208
6,532
alexey-milovidov
Do you mean total number of threads in background pool or the number of active threads (that are doing merges or fetches)? You can check the status of fetching parts (what parts are currently fetched) in `system.replication_queue` table.
2019-08-19T16:37:09
2019-08-19T16:37:09
{}
MEMBER
ClickHouse
ClickHouse
522,723,258
6,552
alexey-milovidov
You can send a pull request with this change and it will be tested automatically.
2019-08-19T19:36:35
2019-08-19T19:36:35
{}
MEMBER
ClickHouse
ClickHouse
522,741,711
6,514
alexey-milovidov
Fixed ``` ================== WARNING: ThreadSanitizer: data race (pid=257651) Read of size 8 at 0x7b2c00130d10 by thread T76: #0 boost::multi_index::detail::sequenced_index<boost::multi_index::detail::nth_layer<1, DB::ColumnDescription, boost::multi_index::indexed_by<boost::multi_index::sequenced<boost::mult...
2019-08-19T20:31:14
2019-08-19T20:31:14
{}
MEMBER
ClickHouse
ClickHouse
522,970,636
6,506
astudnev
No, this is not readonly. Others must NOT read from this replica
2019-08-20T11:25:54
2019-08-20T11:25:54
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,507,396
6,534
Sundarcj
For first use case, the thing is in my table user can transform any column
2019-08-19T10:21:01
2019-08-19T10:21:01
{}
NONE
ClickHouse
ClickHouse
522,510,306
6,425
vzakaznikov
@KochetovNicolai, you are right SIGCHLD has nothing to do with it. I confused it with SIGHUP which clickhouse-client actually does not ignore. We should try killing bash with SIGHUP so that it re-sends it to all its children. https://www.gnu.org/software/bash/manual/html_node/Signals.html#Signals > The shell exi...
2019-08-19T10:29:33
2019-08-19T10:29:33
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,850,195
6,549
thyn
> On second question - can you grep logs for strings 20190729_69485_124392_9_196813 and 20190729_110928_124392_3_196816 and attach the results? ` 2019.08.19 11:09:17.061134 [ 54 ] {47ccdf53-d8ad-48d2-b846-7123bbc15559} <Error> executeQuery: Code: 49, e.displayText() = DB::Exception: Part 20190729_69485_124392_9_1...
2019-08-20T04:41:07
2019-08-20T04:41:19
{}
NONE
ClickHouse
ClickHouse
522,960,787
6,489
alesapin
Looks like code is ready. Now I'll try to write tests for all introduced features.
2019-08-20T10:51:57
2019-08-20T10:51:57
{}
MEMBER
ClickHouse
ClickHouse
522,997,563
6,565
4ertus2
How much parts are in your v_url_user_category_h_2 201907 partition (select count() from system.parts where partition_id = '201907' and active)? You may OPTIMIZE v_url_user_category_h_2 to make one part for the partition and recheck query. There's also different partitioning for source and MV tables. It could make s...
2019-08-20T12:46:20
2019-08-20T12:50:45
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,469,885
6,528
alex-krash
My PR on `neighbour()` function is not merged yet, so we may apply some additional logic here. But it is not clear for me, what behavior you expect? Could you, please, provide an example of input-output?
2019-08-19T08:35:01
2019-08-19T08:35:01
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,576,650
6,534
4ertus2
The general OLAP solution is to have big fact table that is immutable and small tables that could be joined to it. It also called as star schema. Then you are able to change joined tables as needed. If you have lots of mutations on the fact table or very big dimensions it's something strange and you need some special s...
2019-08-19T13:31:31
2019-08-19T13:31:31
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,582,391
6,425
vzakaznikov
@KochetovNicolai, no I don't mind but I think we are very close to making it solid. I will provide you a diff by the end of the day that should fix this last issue and all tests should be good to go. The small fix is in the reader method and change "except OSError" to just "except" so that all exceptions are ignored ...
2019-08-19T13:45:42
2019-08-19T13:45:42
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,802,029
6,514
alexey-milovidov
Something like race condition of system.parts with DROP table: ``` (gdb) bt #0 __GI___pthread_mutex_lock (mutex=0x14b0) at ../nptl/pthread_mutex_lock.c:65 #1 0x0000000016c0173b in std::__1::__libcpp_mutex_lock (__m=0x14b0) at ../contrib/libcxx/include/__threading_support:255 #2 std::__1::mutex::lock (this=0x1...
2019-08-20T00:17:21
2019-08-20T00:17:21
{}
MEMBER
ClickHouse
ClickHouse
522,840,707
6,535
filimonov
Actually it combines the value but will still work ineffecient. Instead you can create a simple key-value dictionary or table with engine=Join and just do `alter table update col1=dictGet..where 1` (or with joinGet in case of table with engine=Join).
2019-08-20T03:44:37
2019-08-20T03:44:37
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,842,909
6,549
filimonov
On first question: What filesystem is used? Is the directory (listed in exeption message) writable for clickhouse user?
2019-08-20T03:57:48
2019-08-20T03:57:48
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,869,497
6,508
filimonov
Related #6104
2019-08-20T06:14:05
2019-08-20T06:14:05
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,870,022
3,492
filimonov
Related #6104
2019-08-20T06:16:14
2019-08-20T06:16:14
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,669,105
6,532
kreuzerkrieg
@alexey-milovidov Not sure I got your question. After setting background_pool_size to 64 I got following Pay attention, exactly 10 minutes, from finding last needed part and starting to download It doesnt happen if the background_pool_size is set to 16 Never mind, found the problem, my bad
2019-08-19T17:10:01
2019-08-20T06:36:23
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,889,712
6,561
blinkov
@BayoNet do you have a ticket for this?
2019-08-20T07:24:38
2019-08-20T07:24:38
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,937,727
6,563
akuzm
A second error occurred when printing stack trace for the first error. This should be somewhat alleviated by the PR #6529, that is available in v19.14.1.1095-testing. @alexey-milovidov FYI
2019-08-20T09:40:22
2019-08-20T09:40:22
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,515,856
6,540
alesapin
Fix https://github.com/yandex/ClickHouse/pull/6543/files
2019-08-19T10:46:04
2019-08-19T10:46:04
{}
MEMBER
ClickHouse
ClickHouse
522,850,788
6,549
thyn
> On first question: What filesystem is used? Is the directory (listed in exeption message) writable for clickhouse user? I have one ext4 formatted partition on nvme 1 Tb disk It's mounted on /data/d1 /data/d1/clickhouse is linked to /data/clickhouse Clickhouse data path is /data/clickhouse All data is cho...
2019-08-20T04:44:32
2019-08-20T04:45:39
{}
NONE
ClickHouse
ClickHouse
522,968,420
5,834
abyss7
Implemented in 19.14+
2019-08-20T11:18:34
2019-08-20T11:18:34
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,492,613
6,539
4ertus2
There's only count(distinct ) that is replaced with uniqExact() (according to count_distinct_implementation setting). You may try to make sum over subselect with group by.
2019-08-19T09:36:51
2019-08-19T09:37:52
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,584,342
6,543
alesapin
Fail is not connected with changes (known bug in CI)
2019-08-19T13:50:18
2019-08-19T13:50:18
{}
MEMBER
ClickHouse
ClickHouse
522,621,397
6,549
thyn
I was trying to optimize partition (other with problem) `OPTIMIZE TABLE actions.actions PARTITION '2019-07-29' FINAL` I'm getting `Code: 49. DB::Exception: Received from localhost:9000. DB::Exception: Part 20190729_69485_124392_9_196813 intersects next part 20190729_110928_124392_3_196816 (state Committed). It...
2019-08-19T15:12:45
2019-08-19T15:16:59
{}
NONE
ClickHouse
ClickHouse
522,723,868
6,552
alexey-milovidov
We avoid to include `<mysql/mysql.h>` in header files to avoid namespace pollution. Better to include in `.cpp` whenever possible. So, we use forward declarations.
2019-08-19T19:38:28
2019-08-19T19:38:28
{}
MEMBER
ClickHouse
ClickHouse
522,730,841
4,639
alexey-milovidov
It is only implemented for MergeTree, but not for ReplicatedMergeTree.
2019-08-19T19:59:31
2019-08-19T19:59:31
{}
MEMBER
ClickHouse
ClickHouse
522,785,185
6,548
ghost
@alexey-milovidov Thank you for your comment. I fixed the test.
2019-08-19T22:57:43
2019-08-19T22:57:43
{}
NONE
ClickHouse
ClickHouse
522,853,561
6,535
filimonov
For dictionaries - use one of dictGet... functions. For engine=Join - use joinGet. Performance is almost the same. Engine=Join is bit more handy (no XML config, easy to fill / maintain). Dictionaries consume more memory.
2019-08-20T04:59:46
2019-08-20T04:59:46
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,853,673
6,535
Sundarcj
Okay thank u
2019-08-20T05:00:19
2019-08-20T05:00:19
{}
NONE
ClickHouse
ClickHouse
522,881,699
5,387
manish53
@alexey-milovidov Please review this. I have removed conflicts.
2019-08-20T06:57:39
2019-08-20T06:57:39
{}
NONE
ClickHouse
ClickHouse
522,881,718
6,561
filimonov
In upcoming releases (probably in 19.14 and higher) special syntax for that will appear: ```ALTER TABLE ... MATERIALIZE INDEX ... IN PARTITION ...``` See: https://github.com/yandex/ClickHouse/pull/5053/files?file-filters%5B%5D=.md
2019-08-20T06:57:43
2019-08-20T06:57:43
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,969,732
2,248
stavrolia
https://clickhouse.yandex/docs/en/query_language/functions/geo/
2019-08-20T11:22:58
2019-08-20T11:22:58
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,970,185
6,506
filimonov
Sounds like forced `is_readonly`.
2019-08-20T11:24:23
2019-08-20T11:24:23
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,531,392
6,484
alexey-milovidov
@JonLeeCSDN `cluster-1` is interpreted as arithmetic expression. You have to put in in backticks: ``` `cluster-1` ``` or use different cluster name.
2019-08-19T11:36:14
2019-08-19T11:36:30
{ "heart": 2 }
MEMBER
ClickHouse
ClickHouse
522,584,090
6,474
4ertus2
Your crashes are in LFAlloc that were experimental feature. I think update to 19.11 would be the best solution.
2019-08-19T13:49:44
2019-08-19T13:49:44
{ "+1": 2 }
CONTRIBUTOR
ClickHouse
ClickHouse
522,595,503
6,528
filimonov
Now: ``` SELECT *, runningDifference(x) FROM ( SELECT number, arrayJoin([1, 2, 3, 5, 8]) AS x FROM numbers(3) ) ┌─number─┬─x─┬─runningDifference(x)─┐ │ 0 │ 1 │ 0 │ │ 0 │ 2 │ 1 │ │ 0 │ 3 │ 1 │ ...
2019-08-19T14:17:40
2019-08-19T14:17:40
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,606,267
6,451
CurtizJ
Initially TTL was implemented with (maybe a bit confusing) logic, that parts created before `ALTER TABLE ... MODIFY TTL` would never be dropped by TTL. It was changed in PR #6274 and now those old parts can be filtered by TTL with `OPTIMIZE ... FINAL` query. This patch will be available in 19.14 release. NOTE: Even no...
2019-08-19T14:42:08
2019-08-19T14:42:08
{ "+1": 2 }
MEMBER
ClickHouse
ClickHouse
522,501,473
6,425
KochetovNicolai
> I think the fundamental issue is that clickhouse-client process ignores SIGCHLD. Do you know why? @vzakaznikov I know almost nothing about unix signals. But I have just googled that SIGCHLD is sent to parent process by os after child have finished (and usually should be ignored). Why do you think it's related?
2019-08-19T10:02:39
2019-08-19T10:02:39
{}
MEMBER
ClickHouse
ClickHouse
522,573,639
6,538
alexey-milovidov
clickhouse-odbc-bridge is always built as a separate binary (that's what it intended for). If we have some misleading option to include it in `clickhouse` binary, this option should be removed.
2019-08-19T13:24:49
2019-08-19T13:24:49
{}
MEMBER
ClickHouse
ClickHouse
522,594,079
6,405
4ertus2
Try to check default_database in your config files. It most probable than it was chenged to ```DEFAULT```.
2019-08-19T14:14:12
2019-08-19T14:14:12
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,475,534
6,534
4ertus2
> So, is there any other better way to do it? For one-to-one updates you may try to use ALIASES with expressions: https://clickhouse.yandex/docs/en/query_language/create/#create-default-values Then you are able to select from this virtual column instead of source one. > Second Usecase It's something very spec...
2019-08-19T08:50:45
2019-08-19T08:50:54
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,553,031
6,425
KochetovNicolai
@vzakaznikov Tests still fail sometimes. For example on release in prev commit. Wouldn't you mind if I rewrite all tests without `uexpect.py` and so on? As I did with live_view_watch_http.
2019-08-19T12:37:22
2019-08-19T12:37:38
{}
MEMBER
ClickHouse
ClickHouse
522,843,854
6,549
filimonov
On second question - can you grep logs for strings 20190729_69485_124392_9_196813 and 20190729_110928_124392_3_196816 and attach the results?
2019-08-20T04:03:14
2019-08-20T04:03:14
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,863,996
6,561
blinkov
Run `OPTIMIZE TABLE xxx FINAL;` after adding index to build it for old data.
2019-08-20T05:50:45
2019-08-20T05:50:45
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,882,234
5,388
manish53
@alexey-milovidov This is the continuation of PR #5387
2019-08-20T06:59:31
2019-08-20T06:59:31
{}
NONE
ClickHouse
ClickHouse
522,933,691
6,553
akuzm
Do we really have to disable parallel parsings for so many tests? For several-line examples that are usually tested, parallel parsing doesn't make any difference. And I'm not sure what difference it makes for big files, so maybe we just have to leave it always on. If we do have to disable it for the tests, we can d...
2019-08-20T09:29:02
2019-08-20T09:29:02
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,988,297
6,565
4ertus2
It could be realted with #4510
2019-08-20T12:19:57
2019-08-20T12:19:57
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,484,566
2,764
4ertus2
A use case for ANTI JOIN #6534
2019-08-19T09:15:09
2019-08-19T09:15:09
{}
CONTRIBUTOR
ClickHouse
ClickHouse
513,194,815
4,631
filimonov
@florida-hehe can you please try to reproduce that on 19.11.7?
2019-07-19T11:28:03
2019-08-19T10:34:44
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,576,607
6,538
orivej
Even if odbc-bridge should not be a part of clickhouse-lib, this PR should be merged because it removes the second copy of odbc-bridge.cpp from the clickhouse-odbc-bridge program: 1) https://github.com/yandex/ClickHouse/blob/v19.13.2.19-stable/dbms/programs/odbc-bridge/CMakeLists.txt#L8 2) https://github.com/yandex/C...
2019-08-19T13:31:24
2019-08-19T13:31:24
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,592,860
6,535
Sundarcj
okay, consider this following scenario. I need to change all the values in the column. so I run a loop `LEN = somemillion or some billion; for(long i=0;i<LENi++){ connetion.createStatement.execute("alter table sample update columnX = "+_someAlgorithm()+" where row_id="+i"); }` so how many times the partition ...
2019-08-19T14:11:09
2019-08-19T14:11:46
{}
NONE
ClickHouse
ClickHouse
522,637,781
6,425
vzakaznikov
Nicolai (@KochetovNicolai), here is the diff ``` $ git diff diff --git a/dbms/tests/queries/0_stateless/helpers/uexpect.py b/dbms/tests/queries/0_stateless/helpers/uexpect.py index 930a45a099..f71b32a53e 100644 --- a/dbms/tests/queries/0_stateless/helpers/uexpect.py +++ b/dbms/tests/queries/0_stateless/helpers/...
2019-08-19T15:50:17
2019-08-19T15:50:17
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,787,031
6,425
vzakaznikov
Nicolai (@KochetovNicolai), looking at the results of the last commit there are no more failing live view tests. Hopefully, you don't have to rewrite any more of my tests :). I only see the 00395_nullable as failing. Is there a way for you to re-run all the tests a few more times to make sure we are solid?
2019-08-19T23:05:51
2019-08-19T23:05:51
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,870,356
2,184
filimonov
Related #6104
2019-08-20T06:17:40
2019-08-20T06:17:40
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,950,682
6,563
yingfeng
Then what about such queries: ``` SELECT multiIf(id = 1, arrayPushBack([],id), NULL) FROM trips ``` Which will also lead to segmentation error? @akuzm
2019-08-20T10:17:44
2019-08-20T10:17:44
{}
NONE
ClickHouse
ClickHouse
522,991,033
215
filimonov
@philodavies please, give a simple sample.
2019-08-20T12:27:50
2019-08-20T12:27:50
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,566,652
6,483
4ertus2
Developer note: it's possible to get std::terminate() here in ThreadFromGlobalPool::join if DDLWorker dtor called with uninitialized ThreadFromGlobalPool.state It needs some fix.
2019-08-19T13:09:17
2019-08-19T13:10:34
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,592,630
6,534
filimonov
First use case is not clear. You want to allow user to modify data stored in table in any way? How would you imagine that? I.e. you have (let's say) column with 500Gb of strings. And you want to allow user to rewrite that 500Gb of strings? What are better options than just read-process-write back? There are severa...
2019-08-19T14:10:33
2019-08-19T14:11:01
{}
CONTRIBUTOR
ClickHouse
ClickHouse
522,679,729
6,474
iameugenejo
Thank you. I upgraded it to 19.11.7.40 and confirmed the issue is gone.
2019-08-19T17:37:51
2019-08-19T17:37:51
{}
NONE
ClickHouse
ClickHouse
522,780,975
6,548
alexey-milovidov
Need to update tests.
2019-08-19T22:39:24
2019-08-19T22:39:24
{}
MEMBER
ClickHouse
ClickHouse
522,824,809
6,560
alexey-milovidov
We are using gcc 9.1.0 for production builds right now: https://clickhouse-builds.s3.yandex.net/6550/9e2537a7d09ebee108ea2e68a6d6b2f282f528f1/build_log_490262189_1566238254.txt
2019-08-20T02:19:48
2019-08-20T02:19:48
{}
MEMBER
ClickHouse
ClickHouse
522,872,566
3,196
filimonov
currentDatabase part relates #6104
2019-08-20T06:25:33
2019-08-20T06:25:33
{}
CONTRIBUTOR
ClickHouse
ClickHouse
511,814,009
5,999
filimonov
documented? didn't see that. Indeed it can hit the performance in cases where TSV has a lot of zeros - probably testing performance degradation only for that single border case should be ok.
2019-07-16T13:24:08
2019-07-16T13:24:08
{}
CONTRIBUTOR