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
553,652,217
7,765
SreekanthMannari
No. I haven't restarted. But config.xml in "preprocessed_configs" folder shows my change.
2019-11-13T23:27:03
2019-11-13T23:27:03
{}
NONE
ClickHouse
ClickHouse
553,652,759
7,765
den-crane
>But config.xml in "preprocessed_configs" folder shows my change. It does not matter. All parameters from config.xml (except cluster & dictionaries configurations) require CH reboot to apply.
2019-11-13T23:28:59
2019-11-13T23:29:13
{ "laugh": 1 }
CONTRIBUTOR
ClickHouse
ClickHouse
554,123,047
7,392
SaltTan
If I set `enable_scalar_subquery_optimization` to anything I will get `Exception: Unknown setting enable_scalar_subquery_optimization` from older versions in distributed queries.
2019-11-14T22:59:49
2019-11-14T22:59:49
{}
CONTRIBUTOR
ClickHouse
ClickHouse
554,158,414
7,392
den-crane
@amosbird = false Yes, the upgraded server starts to respond with errors `Exception: Unknown setting enable_scalar_subquery_optimization` because not upgraded shards know nothing about this setting. You have to change default value to `false` as always for new settings.
2019-11-15T00:55:12
2019-11-15T01:00:37
{}
CONTRIBUTOR
ClickHouse
ClickHouse
554,358,565
7,788
filimonov
See #7757 (fixed already).
2019-11-15T13:29:17
2019-11-15T13:29:35
{}
CONTRIBUTOR
ClickHouse
ClickHouse
554,418,689
7,757
Tasselmi
so now how can I input multi-line sql code? can you guys give me a example? I have work to accomplish. thanks
2019-11-15T16:01:17
2019-11-15T16:01:17
{}
NONE
ClickHouse
ClickHouse
553,461,771
7,757
amosbird
https://github.com/ClickHouse/ClickHouse/pull/7760
2019-11-13T15:46:56
2019-11-13T15:46:56
{}
COLLABORATOR
ClickHouse
ClickHouse
553,586,721
7,748
den-crane
Engine Log. 1000 tables. Elapsed: 0.047 sec. I think partition pruning optimization. https://github.com/ClickHouse/ClickHouse/pull/2468
2019-11-13T20:26:15
2019-11-13T20:26:15
{}
CONTRIBUTOR
ClickHouse
ClickHouse
553,439,041
2,833
develar
Looks like problem in `clickhouse-go` compress logic. I use latest version from master (6b1ba590788873fd6993f50f823bb90274a23ba5, not released). With `compress=1` and native lz4 (build tag `clz4`) I get this error constantly. Works correctly without `compress=1`. update: no, the same problem without compress...
2019-11-13T14:56:11
2019-11-14T06:16:00
{}
NONE
ClickHouse
ClickHouse
553,841,956
2,161
MeteHanC
> Thanks a lot guys. I'm new beginner to ClickHouse and also confusing like @filimonov. > And I have a question, how can we determine the size of a shard ? OR how many shards should we use ? I'm building a real-time event analytic for 50 millions events per day with 6 server (3 SSD and 3 HDD) > Thanks again. You d...
2019-11-14T11:14:13
2019-11-14T11:14:13
{ "+1": 6, "rocket": 1 }
CONTRIBUTOR
ClickHouse
ClickHouse
554,172,068
7,773
vzakaznikov
Adding a simple test case to reproduce. ``` user-node :) CREATE TABLE myints (a Int32) Engine=Memory; CREATE TABLE myints ( `a` Int32 ) ENGINE = Memory Ok. 0 rows in set. Elapsed: 0.009 sec. user-node :) CREATE LIVE VIEW lv AS SELECT a FROM (SELECT a FROM myints) CREATE LIVE VIEW lv AS SELE...
2019-11-15T01:47:58
2019-11-15T01:47:58
{}
CONTRIBUTOR
ClickHouse
ClickHouse
554,306,202
7,767
filimonov
May be you should decrease `query_profiler_real_time_period_ns` / `query_profiler_cpu_time_period_ns` ? By default both are set to 1000000000, i.e. 1 second, so if your queries finishes faster than 1 sec, you will not get anything in `trace_log`. Try longer queries or smaller values for `query_profiler_real_time_per...
2019-11-15T10:31:34
2019-11-15T10:33:24
{}
CONTRIBUTOR
ClickHouse
ClickHouse
554,397,716
7,704
guanlisheng
hi @filimonov i think [ArrowColumnToCHColumn.cpp](https://github.com/ClickHouse/ClickHouse/blob/817c93f5d05ad0c57663d40d996fefe6df15dd40/dbms/src/Processors/Formats/Impl/ArrowColumnToCHColumn.cpp#L55) had explained the array is not supported yet. also the same desc from its [documentation](https://clickhouse.yan...
2019-11-15T15:14:35
2019-11-15T15:14:35
{}
NONE
ClickHouse
ClickHouse
554,637,420
7,704
guanlisheng
or we just just a line as like `{arrow::Type::LIST, "Array"},`
2019-11-16T13:27:47
2019-11-16T13:27:47
{}
NONE
ClickHouse
ClickHouse
554,646,018
7,803
Tasselmi
> for now CH supports only > > ``` > WITH (SELECT dt, app_id, count() FROM test.test GROUP BY dt, app_id) AS tba > SELECT * FROM tba > UNION ALL > WITH (SELECT dt, app_id, count() FROM test.test GROUP BY dt, app_id) AS tba > SELECT * FROM tba > ``` I tested, and it does not support at all. `Table tba doesn...
2019-11-16T15:14:16
2019-11-16T15:15:26
{}
NONE
ClickHouse
ClickHouse
553,640,957
7,757
alexey-milovidov
@amosbird If we will made it optional, it will not be used in 99.99% cases, and the code is non-trivial, so better to remove the code at all.
2019-11-13T22:51:03
2019-11-13T22:51:03
{}
MEMBER
ClickHouse
ClickHouse
553,651,001
7,765
den-crane
Did you restart CH after setting max_concurrent_queries ?
2019-11-13T23:22:46
2019-11-13T23:23:01
{}
CONTRIBUTOR
ClickHouse
ClickHouse
554,300,431
7,784
filimonov
Workaround: `set enable_optimize_predicate_expression=0;` It looks like that optimization should be disabled for WITH FILL. @zhang2014 what do you think?
2019-11-15T10:14:49
2019-11-15T10:19:42
{}
CONTRIBUTOR
ClickHouse
ClickHouse
554,304,903
7,782
akuzm
I only briefly looked at the PR, so a superficial question: when do we need iteration over hash table in the Set? The reason I'm asking is that I hope we can avoid the `forEachCell` changes, because they are somewhat complicated.
2019-11-15T10:27:56
2019-11-15T10:27:56
{}
CONTRIBUTOR
ClickHouse
ClickHouse
554,185,876
7,767
anheicang
> You have to enable logging in a session `set log_queries = 1;` > or in users's profie ` <log_queries>1</log_queries>` log_queries already enable, and query_log,metric_log all in system,only trace_log not
2019-11-15T02:46:12
2019-11-15T02:46:12
{}
NONE
ClickHouse
ClickHouse
554,433,283
7,783
amosbird
Also makes `INCLUDE_DEBUG_HELPERS` more useful.
2019-11-15T16:36:02
2019-11-15T16:36:02
{}
COLLABORATOR
ClickHouse
ClickHouse
554,481,510
6,847
CurtizJ
Accidentally backported fix for feature, that is missing in 19.14.
2019-11-15T18:46:06
2019-11-15T18:46:06
{}
MEMBER
ClickHouse
ClickHouse
553,391,833
7,753
4ertus2
It's possible to make T64 codec for fixed sized Arrays that use own min-max pairs for each array position. It's also possible to make the codec for general Arrays with one min-max for all the Array's positions. Both cases have some restrictions. I think we should not add such a support fo Arrays now but we may imple...
2019-11-13T12:54:14
2019-11-13T12:54:14
{}
CONTRIBUTOR
ClickHouse
ClickHouse
553,442,212
7,758
4ertus2
#6836
2019-11-13T15:03:20
2019-11-13T15:03:20
{}
CONTRIBUTOR
ClickHouse
ClickHouse
553,672,822
7,757
amosbird
Ok, I've changed it to only enable `bracketed-paste` when `multiquery` is enabled and `multiline` is disabled. This should be less intrusive.
2019-11-14T00:50:14
2019-11-14T01:04:16
{}
COLLABORATOR
ClickHouse
ClickHouse
554,219,030
7,769
amosbird
PREWHERE in ClickHouse can serve two optimization purposes. 1. Reduce disk read IO This is the main purpose of PREWHERE. When the prewhere predicate has low selectivity (f.g. 10%), and non prewhere columns to be scanned is large, PREWHERE reduces a large amount of disk read IO, and saves some efforts to do filter...
2019-11-15T05:31:45
2019-11-15T05:33:38
{}
COLLABORATOR
ClickHouse
ClickHouse
554,320,801
7,772
akuzm
Thanks for the data, can reproduce locally. I get 0 rows even without grouping, so this is probably related to something else: ``` :) SELECT count() as views FROM hit_sh WHERE countryID = 33 and hitDate>=toDate('2019-11-07') AND bitAnd(status, 0x10) AND zoneID IN (1339803993,1200739130,1317373562,2136862789,1301...
2019-11-15T11:15:18
2019-11-15T11:15:18
{}
CONTRIBUTOR
ClickHouse
ClickHouse
554,354,443
7,784
zhang2014
@filimonov disabled it in #7789
2019-11-15T13:15:33
2019-11-15T13:15:33
{ "+1": 1 }
CONTRIBUTOR
ClickHouse
ClickHouse
554,221,308
7,392
amosbird
Ok, I'm not familiar with cluster upgrading, especially seemingless ones. If defaulting the setting to false is a proper fix, I'm fine with that. As I said, "enable_scalar_subquery_optimization is set to true by default in order to get the perf test result."
2019-11-15T05:43:01
2019-11-15T05:43:01
{}
COLLABORATOR
ClickHouse
ClickHouse
553,332,519
7,748
filimonov
Can you also do ``` SET send_logs_level='trace'; ``` before running your select, and compare the output / find the place where it get slower.
2019-11-13T10:08:37
2019-11-13T10:08:37
{}
CONTRIBUTOR
ClickHouse
ClickHouse
553,675,184
7,765
SreekanthMannari
Thanks. Setting worked after CH reboot. Does reboot needed even for User.xml changes like max memory settings?
2019-11-14T01:00:32
2019-11-14T01:00:32
{}
NONE
ClickHouse
ClickHouse
554,002,943
7,666
SaltTan
I tried 19.17.2.4 and got an exception for both version() and uptime(): DB::Exception: Block structure mismatch in MergingSorted stream: different values of constants, actual: '19.13.7.57', expected: '19.17.2.4' DB::Exception: Block structure mismatch in MergingSorted stream: different values of constants, actual...
2019-11-14T17:53:02
2019-11-14T17:53:02
{}
CONTRIBUTOR
ClickHouse
ClickHouse
554,052,138
7,080
cw9
Hi, is there any timeline for supporting this? This would be a great feature.
2019-11-14T19:53:19
2019-11-14T19:53:19
{}
CONTRIBUTOR
ClickHouse
ClickHouse
554,052,497
2,833
develar
Dirty workaround: as suggested above (https://github.com/ClickHouse/ClickHouse/issues/2833#issuecomment-414461705), add following options in DSN `&send_timeout=30000&receive_timeout=3000` (at least, works for me).
2019-11-14T19:54:11
2019-11-14T19:54:11
{}
NONE
ClickHouse
ClickHouse
554,136,999
7,392
den-crane
> `enable_scalar_subquery_optimization` is set to `true` @amosbird you made upgrade of a cluster is impossible.
2019-11-14T23:42:19
2019-11-14T23:42:19
{}
CONTRIBUTOR
ClickHouse
ClickHouse
554,266,144
7,775
rochoa
If you need better precision, you could go with toFloat64. See the example below. ``` SELECT toFloat32('23.739694'), toFloat64('23.739694') ┌─toFloat32('23.739694')─┬─toFloat64('23.739694')─┐ │ 23.739695 │ 23.739694 │ └────────────────────────┴────────────────────────┘ ...
2019-11-15T08:39:17
2019-11-15T08:39:17
{}
NONE
ClickHouse
ClickHouse
554,297,913
7,772
saniapro
I make a small sample set on yours table structure that repeats bug on latest 19.16.3.6 https://fex.net/s/rvt9lbx ``` SELECT countryID,count() as views FROM abn.hit_sh_test_case WHERE hitDate>=toDate('2019-11-07') AND bitAnd(status, 0x10) AND zoneID IN (1339803993,1200739130,1317373562,2136862789,1301506478,1961629...
2019-11-15T10:07:47
2019-11-15T10:12:51
{ "rocket": 1 }
NONE
ClickHouse
ClickHouse
554,311,283
7,714
filimonov
Maybe some part get damaged for some reason? Can data be read with simple selects, like: `SELECT count(ignore(*)) from test` If not - may be you can locate a part causing that? Can you try to create a reproducible sample?
2019-11-15T10:46:41
2019-11-15T10:47:30
{}
CONTRIBUTOR
ClickHouse
ClickHouse
554,651,058
7,803
den-crane
I see now. CH `WITH` supports scalars only. with (select 1) as x select x with (select 1) as x select x union all with (select 1) as x select x select y from (with (select 1) as x select x as y)
2019-11-16T16:16:05
2019-11-16T16:16:05
{ "confused": 1 }
CONTRIBUTOR
ClickHouse
ClickHouse
553,458,103
7,757
amosbird
@akuzm multi-statement pasting now requires `-n` after https://github.com/ClickHouse/ClickHouse/pull/7492 . (I wonder if that should be the default value) > Also you always need to press Enter after pasting some query. @filimonov hmm, I favor the opposite, to validate (or do small changes to) the query that has...
2019-11-13T15:38:48
2019-11-13T15:38:48
{}
COLLABORATOR
ClickHouse
ClickHouse
554,026,506
7,172
CurtizJ
Backport conflict to 19.14
2019-11-14T18:50:07
2019-11-14T18:50:07
{}
MEMBER
ClickHouse
ClickHouse
554,346,385
7,784
ghuname
I confirm that workaround with select ... settings enable_optimize_predicate_expression=0 works.
2019-11-15T12:46:34
2019-11-15T12:46:34
{}
NONE
ClickHouse
ClickHouse
554,465,104
7,792
vzakaznikov
Another option. Instead of adding `live_view_context` is to update dbms/src/Interpreters/InterpreterSelectWithUnionQuery.cpp to something like ```c++ Block InterpreterSelectWithUnionQuery::getSampleBlock( const ASTPtr & query_ptr, const Context & context) { if context.hasQueryContext() { ...
2019-11-15T18:01:00
2019-11-15T18:01:00
{}
CONTRIBUTOR
ClickHouse
ClickHouse
554,645,382
7,794
Tasselmi
> Just change timeout. > > set distributed_ddl_task_timeout = 9000; THANKS. PLEASE HELP ME IN #7803
2019-11-16T15:06:18
2019-11-16T15:06:18
{}
NONE
ClickHouse
ClickHouse
553,358,203
1,825
BayoNet
Is this documentation correct? https://clickhouse.yandex/docs/en/query_language/show/#show-processlist
2019-11-13T11:16:18
2019-11-13T11:16:18
{}
CONTRIBUTOR
ClickHouse
ClickHouse
553,832,348
1,784
xmariachi
Hi @NanoBjorn, yes, I think so. It works fine with duplicates on smaller table sizes.
2019-11-14T10:46:34
2019-11-14T10:46:34
{}
NONE
ClickHouse
ClickHouse
554,155,803
7,392
amosbird
@SaltTan but `enable_scalar_subquery_optimization` only affects newer version of clickhouse. Older versions should behave as if it is set to `false` @den-crane Hmm, did you try upgrading with that setting = `false` using config file?
2019-11-15T00:45:13
2019-11-15T00:45:13
{}
COLLABORATOR
ClickHouse
ClickHouse
554,197,821
7,748
Girgitt
> Can you try to minimize the query? I.e. remove the parts which does not change that problem appearance I had no luck in minimizing the query so far - it seems like any field in select part or where part of the query contributes to the initial delay. The less fields included the shorter delay but it's always there....
2019-11-15T03:38:23
2019-11-15T03:38:23
{}
NONE
ClickHouse
ClickHouse
554,399,052
7,772
CurtizJ
I found that issue is caused by replacing `&&` to `&` [here](https://github.com/ClickHouse/ClickHouse/blob/817c93f5d05ad0c57663d40d996fefe6df15dd40/dbms/src/Functions/FunctionsLogical.h#L66). As I understand it was needed to implement ternary logic for and/or. So, now all bit functions are interpreted wrong in terms of...
2019-11-15T15:17:56
2019-11-15T15:17:56
{}
MEMBER
ClickHouse
ClickHouse
553,568,160
7,748
filimonov
Can you try to minimize the query? I.e. remove the parts which does not change that problem appearence
2019-11-13T19:43:03
2019-11-13T19:43:03
{}
CONTRIBUTOR
ClickHouse
ClickHouse
553,795,722
7,769
zhang2014
How does this work when the `where expression` cannot filter large amounts of data ? in our scenario, we always put `where expression` that are not included in the `primary key` into `prewhere expression`.
2019-11-14T09:12:39
2019-11-14T09:23:18
{}
CONTRIBUTOR
ClickHouse
ClickHouse
553,887,049
7,772
den-crane
Can you provide a sample of data or make reproducible example?
2019-11-14T13:27:03
2019-11-14T13:27:03
{}
CONTRIBUTOR
ClickHouse
ClickHouse
554,321,576
7,610
filimonov
> What's the root cause of this damaged part? You can try to find fist mention of that part name in the clickhouse logs. Maybe something happened at that moment? Maybe there was a problem with filesystem (like lack of disk space), maybe there was some hard restart or something like that. Please update clickhouse ...
2019-11-15T11:17:34
2019-11-15T11:17:34
{}
CONTRIBUTOR
ClickHouse
ClickHouse
554,322,849
7,609
filimonov
Can you please paste the text of queries and output, instead of screenshots? What is the version of the clickhouse? Did you change some settings from their default?
2019-11-15T11:21:36
2019-11-15T11:21:36
{}
CONTRIBUTOR
ClickHouse
ClickHouse
554,464,889
7,641
CurtizJ
backport conflict in 19.15
2019-11-15T18:00:26
2019-11-15T18:00:26
{}
MEMBER
ClickHouse
ClickHouse
554,637,529
7,704
guanlisheng
@proller and @alexey-milovidov , thoughts?
2019-11-16T13:29:05
2019-11-16T13:29:05
{ "eyes": 1 }
NONE
ClickHouse
ClickHouse
553,379,260
7,540
akuzm
The original query is `SELECT toDecimal32OrZero(CAST(-7174046, 'String'), 6)` Another log for `SELECT bitTestAny(-3, 94874)` https://clickhouse-test-reports.s3.yandex.net/0/30acb1b653cbc21fbd8b873023a9f9ace0715e1a/functional_stateless_tests_(ubsan)/stderr.log
2019-11-13T12:16:56
2019-11-13T12:16:56
{}
CONTRIBUTOR
ClickHouse
ClickHouse
553,942,440
7,778
filimonov
Also #6387
2019-11-14T15:37:26
2019-11-14T15:37:26
{}
CONTRIBUTOR
ClickHouse
ClickHouse
554,317,632
7,704
filimonov
It would be good if you will provide more info: docs & samples (data / schema files) Code is here: https://github.com/ClickHouse/ClickHouse/blob/master/dbms/src/Processors/Formats/Impl/ArrowColumnToCHColumn.cpp Also #4903
2019-11-15T11:05:17
2019-11-15T11:05:17
{}
CONTRIBUTOR
ClickHouse
ClickHouse
554,360,062
7,757
Tasselmi
why not just learn from `scala` or `spark-shell`? `:paste` to start copy code , and `ctrl+d` to start interpret.... just advice. actually `\` is not friendly , because I have to remove all the `\` to run in `tabix` or other program.
2019-11-15T13:34:01
2019-11-15T13:34:22
{}
NONE
ClickHouse
ClickHouse
553,464,624
7,757
akuzm
I think we should not change the default behavior without a strong reason, even more so if we're changing it to something different from every other console client.
2019-11-13T15:53:20
2019-11-13T15:53:20
{}
CONTRIBUTOR
ClickHouse
ClickHouse
553,681,296
7,765
den-crane
>Does reboot needed even for User.xml changes like max memory settings? No, changes in user.xml does not need reboot.
2019-11-14T01:27:11
2019-11-14T01:27:11
{}
CONTRIBUTOR
ClickHouse
ClickHouse
553,716,166
7,767
den-crane
You have to enable logging in a session `set log_queries = 1;` or in users's profie ` <log_queries>1</log_queries>`
2019-11-14T04:11:34
2019-11-14T04:11:34
{}
CONTRIBUTOR
ClickHouse
ClickHouse
553,885,156
7,775
den-crane
The same as greatCircleDistance https://clickhouse.yandex/docs/en/query_language/functions/geo/#greatcircledistance Calculate the distance between two points on the Earth's surface. ``` create table poi(id Int64, lat Float64, lon Float64) Engine = Memory; insert into poi values(1, -55.755831, -37.617673); ...
2019-11-14T13:22:03
2019-11-14T13:36:09
{}
CONTRIBUTOR
ClickHouse
ClickHouse
553,909,155
6,515
filimonov
@fancyqlx you mean you changed that: https://github.com/ClickHouse/ClickHouse/blob/2ad4df1d6aa2fbd7d3a030b0e9e4fff9dde8600e/dbms/src/DataStreams/AsynchronousBlockInputStream.h#L99 ?
2019-11-14T14:22:11
2019-11-14T14:22:11
{}
CONTRIBUTOR
ClickHouse
ClickHouse
553,944,422
4,190
SychovOleh
It is a really easy task. But, unfortunately, looks like no one wants to fix it.
2019-11-14T15:41:35
2019-11-14T15:41:35
{ "confused": 1 }
NONE
ClickHouse
ClickHouse
554,248,825
7,459
akuzm
This PR was split in two.
2019-11-15T07:36:25
2019-11-15T07:36:25
{}
CONTRIBUTOR
ClickHouse
ClickHouse
554,324,417
7,537
filimonov
@chijiajian what is the exact version (and distribution) of mysql you use? Any non-default setting on mysql side? Was it one-time issue, or it reproduces every time?
2019-11-15T11:26:47
2019-11-15T11:26:47
{}
CONTRIBUTOR
ClickHouse
ClickHouse
554,330,533
7,772
akuzm
It gets better, just adding `zoneID != -1` breaks it: ``` :) SELECT distinct(zoneID) FROM hit_sh WHERE hitDate>=toDate('2019-11-07') AND bitAnd(status, 0x10) AND zoneID IN (1339803993,1200739130,1317373562,2136862789,1301506478,1961629277) ┌─────zoneID─┐ │ 1961629277 │ │ 1301506478 │ └────────────┘ ┌─────zone...
2019-11-15T11:49:42
2019-11-15T11:49:42
{ "laugh": 2 }
CONTRIBUTOR
ClickHouse
ClickHouse
554,605,024
7,795
alexey-milovidov
@Akazz Ok, but need to check if there is any difference in performance.
2019-11-16T05:02:06
2019-11-16T05:02:06
{}
MEMBER
ClickHouse
ClickHouse
554,645,316
7,803
den-crane
for now CH supports only ``` WITH (SELECT dt, app_id, count() FROM test.test GROUP BY dt, app_id) AS tba SELECT * FROM tba UNION ALL WITH (SELECT dt, app_id, count() FROM test.test GROUP BY dt, app_id) AS tba SELECT * FROM tba ```
2019-11-16T15:05:37
2019-11-16T15:05:37
{}
CONTRIBUTOR
ClickHouse
ClickHouse
553,703,703
7,729
tlightsky
it seems disappear when i build second time without `-DCMAKE_BUILD_WITH_INSTALL_RPATH=ON`
2019-11-14T03:08:40
2019-11-14T03:08:40
{}
CONTRIBUTOR
ClickHouse
ClickHouse
553,790,793
2,833
orangeops
executeQuery: Code: 209, e.displayText() = DB::Exception: Timeout exceeded while receiving data from client. Waited for 300 seconds, timeout is 300 seconds
2019-11-14T08:59:30
2019-11-14T08:59:30
{}
NONE
ClickHouse
ClickHouse
553,792,492
2,833
develar
I suspect problem is that when clickhouse-go starts transaction, TCP connection is opened. But if you do not commit transaction in 300 seconds, then clickhouse closes connection and on commit you immediately get error because connection is already closed by server.
2019-11-14T09:03:57
2019-11-14T09:04:23
{}
NONE
ClickHouse
ClickHouse
554,025,918
7,772
akuzm
There was a change in 19.14 for UInt8 aggregation keys -- may be related: https://github.com/ClickHouse/ClickHouse/pull/6298
2019-11-14T18:48:40
2019-11-14T18:48:40
{}
CONTRIBUTOR
ClickHouse
ClickHouse
554,262,732
7,772
saniapro
Also interesting that query without AND countryID=33 returns whole set of countries and proper counts. About dump of rows - I'll try to make set of records that repeats bug, because whole dir of hit_sh is 450Gb. Where can I download it?
2019-11-15T08:27:59
2019-11-15T08:27:59
{}
NONE
ClickHouse
ClickHouse
554,324,999
7,546
filimonov
BTW: 'official' mac build from CI have some issues (sounds like readline-related) in clickhouse-client
2019-11-15T11:28:42
2019-11-15T11:28:42
{}
CONTRIBUTOR
ClickHouse
ClickHouse
553,472,307
7,757
amosbird
Fair enough. I've made it optional and default to false.
2019-11-13T16:10:13
2019-11-13T16:10:13
{}
COLLABORATOR
ClickHouse
ClickHouse
553,582,918
7,748
den-crane
`for i in {1..1000}; do echo "CREATE TABLE default.PAR_${i} ( d Date DEFAULT CAST(0, 'Date'),ts UInt32,ids UInt32,tss Int32,stat UInt16,val Float32)ENGINE = MergeTree(d, (ids, ts), 8192)"|clickhouse-client; done` original query: Elapsed: 11.748 sec. ``` select min(ids), toStartOfFiveMinute(toDateTime...
2019-11-13T20:16:23
2019-11-13T20:16:23
{ "+1": 1 }
CONTRIBUTOR
ClickHouse
ClickHouse
553,702,400
7,730
tlightsky
@filimonov here it is https://github.com/ClickHouse/ClickHouse/pull/7766
2019-11-14T03:02:14
2019-11-14T03:02:31
{}
CONTRIBUTOR
ClickHouse
ClickHouse
554,276,268
7,772
filimonov
Can you try to minimize the data set first? For example, try smth like: ```sql CREATE TABLE abn.hit_sh_test_case ( `hitDate` Date, `zoneID` UInt32, `campaignID` UInt32, `countryID` UInt8, `status` UInt8 ) ENGINE = MergeTree(hitDate, campaignID, 8192); INSERT INTO abn.hit_sh_tes...
2019-11-15T09:10:09
2019-11-15T09:10:09
{}
CONTRIBUTOR
ClickHouse
ClickHouse
554,605,111
7,795
alexey-milovidov
I suspect that usual WHERE clause with comparison and logical expressions may slow down.
2019-11-16T05:03:20
2019-11-16T05:03:20
{}
MEMBER
ClickHouse
ClickHouse
554,634,842
7,803
Tasselmi
in hive, I can write it like this and it works fine : ``` WITH tba AS (SELECT * FROM test.test LIMIT 10) SELECT * FROM tba UNION ALL SELECT * FROM tba ```
2019-11-16T12:54:44
2019-11-16T12:54:44
{}
NONE
ClickHouse
ClickHouse
553,360,064
7,745
BayoNet
Thanx for your comments :-). I'll review these aricles and try to make them more clear and consistent.
2019-11-13T11:21:43
2019-11-13T11:21:43
{}
CONTRIBUTOR
ClickHouse
ClickHouse
553,510,886
6,005
richardartoul
I guess the `LowCardinality` native format does have flags for re-using the same dictionary as the previous block, so I guess that could be used where the first array element has the dictionary and all subsequent arrays just set the flag to keep reusing the previous dictionary?
2019-11-13T17:29:40
2019-11-13T17:29:40
{}
NONE
ClickHouse
ClickHouse
553,810,790
7,760
alexey-milovidov
We can include it in changelog.
2019-11-14T09:50:06
2019-11-14T09:50:06
{}
MEMBER
ClickHouse
ClickHouse
553,818,376
7,565
ghost
I Added `DROP QUERY_CACHE` to delete all query_caches.
2019-11-14T10:09:53
2019-11-14T10:09:53
{}
NONE
ClickHouse
ClickHouse
553,885,614
7,773
den-crane
cc @vzakaznikov
2019-11-14T13:23:17
2019-11-14T13:23:17
{}
CONTRIBUTOR
ClickHouse
ClickHouse
554,002,105
7,772
filimonov
Highlighted diff of execution ```diff @@ -14,14 +14,10 @@ MergeTreeThread <Trace> ParallelAggregatingBlockInputStream: Aggregating -<Trace> Aggregator: Aggregation method: key8 -<Trace> Aggregator: Aggregation method: key8 -<Trace> Aggregator: Aggregation method: key8 -<Trace> Aggregator: Aggregati...
2019-11-14T17:50:58
2019-11-14T17:50:58
{}
CONTRIBUTOR
ClickHouse
ClickHouse
554,275,852
7,666
KochetovNicolai
Suppose I need to continue fixing it. @alexey-milovidov maybe the correct way to solve it is to expect that `version()` and `uptime()` are not constant even in scope of query?
2019-11-15T09:08:51
2019-11-15T09:08:51
{}
MEMBER
ClickHouse
ClickHouse
554,307,275
7,782
amosbird
Not for now, but maybe used for https://github.com/ClickHouse/ClickHouse/pull/7755 . Also `forEachCell` is used in Join. The reason that this pr contains such changes is because I did it after https://github.com/ClickHouse/ClickHouse/pull/7662 and I don't want to resolve merge conflicts... (and there are two sepa...
2019-11-15T10:34:51
2019-11-15T10:35:43
{}
COLLABORATOR
ClickHouse
ClickHouse
554,651,326
7,803
Tasselmi
> I see now. > CH `WITH` supports scalars only. > > with (select 1) as x select x > > with (select 1) as x select x union all with (select 1) as x select x > > select y from (with (select 1) as x select x as y) cry....
2019-11-16T16:18:54
2019-11-16T16:18:54
{}
NONE
ClickHouse
ClickHouse
553,368,120
7,748
Girgitt
> Can you try latest version first? I did - the latest version is where I noticed the issue and then started testing from the oldest version I knew worked fine (1.1.5365) till the first experiencing the issue (minor delay started in 18.14.8 and the considerable delay started in 18.16.1 - at the same level as in the ...
2019-11-13T11:44:56
2019-11-13T11:44:56
{}
NONE
ClickHouse
ClickHouse
553,422,775
7,757
filimonov
I'm also hitting that last week. Also you always need to press Enter after pasting some query. One workaround is to run clickhouse with -nm parameters, but in that case `exit` command doesn't work (only Ctrl+C) :D I think it's caused by https://github.com/ClickHouse/ClickHouse/pull/7492
2019-11-13T14:16:50
2019-11-13T14:16:50
{}
CONTRIBUTOR
ClickHouse
ClickHouse
554,262,757
7,775
samz406
thank's ,the source data is String type. for exmaple ,the lon is '23.739694', I use toFloat convert float.the convert data is not 23.739694. why is that?
2019-11-15T08:28:05
2019-11-15T08:28:05
{}
NONE
ClickHouse
ClickHouse
554,319,444
7,628
filimonov
@Tasselmi official repo is in Moscow only, but usually, it works very well, maybe there were some temporary network issues. Please retry. For centos, you can also use repo from Altinity (it is not the mirror, it is clickhouse build on CentOS natively): https://github.com/Altinity/clickhouse-rpm-install
2019-11-15T11:11:05
2019-11-15T11:11:05
{}
CONTRIBUTOR
ClickHouse
ClickHouse
554,390,684
7,791
DimasKovas
I didn't find any usages of PODArray.insert with 3 iterators in the code base. Nothing is affected by this PR yet.
2019-11-15T14:56:46
2019-11-15T14:56:46
{}
CONTRIBUTOR
ClickHouse
ClickHouse
554,490,308
7,769
amosbird
https://github.com/yandex/ClickHouse/blob/master/dbms/tests/queries/1_stateful/00091_prewhere_two_conditions.sql#L6 catches another tradeoff. If we put predicates with heavy columns into PREWHERE, we might increase IO. This also means if no filtering is done in prewhere, it will also increase IO (similar to case 2, p...
2019-11-15T19:11:53
2019-11-15T19:21:11
{}
COLLABORATOR
ClickHouse
ClickHouse
554,645,173
7,794
den-crane
Just change timeout. set distributed_ddl_task_timeout = 9000;
2019-11-16T15:03:56
2019-11-16T15:03:56
{}
CONTRIBUTOR
ClickHouse
ClickHouse
544,261,799
7,380
alexey-milovidov
Ok, now it's clear :) It replaces values where condition is met with the previous value that have condition not met (or with the first value if condition was true for all elements before). Looks somewhat similar to `arrayFilter`, but instead removing elements, it fills gaps with the value of previous element. ...
2019-10-20T15:18:30
2019-10-20T15:18:47
{}
MEMBER