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 | 706,638,506 | 15,805 | l1t1 | optimize table without final works .confused
```
optimize table rmt;
Ok.
0 rows in set. Elapsed: 3.848 sec.
select count(distinct id),count(id),now() from rmt;
┌─uniqExact(id)─┬─count(id)─┬───────────────now()─┐
│ 7907336 │ 7907336 │ 2020-10-11 10:08:42 │
└───────────────┴───────────┴───────────... | 2020-10-11T02:11:05 | 2020-10-11T02:11:05 | {} | NONE |
ClickHouse | ClickHouse | 706,502,587 | 15,805 | l1t1 | many ids are not replace merge
```
select count(distinct id),count(id) from rmt;
┌─uniqExact(id)─┬─count(id)─┐
│ 6400000 │ 7500000 │
└───────────────┴───────────┘
``` | 2020-10-10T07:16:20 | 2020-10-10T07:16:20 | {} | NONE |
ClickHouse | ClickHouse | 706,609,029 | 15,821 | azat | There are other hunks like this, will submit another PR | 2020-10-10T20:44:41 | 2020-10-10T20:44:41 | {} | MEMBER |
ClickHouse | ClickHouse | 706,614,531 | 15,805 | l1t1 | thanks | 2020-10-10T21:33:23 | 2020-10-10T21:33:23 | {} | NONE |
ClickHouse | ClickHouse | 706,638,621 | 15,805 | l1t1 | i found a **rule**, if optimize table took many seconds, it should work | 2020-10-11T02:12:59 | 2020-10-11T02:12:59 | {} | NONE |
ClickHouse | ClickHouse | 706,187,794 | 15,750 | alesapin | Not real race | 2020-10-09T13:41:51 | 2020-10-09T13:41:51 | {} | MEMBER |
ClickHouse | ClickHouse | 706,471,510 | 15,001 | IVitamin-C | Hope to get help. Thank you very much | 2020-10-10T02:29:55 | 2020-10-10T02:29:55 | {} | NONE |
ClickHouse | ClickHouse | 706,558,030 | 15,810 | DaiDaDao | I deployed FileBeat to collect logs. And finally found the error. because host is not IP , it's hostname , that's not found , I will find how can fix it, maybe I will write it in system host file.
If anyone encounters the same problem, welcome to discuss with me | 2020-10-10T14:29:15 | 2020-10-10T14:29:15 | {} | NONE |
ClickHouse | ClickHouse | 706,581,431 | 15,609 | bugaevc | > `No such file or directory [/Volumes/SystemRoot/etc/localtime]`
Sounds like your *host* system doesn't have `/etc/localtime`. | 2020-10-10T17:10:52 | 2020-10-10T17:10:52 | {} | NONE |
ClickHouse | ClickHouse | 706,668,221 | 15,828 | azat | >ClickHouse build check — 15/16 builds are OK
Actually unbundled jemalloc cannot be used, since it includes jemalloc_cpp.o that contains new/delete overrides, so as clickhouse:
```
2020-10-11 06:52:23 ld.lld: error: duplicate symbol: operator delete(void*, unsigned long)
2020-10-11 06:52:23 >>> defined at new_d... | 2020-10-11T08:13:26 | 2020-10-11T08:13:26 | {} | MEMBER |
ClickHouse | ClickHouse | 706,175,020 | 9,434 | ucasfl | > `logTrace('message')` - emits trace log message for each block (`LOG_TRACE`), message argument must be constant expression.
The message emits to server log and no return value? @alexey-milovidov | 2020-10-09T13:17:20 | 2020-10-09T13:17:45 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 706,188,083 | 15,698 | alesapin | not real race | 2020-10-09T13:42:23 | 2020-10-09T13:42:23 | {} | MEMBER |
ClickHouse | ClickHouse | 706,282,481 | 13,653 | greenbergjosh | Does anyone have any thoughts on this? I am writing a code generator to generate clickhouse queries, and this issue will require me to carry along superfluous type information so I can do the casts. Any chance this could be corrected at some point? | 2020-10-09T16:35:10 | 2020-10-09T16:35:10 | {} | NONE |
ClickHouse | ClickHouse | 706,362,733 | 15,711 | alexey-milovidov | I hoped it will help but the issues with ZooKeeper cluster in Yandex Metrics have disappeared and now I cannot validate whether it helps. | 2020-10-09T19:25:28 | 2020-10-09T19:25:40 | {} | MEMBER |
ClickHouse | ClickHouse | 706,443,749 | 14,939 | nikitamikhaylov | Merged in #15783 | 2020-10-09T23:34:07 | 2020-10-09T23:34:07 | {} | MEMBER |
ClickHouse | ClickHouse | 706,456,348 | 15,782 | dbawangwei | When the MySQL database is synchronized to clickhouse, the MySQL tables cannot be synchronized without a primary key. In production, many MySQL tables do not have a primary key for various reasons, so data cannot be synchronized to clickhouse. | 2020-10-10T00:40:48 | 2020-10-10T00:40:48 | {} | NONE |
ClickHouse | ClickHouse | 706,571,466 | 15,609 | amosbird | ```
Darling [/Volumes/SystemRoot/home/amos/clickhouse-test]$ ./clickhouse server --config config.xml
Segmentation fault: 11 (core dumped)
```
It doesn't work for current darling. MacOS build is acquired from https://builds.clickhouse.tech/master/macos/clickhouse | 2020-10-10T16:04:27 | 2020-10-10T16:04:27 | {
"+1": 1
} | COLLABORATOR |
ClickHouse | ClickHouse | 706,572,260 | 15,609 | alexey-milovidov | Ok, thank you!
The idea has failed miserably :) | 2020-10-10T16:09:18 | 2020-10-10T16:09:18 | {} | MEMBER |
ClickHouse | ClickHouse | 706,582,407 | 15,609 | amosbird | `/Volumes/SystemRoot/etc/localtime` exists. | 2020-10-10T17:17:12 | 2020-10-10T17:17:12 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 706,604,924 | 15,035 | azat | Initially I marked this PR as a draft since I want to get feedback from the maintainers.
However [looks like](https://github.com/jemalloc/jemalloc/pull/1939#issuecomment-703950546) they don't think that dynamic affinity is the generic case, and using always total number of CPU can cause some issues for [decay](http://... | 2020-10-10T20:11:22 | 2020-10-10T20:11:22 | {} | MEMBER |
ClickHouse | ClickHouse | 706,605,661 | 15,035 | alexey-milovidov | We are going to enable ThreadFuzzer in CI as soon as it will find no more flaky tests - basically as soon as possible. | 2020-10-10T20:18:19 | 2020-10-10T20:18:19 | {} | MEMBER |
ClickHouse | ClickHouse | 706,610,179 | 15,497 | alexey-milovidov | Because it is using `std::to_string` that is extremely slow. | 2020-10-10T20:54:05 | 2020-10-10T20:54:05 | {} | MEMBER |
ClickHouse | ClickHouse | 706,677,862 | 15,833 | l1t1 | ```
SQL> set autot trace;
SQL> select count(*) from rmt where (id,name,dt)in (select id,name,dt from mt);
Elapsed: 00:00:02.09
SQL> set lines 132 pages 50000
SQL> /
Elapsed: 00:00:02.08
Execution Plan
----------------------------------------------------------
Plan hash value: 187692788
------------... | 2020-10-11T09:39:57 | 2020-10-11T09:39:57 | {} | NONE |
ClickHouse | ClickHouse | 706,156,865 | 15,781 | den-crane | 19.17.4.11 - is out of support. Supported versions are 20.7 , 20.8 , 20.9, 20.3 (LTS)
Also need more information. Do you get this error at one server or several?
Can you show stacktraces , or full CH logs ?
| 2020-10-09T12:39:40 | 2020-10-09T16:51:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 706,289,980 | 15,775 | den-crane | Sorry, My mistake. 20.3.19.4 is supported. | 2020-10-09T16:50:44 | 2020-10-09T16:51:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 706,508,799 | 14,528 | alexey-milovidov | Reminds CGI vs FastCGI. | 2020-10-10T08:07:28 | 2020-10-10T08:07:28 | {} | MEMBER |
ClickHouse | ClickHouse | 706,620,678 | 14,414 | filimonov | > 01064_incremental_streaming_from_2_src_with_feedback
fixed in https://github.com/ClickHouse/ClickHouse/pull/11692 | 2020-10-10T22:35:22 | 2020-10-10T22:35:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 706,636,230 | 15,809 | amosbird | > Maybe it is better to add previousQueryId + some client hooks
That sounds no better either :) . previousQueryId is interesting but client now becomes stateful, which is bad. | 2020-10-11T01:42:09 | 2020-10-11T01:42:09 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 706,158,186 | 15,775 | den-crane | 20.3.19.4 is also obsolete. Supported versions are 20.7, 20.8, 20.9.
Can you show an output of the clickhouse-client for this query? | 2020-10-09T12:42:20 | 2020-10-09T12:46:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 706,187,638 | 15,749 | alesapin | Not real race | 2020-10-09T13:41:37 | 2020-10-09T13:41:37 | {} | MEMBER |
ClickHouse | ClickHouse | 706,188,149 | 15,697 | alesapin | not real race | 2020-10-09T13:42:30 | 2020-10-09T13:42:30 | {} | MEMBER |
ClickHouse | ClickHouse | 706,189,415 | 15,619 | alesapin | Test failures and the fix is not important. | 2020-10-09T13:44:54 | 2020-10-09T13:44:54 | {} | MEMBER |
ClickHouse | ClickHouse | 706,227,949 | 15,702 | bohutang | GTID replication mode works for it. | 2020-10-09T14:53:31 | 2020-10-09T14:53:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 706,469,121 | 15,775 | den-crane | Can you show cluster configuration (xml) (remote_servers) for dtstack ? | 2020-10-10T02:11:01 | 2020-10-10T02:11:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 706,543,363 | 15,805 | den-crane | `OPTIMIZE TABLE` initiates an unplanned merge. It's not designed to remove duplicates.
For example your table has 100 partitions and each partition has 100 parts. OPTIMIZE TABLE can just merge 2 of the parts inside one partitions and leave 10000 unmerged parts. Furthermore it can do nothing because of lack of free re... | 2020-10-10T12:41:16 | 2020-10-10T12:41:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 706,568,992 | 13,166 | calebeaires |
<img width="927" alt="Screen Shot 2020-10-10 at 12 48 10" src="https://user-images.githubusercontent.com/794195/95659359-e4b17600-0af6-11eb-959a-a49e085d912c.png">
This function does not exist,
```
select mapPopulateSerie([1,2,4], [11,22,44], 5) as res, toTypeName(res) as type;
``` | 2020-10-10T15:48:45 | 2020-10-10T15:48:45 | {} | NONE |
ClickHouse | ClickHouse | 706,609,938 | 15,497 | alexey-milovidov | It returns empty string for zero. I expect that it should output "0 seconds" instead. | 2020-10-10T20:51:51 | 2020-10-10T20:52:01 | {} | MEMBER |
ClickHouse | ClickHouse | 706,645,867 | 15,776 | baibaichen | I don't find enough documentation about the Pipeline and [sorting](https://clickhousedb.slack.com/archives/CU170QE9H/p1601451196008500), so these examples are actually documentation, I hope they are maintained to exactly reflect the latest usage.
would you please tell me which functional test cover FinishSorting ?... | 2020-10-11T03:54:37 | 2020-10-11T03:54:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 706,223,384 | 15,497 | alexey-milovidov | (You can see it in the result of `ClickHouse special build check`) | 2020-10-09T14:44:44 | 2020-10-09T14:44:44 | {} | MEMBER |
ClickHouse | ClickHouse | 706,284,469 | 15,784 | zvonand | > @zvonand do you mean current stable CH or about code with PR #9404 ?
PR #9404 | 2020-10-09T16:39:34 | 2020-10-09T16:39:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 706,583,745 | 15,609 | alexey-milovidov | `date` works (almost) perfectly on my machine :)
```
milovidov@milovidov-desktop:~/work/ClickHouse/src/Formats$ date
Сб окт 10 20:24:44 MSK 2020
milovidov@milovidov-desktop:~/work/ClickHouse/src/Formats$ darling shell
Darling [/Volumes/SystemRoot/home/milovidov/work/ClickHouse/src/Formats]$ date
суббота, 10 окт... | 2020-10-10T17:25:12 | 2020-10-10T17:25:56 | {} | MEMBER |
ClickHouse | ClickHouse | 706,584,195 | 15,609 | bugaevc | cc @LubosD — sounds like a vchroot bug | 2020-10-10T17:28:27 | 2020-10-10T17:28:27 | {} | NONE |
ClickHouse | ClickHouse | 706,619,040 | 15,757 | filimonov | AST fuzzer - already reported: https://github.com/ClickHouse/ClickHouse/issues/14974 | 2020-10-10T22:18:01 | 2020-10-10T22:18:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 706,655,057 | 15,809 | azat | >but client now becomes stateful
Not the client, since the information is stored in the client context on the server, but agree - explicit query_id is better
Those query_id_formats looks too specific though
>system flush logs settings log_queries=0; -- that bit that was missing that time is updating previousQuer... | 2020-10-11T05:59:53 | 2020-10-11T05:59:53 | {} | MEMBER |
ClickHouse | ClickHouse | 706,281,137 | 15,787 | den-crane | It's an expected behaviour as with any other database.
Indexes works on a prefix. Combined index on two columns (a,b) it's actually is a combine (ab)
So then you search `SERVICE1/SERVICE2/SERVICE3%` in `stnamestarttime` it works effectively.
And ineffectively in case of `starttimestname`.
Most of other databases... | 2020-10-09T16:32:20 | 2020-10-09T16:34:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 706,370,949 | 15,163 | alexey-milovidov | Now it is ready to become prestable. | 2020-10-09T19:45:48 | 2020-10-09T19:45:48 | {} | MEMBER |
ClickHouse | ClickHouse | 706,550,321 | 15,775 | den-crane | As I understood you have 2 nodes. --- 1 shard - 2 replicas.
Then you config have to be
```
<clickhouse_remote_servers>
<dtstack>
<shard>
<internal_replication>true</internal_replication>
<replica>
<host>172.16.8.96</host>
... | 2020-10-10T13:34:57 | 2020-10-10T13:36:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 706,571,169 | 13,597 | amosbird | You can just tweak your PATH env to contain a working objcopy and cmake will find it. | 2020-10-10T16:02:56 | 2020-10-10T16:02:56 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 706,676,673 | 15,833 | l1t1 | same test in oracle
```
Oracle Database 18c Express Edition Release 18.0.0.0.0 - Production
Version 18.4.0.0.0
SQL> set timi on
SQL> create table rmt(id int,name int, dt date);
Table created.
Elapsed: 00:00:00.46
SQL> insert into rmt select level,level, sysdate from dual connect by level<=1e6;
100000... | 2020-10-11T09:29:36 | 2020-10-11T09:29:36 | {} | NONE |
ClickHouse | ClickHouse | 706,236,716 | 15,757 | filimonov | Now it's on top of #15786 | 2020-10-09T15:07:44 | 2020-10-09T15:07:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 706,358,716 | 15,454 | azat | > LAYOUT(HASHED());
<snip>
>Summary:
>PREALLOCATE() works faster
Note that you are using HASHED here, over SPARSE_HASHED, the difference for it should be even more significant | 2020-10-09T19:15:40 | 2020-10-09T19:15:40 | {} | MEMBER |
ClickHouse | ClickHouse | 706,464,504 | 15,775 | yisany | two node: 172-16-8-151 & 172-16-8-96
this is the output of the clickhouse-client on 151
```
172-16-8-151 :) select * from system.replicas where database='smart_3' and table='source_tags';\G
SELECT *
FROM system.replicas
WHERE (database = 'smart_3') AND (table = 'source_tags')
Row 1:
──────
database: ... | 2020-10-10T01:36:57 | 2020-10-10T01:36:57 | {} | NONE |
ClickHouse | ClickHouse | 706,583,526 | 15,609 | alexey-milovidov | BTW, on my machine it does not even start.
The fact that clickhouse local and client at least started indicates that there is a chance... | 2020-10-10T17:23:52 | 2020-10-10T17:23:52 | {} | MEMBER |
ClickHouse | ClickHouse | 706,655,647 | 15,809 | amosbird | > Those query_id_formats looks too specific though
The default is `Query id: xxx`.
The idea of `query_id_formats` is inspired from impala-client (so as presto, sparksql), which prints a query profiling url for every query. The url presents a nice dynamic DAG data flow, which excels our progress bar. We'll definit... | 2020-10-11T06:07:31 | 2020-10-11T06:15:36 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 706,187,909 | 15,752 | alesapin | Not real race | 2020-10-09T13:42:03 | 2020-10-09T13:42:03 | {} | MEMBER |
ClickHouse | ClickHouse | 706,462,155 | 15,787 | moweonlee | Thank you. @den-crane
Thank you for good references too.
| 2020-10-10T01:19:13 | 2020-10-10T01:19:13 | {} | NONE |
ClickHouse | ClickHouse | 706,470,391 | 15,775 | yisany | config.xml
```xml
<?xml version="1.0"?>
<yandex>
<logger>
<level>information</level>
<log>/data/logs/clickhouse/clickhouse-server.log</log>
<errorlog>/data/logs/clickhouse/clickhouse-server.err.log</errorlog>
<size>1000M</size>
<count>10</count>
</logger>
... | 2020-10-10T02:20:38 | 2020-10-10T02:20:38 | {} | NONE |
ClickHouse | ClickHouse | 706,541,869 | 15,802 | den-crane | >The table contains 14521 partitions and 119 columns.
Then slow start is expected. Especially on slow HDD or IOPS limited cloud disks (like AWS EBS GP1)
You should wait longer for a start minutes. I've heard that in case of 100000 CH can start up to 40 minutes.
This is expected by design. Because CH reads reads e... | 2020-10-10T12:30:11 | 2020-10-10T12:30:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 706,590,568 | 15,609 | bugaevc | Alright, here's the fix for the localtime bug:
```diff
diff --git a/src/kernel/emulation/linux/unistd/faccessat.c b/src/kernel/emulation/linux/unistd/faccessat.c
index cfe6c925c..f1ba2820b 100644
--- a/src/kernel/emulation/linux/unistd/faccessat.c
+++ b/src/kernel/emulation/linux/unistd/faccessat.c
@@ -15,7 +15... | 2020-10-10T18:18:31 | 2020-10-10T18:18:31 | {
"+1": 3
} | NONE |
ClickHouse | ClickHouse | 706,604,535 | 13,166 | ildus | @calebeaires Series, not Serie, and closed pull requests not a good place to put your issues. | 2020-10-10T20:08:02 | 2020-10-10T20:08:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 706,196,196 | 15,772 | tavplubix | Integration tests (release) - test_host_ip_change/test.py::test_ip_change_update_dns_cache - flacky | 2020-10-09T13:57:12 | 2020-10-09T13:57:12 | {} | MEMBER |
ClickHouse | ClickHouse | 706,472,454 | 15,802 | y45398jp123 | Thanks for your reply @den-crane
the table contains 14521 partitions and 119 columns.
do i have to wait for longer time after restart the server before trying to connect? have no idea...
Thanks. | 2020-10-10T02:38:42 | 2020-10-10T02:38:42 | {} | NONE |
ClickHouse | ClickHouse | 706,506,489 | 15,799 | alexey-milovidov | It helps to lower the number of broken links from 2 to 1. | 2020-10-10T07:48:06 | 2020-10-10T07:48:06 | {} | MEMBER |
ClickHouse | ClickHouse | 706,563,686 | 15,757 | filimonov | Yep. Can't reproduce that locally too, but from logs, it seems caused by some kind of race between consumer initialization and topic creation (done implicitly by the producer).
Since rdkafka 1.5 consumer don't create topic automatically and first attempt to start consuming can return error, and on the next attempt l... | 2020-10-10T15:08:08 | 2020-10-10T15:12:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 706,605,805 | 15,035 | alexey-milovidov | I'm Ok with the changes - merge as is. | 2020-10-10T20:19:53 | 2020-10-10T20:19:53 | {} | MEMBER |
ClickHouse | ClickHouse | 706,606,610 | 15,497 | alexey-milovidov | It does not pass clang-tidy static analyzer:
```
2020-10-10 19:47:51 /build/obj-x86_64-linux-gnu/../src/Functions/formatReadableTimeDelta.cpp:38:13: error: the 'empty' method should be used to check for emptiness instead of 'size' [readability-container-size-empty,-warnings-as-errors]
2020-10-10 19:47:51 i... | 2020-10-10T20:27:58 | 2020-10-10T20:27:58 | {} | MEMBER |
ClickHouse | ClickHouse | 706,611,857 | 15,766 | adevyatova | > никак там проблелы и запятые вообще не используются
Так было написано в исходном ПР. Исправила. | 2020-10-10T21:09:27 | 2020-10-10T21:09:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 706,503,714 | 15,805 | l1t1 | if i use insert values and optimize table, replace merge ok
```
insert into rmt values(-1,-1,now());
optimize table rmt;
select min(id),max(id) ,dt from rmt group by dt;
┌─min(id)─┬─max(id)─┬──────────────────dt─┐
│ 3 │ 9999999 │ 2020-10-10 15:05:47 │
│ 1 │ 9999996 │ 2020-10-10 15:06:48 │
│ ... | 2020-10-10T07:25:17 | 2020-10-10T07:25:17 | {} | NONE |
ClickHouse | ClickHouse | 706,580,675 | 15,609 | amosbird | More tests.
```
Darling [/Volumes/SystemRoot/home/amos/clickhouse-test]$ ps aux | tail -n +2 | awk '{ printf("%s\t%s\n", $1, $4) }' | ./clickhouse local -S "user String, mem Float64" -q "SELECT user, round(sum(mem), 2) as mem_t>
Processing configuration file 'config.xml'.
Include not found: clickhouse_remote_serv... | 2020-10-10T17:05:13 | 2020-10-10T17:05:13 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 706,583,529 | 15,609 | bugaevc | Hmm, there's something fishy going on for sure:
```
Darling [~]$ xtrace date
[82] open_nocancel("/usr/share/locale/en_GB.UTF-8/LC_TIME", O_RDONLY, 0) -> 3
[82] fstat64(3, 0x7fffffdfe9d8) -> 0
[82] read_nocancel(3, 0x7fae4c403d7c, 377) -> 377
[82] close_nocancel(3) -> 0
[82] gettimeofday(0x7fffffdfefe0, NULL, N... | 2020-10-10T17:23:56 | 2020-10-10T17:23:56 | {} | NONE |
ClickHouse | ClickHouse | 706,187,852 | 15,751 | alesapin | Not real race | 2020-10-09T13:41:57 | 2020-10-09T13:41:57 | {} | MEMBER |
ClickHouse | ClickHouse | 706,187,997 | 15,699 | alesapin | Not real race | 2020-10-09T13:42:12 | 2020-10-09T13:42:12 | {} | MEMBER |
ClickHouse | ClickHouse | 706,197,971 | 15,773 | tavplubix | ClickHouse build check - fails in 20.7 branch
Integration tests (release) - test_mysql_protocol/test.py::test_mysql_federated - flacky | 2020-10-09T14:00:33 | 2020-10-09T14:00:33 | {} | MEMBER |
ClickHouse | ClickHouse | 706,298,943 | 15,793 | den-crane | Never saw a such feature before.
Default for a type? Not a column? Null for any numeric with 0, Null for any date with 1970-01-01 ? | 2020-10-09T17:09:03 | 2020-10-09T17:09:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 706,615,690 | 15,809 | azat | >ClickHouse contains extraordinary capabilities to do introspecting, however, it requires query_id to access related metrics. With custom http handlers and this PR, it makes perf tuning much more pleasant.
Interesting, when I thought about this I've come up with [`previousQueryId`](https://github.com/azat/ClickHouse... | 2020-10-10T21:44:50 | 2020-10-10T21:48:15 | {} | MEMBER |
ClickHouse | ClickHouse | 706,617,553 | 15,497 | alexey-milovidov | Now it's 12 million rows/s. per CPU core. | 2020-10-10T22:04:45 | 2020-10-10T22:04:45 | {} | MEMBER |
ClickHouse | ClickHouse | 706,618,995 | 14,974 | filimonov | Same: https://clickhouse-test-reports.s3.yandex.net/15757/60f4e561fb986e51d831447cbf363934235953c3/fuzzer/report.html#fail1 | 2020-10-10T22:17:36 | 2020-10-10T22:17:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 706,188,425 | 15,620 | alesapin | Integration tests fixed in branches | 2020-10-09T13:43:03 | 2020-10-09T13:43:03 | {} | MEMBER |
ClickHouse | ClickHouse | 706,283,161 | 15,784 | den-crane | @zvonand do you mean current stable CH or about code with PR #9404 ? | 2020-10-09T16:36:42 | 2020-10-09T16:36:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 706,486,364 | 15,087 | zhouyao1994 | > @sundy-li Thanks.
>
> > 1. How to change the current zookeeper cluster to another one smoothly? We can't use zk copy to migrate the znodes. The only way I can think of is to copy all data from old tables to the new tables.
>
> True, we can't copy data of zookeeper cluster to another one. If the multi zookeeper ... | 2020-10-10T04:38:51 | 2020-10-10T04:38:51 | {} | NONE |
ClickHouse | ClickHouse | 696,481,040 | 15,087 | sundy-li | @fastio Good feature. We are also suffering from this problem， there are nearly 20 million znodes in our cluster, and high qps often leads `zxid` overflows.
But
1. How to change the current zookeeper cluster to another one smoothly? We can't use zk copy to migrate the znodes. The only way I can think of is t... | 2020-09-22T02:46:25 | 2020-10-10T07:19:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 706,505,307 | 15,757 | alesapin | Seems like `test_storage_kafka/test.py::test_kafka_produce_key_timestamp (✕5)` indeed flaky? | 2020-10-10T07:38:18 | 2020-10-10T07:38:18 | {} | MEMBER |
ClickHouse | ClickHouse | 706,531,062 | 15,776 | baibaichen | By reading comments from COW.h I can fix the null ptr
```c++
for (size_t i = 0; i < block.columns(); ++i)
{
MutableColumnPtr ptr = IColumn::mutate(std::move(res_block.getByPosition(i).column));
ptr->insertRangeFrom(*block.getByPosition(i).column.get(), 0, block.rows());
res_block.getByPosition(i).colu... | 2020-10-10T11:04:11 | 2020-10-10T11:04:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 706,610,122 | 15,497 | alexey-milovidov | Performance is 1.66 million rows/s. per CPU core,
```
12,45% libc-2.30.so [.] __GI___printf_fp_l
9,51% clickhouse [.] DB::(anonymous namespace)::FunctionFormatReadableTimeDelta::formatReadableTimeDelta
9,14% clickhouse [.... | 2020-10-10T20:53:34 | 2020-10-10T20:53:34 | {} | MEMBER |
ClickHouse | ClickHouse | 706,677,298 | 15,832 | y45398jp123 | looks like a problem caused by networking, could you check the config.xml and user.xml?
the <listener_host> tag and <ip> tag? | 2020-10-11T09:34:57 | 2020-10-11T09:34:57 | {} | NONE |
ClickHouse | ClickHouse | 706,470,595 | 15,802 | den-crane | 40 billion records -- it's not big table. There are installation with 40 trillions.
CH start time depends on number of parts (table partitioning).
Can you show
`select count() from system.parts`?
`select count() from system.columns where tables = ....` ? | 2020-10-10T02:22:20 | 2020-10-10T02:23:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 706,507,220 | 15,805 | l1t1 | insert value is not always work
```
insert into rmt values(-1,-1,now());
optimize table rmt;
select count(distinct id),count(id) from rmt;
┌─uniqExact(id)─┬─count(id)─┐
│ 6400001 │ 8900002 │
└───────────────┴───────────┘
``` | 2020-10-10T07:54:35 | 2020-10-10T07:56:07 | {} | NONE |
ClickHouse | ClickHouse | 706,531,785 | 15,776 | alexey-milovidov | As the test is not automated (It's not a unit test, it's not a functional test. I wouldn't even say that it's a test - it's just an example program), the simplest solution is to remove it. FinishSorting transform is well covered by functional tests. | 2020-10-10T11:10:08 | 2020-10-10T11:10:08 | {} | MEMBER |
ClickHouse | ClickHouse | 706,665,420 | 15,833 | l1t1 | the execute plan for two sql
```
explain select count(*) from rmt where (id,name,dt)in (select id,name,dt from mt);
┌─explain──────────────────────────────────────────────────┐
│ Expression (Projection) │
│ Expression (Before ORDER BY and SELECT) │
│ Aggre... | 2020-10-11T07:46:34 | 2020-10-11T09:13:28 | {} | NONE |
ClickHouse | ClickHouse | 706,188,609 | 15,756 | CurtizJ | > Does it also fix #13993 ?
No. | 2020-10-09T13:43:22 | 2020-10-09T13:43:22 | {} | MEMBER |
ClickHouse | ClickHouse | 706,188,648 | 15,618 | alesapin | Integration test failure not related | 2020-10-09T13:43:28 | 2020-10-09T13:43:28 | {} | MEMBER |
ClickHouse | ClickHouse | 706,194,713 | 9,434 | alexey-milovidov | Yes. It can return constant UInt8 column with value zero. | 2020-10-09T13:54:23 | 2020-10-09T13:54:23 | {} | MEMBER |
ClickHouse | ClickHouse | 706,230,054 | 15,265 | bohutang | You can check it with `change master to ...` using this user | 2020-10-09T14:55:27 | 2020-10-09T14:55:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 706,556,156 | 15,810 | DaiDaDao | And I changed my Zk-cluster config , restart the CK-cluster, DDL opt task-queue always in the old ZK.... | 2020-10-10T14:18:35 | 2020-10-10T14:18:35 | {} | NONE |
ClickHouse | ClickHouse | 706,583,987 | 15,609 | alexey-milovidov | But it cannot access /etc/localtime:
```
[450] access("/etc/localtime", 4) -> ENOENT
``` | 2020-10-10T17:26:44 | 2020-10-10T17:26:44 | {} | MEMBER |
ClickHouse | ClickHouse | 706,638,149 | 15,805 | l1t1 | adding final clause also not solved the problem yet.
```
select count(distinct id),count(id) from rmt;
┌─uniqExact(id)─┬─count(id)─┐
│ 7571429 │ 8357143 │
└───────────────┴───────────┘
1 rows in set. Elapsed: 1.442 sec. Processed 8.36 million rows, 33.43 MB (5.80 million rows/s., 23.19 MB/s.)
optimiz... | 2020-10-11T02:05:16 | 2020-10-11T02:05:16 | {} | NONE |
ClickHouse | ClickHouse | 706,648,120 | 15,497 | filipecaixeta | Thanks for the help =D | 2020-10-11T04:26:59 | 2020-10-11T04:26:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 709,953,242 | 16,036 | tavplubix | Integration tests - test_multiple_disks - probably wrong exception code:
```
2020.10.16 05:40:14.403111 [ 43 ] {b3e6aaa6-68de-44bd-bc29-6d8ea66aeaab} <Fatal> : Logical error: 'Destination table default.test_move_across_policies_does_not_work2 (e2b0fa63-fc75-48e1-8ecc-9d9a0337e475) should have the same storage policy ... | 2020-10-16T10:05:15 | 2020-10-16T10:05:15 | {} | MEMBER |
ClickHouse | ClickHouse | 710,052,224 | 16,043 | filimonov | BTW
```
Fetching submodule contrib/grpc/third_party/boringssl ...
```
I think we may need to adjust that to avoid having 2 copies. | 2020-10-16T13:37:29 | 2020-10-16T13:37:29 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 710,461,738 | 16,041 | alexey-milovidov | @amosbird @KochetovNicolai I suddenly remember that I wanted to name this function `untuple`.
(when I was thinking about this function very long time ago)
Why?
Because tupleFlatten is too similar name to arrayFlatten. And arrayFlatten does deep flattening of arrays, like
`arrayFlatten([[[1], []], [[2, 3]]]) =... | 2020-10-16T19:18:28 | 2020-10-16T19:19:40 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.