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 | 506,357,704 | 5,772 | alexey-milovidov | Nice feature request! By the way, it is very easy to implement.
PS. Many people wonder, what is the practical use case for the `throwIf`. Can you share, please? | 2019-06-27T13:58:23 | 2019-06-27T13:58:37 | {} | MEMBER |
ClickHouse | ClickHouse | 506,409,924 | 5,755 | alexey-milovidov | > There're places where we check isColumnNullable() and then static_cast it to ColumnNullable.
The method is named `isColumnNullable` exactly for this reason (to check that the column is an instance of `ColumnNullable`). You can rename to `isNullable` for your changes to be consistent. | 2019-06-27T16:06:02 | 2019-06-27T16:06:02 | {} | MEMBER |
ClickHouse | ClickHouse | 506,526,856 | 5,692 | filimonov | Fixed in https://github.com/yandex/ClickHouse/issues/5774 | 2019-06-27T21:59:05 | 2019-06-27T21:59:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 506,592,205 | 1,980 | hwchen | Thanks, this is useful to me. It could be a workaround in special
cases, but I probably wouldn't want to depend on it more.
On Mon, Jun 24, 2019 at 4:33 AM Alexander Rodin
<notifications@github.com> wrote:
>
> @hwchen It is possible to COALESCE nullable values to some "empty" value, then it would work with groupArray ... | 2019-06-28T04:07:22 | 2019-06-28T04:07:22 | {} | NONE |
ClickHouse | ClickHouse | 506,691,212 | 5,756 | alesapin | Fix https://github.com/yandex/ClickHouse/pull/5786. | 2019-06-28T10:47:01 | 2019-06-28T10:47:01 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 506,739,201 | 5,783 | den-crane | >Cannot execute query in readonly mode.
This is about user privileges. Probably current user is RO at leader replica because "drop partition" replicate-able. | 2019-06-28T13:44:17 | 2019-06-28T13:44:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 506,818,659 | 5,787 | akuzm | The performance test failure seems to be not related to the changes. I restarted the tests: https://nda.ya.ru/3UxsBF | 2019-06-28T17:39:40 | 2019-06-28T17:39:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 506,961,557 | 5,798 | alexey-milovidov | Leaked submodule. | 2019-06-29T14:36:57 | 2019-06-29T14:36:57 | {} | MEMBER |
ClickHouse | ClickHouse | 507,008,427 | 5,794 | blinkov | > а не другого языка
@davydovska имелся ввиду в первую очередь английский https://github.com/yandex/ClickHouse/blob/master/docs/en/query_language/agg_functions/reference.md
| 2019-06-30T05:13:51 | 2019-06-30T05:13:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 507,064,211 | 5,805 | alexey-milovidov | I've noticed that the order of coordinates was not consistent with existing functions. I have to change it.
https://stackoverflow.com/questions/7309121/preferred-order-of-writing-latitude-longitude-tuples | 2019-06-30T20:16:17 | 2019-06-30T20:16:17 | {} | MEMBER |
ClickHouse | ClickHouse | 507,071,559 | 5,387 | alexey-milovidov | @proller will update Poco library. | 2019-06-30T22:10:08 | 2019-06-30T22:10:08 | {} | MEMBER |
ClickHouse | ClickHouse | 506,141,225 | 5,618 | Aricliang | ┌─name─────────────────────────────────────┬─value───────┬
│ distributed_aggregation_memory_efficient │ 1 │
│ max_bytes_before_external_group_by │ 8000000000 │
│ max_bytes_before_external_sort │ 8000000000 │
│ max_memory_usage │ 16000000000 │
│ max_memory_usage_f... | 2019-06-27T03:23:56 | 2019-06-27T03:24:53 | {} | NONE |
ClickHouse | ClickHouse | 506,389,414 | 5,369 | danlark1 | https://github.com/yandex/ClickHouse/pull/5775 | 2019-06-27T15:15:33 | 2019-06-27T15:15:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 506,967,296 | 5,464 | alexey-milovidov | Reopened, because new tgz packages have not tested.
I don't see install script inside. | 2019-06-29T15:55:57 | 2019-06-29T15:55:57 | {} | MEMBER |
ClickHouse | ClickHouse | 506,086,776 | 5,761 | den-crane | To achieve even better compression rate check CODECs https://clickhouse.yandex/docs/en/query_language/create/#column-compression-codecs
And Low Cardinality (Strings) https://www.altinity.com/blog/2019/3/27/low-cardinality
| 2019-06-26T23:59:31 | 2019-06-26T23:59:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 506,368,202 | 5,772 | thyn | We use clickhouse for log storage and monitoring
current query
```sql
select concat(r0,r1) as text, if (text='','SUCCESS', text) as result from(
select count(1) as logMessages,
if (logMessages=0,'no logs','') as r0,
countIf(message='connection error') as errors,
if (errors>0,'to many errors','') as r1 from l... | 2019-06-27T14:24:11 | 2019-06-27T14:24:43 | {} | NONE |
ClickHouse | ClickHouse | 506,588,631 | 5,724 | MironAtHome | Я вот шел и думал, что наваял здесь нехорошего. Но в Вашем возрасте, молодые люди, пора знать, что такое "источники". Это то, что Ваша компания хранит здесь, на этом самом ресурсе.
Может Вы так выразили свой сарказм. Не знаю.
Но мало того, что эти самые источники на английском, так ещё и хранятся не на собственном Ян... | 2019-06-28T03:42:41 | 2019-06-28T03:42:41 | {} | NONE |
ClickHouse | ClickHouse | 506,988,708 | 5,794 | alexey-milovidov | Давайте сделаем такое же изменение для остальных языков и можно мержить. | 2019-06-29T21:34:42 | 2019-06-29T21:34:42 | {} | MEMBER |
ClickHouse | ClickHouse | 507,106,452 | 5,809 | maskshell | :````) this setting is still in testing ... not committed in stable release yet
but it truly affected the current edition. | 2019-07-01T03:52:39 | 2019-07-01T03:52:39 | {} | NONE |
ClickHouse | ClickHouse | 506,372,420 | 5,761 | blind-oracle | > Both tables filled with same data using: INSERT INTO X SELECT * FROM remote(...) LIMIT 100000000
Are you sure the data is the same? If rows came in different order it might be different. Can you seed one server first and then the second from the first?
Also try specifying the codec manually when creating the ta... | 2019-06-27T14:34:37 | 2019-06-27T14:34:37 | {} | NONE |
ClickHouse | ClickHouse | 506,653,713 | 5,771 | mfridental | I can build now, thanks! | 2019-06-28T08:38:06 | 2019-06-28T08:38:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 506,707,143 | 5,742 | alexey-milovidov | > Code: 202. DB::Exception: Received from localhost:9000, ::1. DB::Exception: Too many simultaneous queries. Maximum: 100.
FYI. This message cannot relate to stress test failure. It fails only in case of sanitizer reports or server crash. | 2019-06-28T11:53:48 | 2019-06-28T11:53:58 | {} | MEMBER |
ClickHouse | ClickHouse | 506,960,058 | 5,784 | alexey-milovidov | > Allocator: Cannot mmap 128.00 MiB., errno: 12, strerror: Cannot allocate memory
This error clearly states that you don't have enough RAM.
Not a bug.
> count(distinct ring) uniq
You can replace to `uniqCombined(ring) AS uniq` to save memory. | 2019-06-29T14:16:01 | 2019-06-29T14:16:01 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 507,045,925 | 5,804 | den-crane | toInt32
SELECT intDiv(toInt32(toStartOfMinute(toDateTime('2019-06-29 00:00:00'))), 60);
It's not because of TZ. And UCT it's the same as UTC (just a symlink in tzdata2019.2) and can be overridden by TZ variable (new feature of CH). | 2019-06-30T15:37:20 | 2019-06-30T15:37:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 506,093,097 | 5,761 | ramazanpolat | Ok, there is something wrong here.
I've created same tables with same settings and put same data on both installation.
Even if they have exact same setting, table size differs considerably. **It is 0.7 GB vs 9.4 GB**.
On one server:
```sql
ch1:) SHOW CREATE TABLE X
CREATE TABLE default.X (...)
ENGINE = MergeTre... | 2019-06-27T00:33:45 | 2019-06-27T00:33:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 506,429,486 | 5,764 | alexey-milovidov | Do you want to make something related to this pull request? | 2019-06-27T17:00:55 | 2019-06-27T17:00:55 | {} | MEMBER |
ClickHouse | ClickHouse | 506,649,391 | 5,756 | alesapin | > I've seen same errors on other codec when getMaxCompressedDataSize() returned wrong value. It probably should returns
>
> ```
> + 2 * sizeof(T)
> ```
Delta header size is actually two bytes regardless of data type: https://github.com/yandex/ClickHouse/blob/master/dbms/src/Compression/CompressionCodecDelta.cpp... | 2019-06-28T08:24:13 | 2019-06-28T08:24:13 | {} | MEMBER |
ClickHouse | ClickHouse | 506,870,581 | 5,724 | filimonov | git - децентрализованеая система. Каждый кто выполнил команду git clone имеет на своём компьютере полную копию всех источников с полной историей и в течение минут можно с единственного компьютера восстановить весь репозитарий.
Github и английский язык это просто удобные инструменты для международного общения (наприм... | 2019-06-28T20:42:14 | 2019-06-28T20:43:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 506,960,651 | 5,781 | alexey-milovidov | Now you can use `IPv6` data type, that is designed specifically for IPv6 addresses.
To transform String or FixedString(16) to IPv6, write `toIPv6(x)`.
You can also transform `FixedString(16)` to text String with `IPv6NumToString(x)`. | 2019-06-29T14:23:42 | 2019-06-29T14:23:42 | {} | MEMBER |
ClickHouse | ClickHouse | 506,974,647 | 5,687 | alexey-milovidov | Sorry for the delay, @abyss7 forgot about this task. | 2019-06-29T17:41:00 | 2019-06-29T17:41:00 | {} | MEMBER |
ClickHouse | ClickHouse | 506,359,662 | 5,771 | alexey-milovidov | If you don't use the `compile` and `compile_expressions` settings (they are disabled by default), you can disable LLVM by adding
`-D ENABLE_EMBEDDED_COMPILER=0` to your `cmake` invocation. | 2019-06-27T14:03:03 | 2019-06-27T14:03:23 | {} | MEMBER |
ClickHouse | ClickHouse | 506,363,657 | 5,717 | alexey-milovidov | Ok, thank you!
PS.
```
milovidov@milovidov-Latitude-E7470:~$ docker run --rm -it ubuntu:bionic ldd --version
ldd (Ubuntu GLIBC 2.27-3ubuntu1) 2.27
Copyright (C) 2018 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY ... | 2019-06-27T14:13:10 | 2019-06-27T14:13:10 | {} | MEMBER |
ClickHouse | ClickHouse | 506,380,013 | 5,376 | akuzm | After reading more about this, it seems that pgbouncer does not support prepared statements in transaction pooling mode, so the error is expected.
I tried installing psqlodbc and configuring a simple data source:
```
# cat /etc/odbc.ini
[postgres]
Driver = /usr/lib/x86_64-linux-gnu/odbc/psqlodbcw.so
Servername =... | 2019-06-27T14:52:46 | 2019-06-27T14:52:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 506,483,995 | 5,755 | 4ertus2 | Semantic changes are reverted.
PR removes ```IColumn.isColumnNullable()``` and ```IColumn.isColumnConst()```, moves ```isNullable()``` and ```isColumnConst()``` outside of IColumn and checks old ```x.isColumnNullable()``` with ```checkColumn<ColumnNullable>(x)```. ```isNullable()``` returns true for ColumnNullable a... | 2019-06-27T19:43:20 | 2019-06-27T19:46:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 506,539,123 | 5,761 | ramazanpolat | @den-crane
> can you show
>
> select uniq(desc) from ...
> from both servers ?
Sorry, I dropped tables and restarted CH.
After restarting CH, I cannot reproduce the problem. I will update this thread if I can repro. | 2019-06-27T22:51:57 | 2019-06-27T22:51:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 506,953,902 | 5,796 | alexey-milovidov | > By the way why not use requirements.txt for Python dependencies? @alexey-milovidov
No idea, I'm not a Python developer. Do what you want. | 2019-06-29T12:37:01 | 2019-06-29T12:37:01 | {} | MEMBER |
ClickHouse | ClickHouse | 507,059,415 | 5,212 | alexey-milovidov | @vitlibar went on vacation. I will reassign this task to someone else. | 2019-06-30T19:07:21 | 2019-06-30T19:07:21 | {} | MEMBER |
ClickHouse | ClickHouse | 506,220,406 | 5,717 | Felixoid | Fixed in aur:clickhouse-static | 2019-06-27T07:04:17 | 2019-06-27T07:04:17 | {} | MEMBER |
ClickHouse | ClickHouse | 506,361,716 | 5,717 | Felixoid | ```
ldd (GNU libc) 2.29
Copyright (C) 2019 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
Written by Roland McGrath and Ulrich Drepper.
``` | 2019-06-27T14:08:13 | 2019-06-27T14:08:13 | {} | MEMBER |
ClickHouse | ClickHouse | 506,368,354 | 5,717 | alexey-milovidov | > This issue is not related to gcc-9.
It is related to the newer version of libc in your system.
Actually I was wrong. The issue is related to more strict attributes for builtin function in gcc-9. | 2019-06-27T14:24:35 | 2019-06-27T14:24:35 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 506,373,312 | 5,772 | alexey-milovidov | What version of ClickHouse do you use and what is the result of the following commands?
```
~/ClickHouse$ clickhouse local --query "SELECT throwIf(1)"
Code: 395, e.displayText() = DB::Exception: Value passed to 'throwIf' function is non zero (version 19.10.1.671 (official build))
~/ClickHouse$ clickhouse local --... | 2019-06-27T14:36:44 | 2019-06-27T14:36:44 | {} | MEMBER |
ClickHouse | ClickHouse | 506,404,841 | 5,764 | alexey-milovidov | Sorry, isn't it what you need?: https://github.com/yandex/ClickHouse/pull/4555 | 2019-06-27T15:53:41 | 2019-06-27T15:53:41 | {} | MEMBER |
ClickHouse | ClickHouse | 506,405,841 | 5,764 | alexey-milovidov | What you can do?
- add `arrayFlatten` as a synonim to `flatten`;
- add documentation (probably missing documentation is the issue);
- add your tests just in case;
- compare implementation;
- compare performance; | 2019-06-27T15:56:02 | 2019-06-27T15:56:02 | {} | MEMBER |
ClickHouse | ClickHouse | 506,415,749 | 5,761 | den-crane | can you show
select uniq(desc) from ...
from both servers ? | 2019-06-27T16:21:51 | 2019-06-27T16:21:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 506,450,040 | 5,777 | alexey-milovidov | notEmpty. | 2019-06-27T18:01:09 | 2019-06-27T18:01:09 | {} | MEMBER |
ClickHouse | ClickHouse | 506,930,839 | 5,794 | blinkov | @davydovska а почему это важно? в любом случае если править, то и остальные языки тоже | 2019-06-29T06:07:28 | 2019-06-29T06:07:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 506,961,552 | 5,798 | alexey-milovidov | Missing changelog. | 2019-06-29T14:36:51 | 2019-06-29T14:36:51 | {} | MEMBER |
ClickHouse | ClickHouse | 507,062,773 | 5,805 | alexey-milovidov | This library is extraordinary slow: less than a million invocations per second:
```
11.14% libm-2.27.so [.] __lroundl
10.28% libm-2.27.so [.] __sincos
8.07% clickhouse [.] _ijkNormalize
7.90% ... | 2019-06-30T19:55:34 | 2019-06-30T19:55:34 | {} | MEMBER |
ClickHouse | ClickHouse | 506,360,024 | 5,761 | blind-oracle | Which columns are affected? Show `SHOW CREATE TABLE` from both servers and `ls -l` from part folder. I don't think that CH versions difference can lead to 6x increase in disk usage, I've not seen this behavior. | 2019-06-27T14:03:57 | 2019-06-27T14:04:39 | {} | NONE |
ClickHouse | ClickHouse | 506,367,088 | 5,761 | ramazanpolat | Ok.
Both tables filled with same data using:
`INSERT INTO X SELECT * FROM remote(...) LIMIT 100000000`
### Server ch1:
```
SELECT version()
┌─version()─┐
│ 19.9.2.4 │
└───────────┘
SHOW CREATE TABLE X
CREATE TABLE default.X (
`cid` UInt64,
`keyword` String,
`ts` DateTime DEFAULT now(),
`rank` ... | 2019-06-27T14:21:30 | 2019-06-27T14:22:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 506,376,999 | 5,761 | ramazanpolat | I'm pretty sure that both tables have same data because I've stopped inserts on the source table on remote server.
This is the sample data.
```
SELECT domain
FROM X
LIMIT 10
┌─domain──────┐
│ otoboom │
│ arabam │
│ arabalar │
│ sifiraracal │
│ takvim │
│ sabah │
│ sahibinden │
... | 2019-06-27T14:45:36 | 2019-06-27T14:45:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 506,815,177 | 5,784 | den-crane | 1. add 8Gb of swap.
2. change CH cache size to 500MB `<mark_cache_size>536870912</mark_cache_size>`
3. change CH user memory settings
```
<yandex>
<profiles>
<default>
<max_bytes_before_external_group_by>2000000000</max_bytes_before_external_group_by>
<max_memory_usage>300000... | 2019-06-28T17:28:16 | 2019-06-28T17:29:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 506,906,141 | 4,469 | kaushik141 | As a work around you can,
1. DETACH the table
2. Modify the table .sql file in metadata to reflect the modification on the column of your choice
3. ATTACH the table | 2019-06-28T23:40:07 | 2019-06-28T23:40:07 | {} | NONE |
ClickHouse | ClickHouse | 506,504,926 | 5,717 | excitoon | Thank you! | 2019-06-27T20:47:17 | 2019-06-27T20:47:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 506,520,491 | 5,618 | den-crane | >should not use the disk? Why still reporting memory limits?
Because final aggregation result should fit into memory.
So for group by SOME_HIGH_CARDINALITY_SET_OF_COLUMNS, ch will do intermediate partial aggregations on a disk, but in the end CH will gather the result into one part in memory. | 2019-06-27T21:35:06 | 2019-06-27T23:57:53 | {
"+1": 2,
"eyes": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 505,770,405 | 5,750 | hczhcz | By the way, we maintain a fork of ClickHouse with some custom functions. We use these shell scripts to generate the files:
```sh
cd dbms/src/AggregateFunctions
{
echo '#include <AggregateFunctions/registerAggregateFunctions.h>'
echo
echo '#include <AggregateFunctions/AggregateFunctionFactory.h>'
... | 2019-06-26T08:13:54 | 2019-06-28T03:41:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 506,719,207 | 5,724 | alexey-milovidov | Имеются ввиду вопросы:
1. "Зачем вы храните исходные коды ClickHouse на сервисе GitHub, принимая во внимание, что этот сервис принадлежит американской корпорации и не имеет отношения к Яндексу, что может вызывать опасения из-за торговых войн?"
2. "Почему мы предпочитаем вести обсуждения по поводу разработки проду... | 2019-06-28T12:39:55 | 2019-06-28T12:39:55 | {} | MEMBER |
ClickHouse | ClickHouse | 506,814,056 | 5,792 | proller | now warning in clang:
../dbms/src/IO/WriteBufferAIO.cpp:277:54: error: result of comparison 'ssize_t' (aka 'long') > 9223372036854775807 is always false [-Werror,-Wtautological-type-limit-compare]
if ((static_cast<ssize_t>(flush_buffer.offset()) > std::numeric_limits<off_t>::max()) ||
~~~~~~~~~~~~~~~~~~... | 2019-06-28T17:24:37 | 2019-06-28T17:24:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 507,051,430 | 5,804 | den-crane | UCT is not typo, it's a legacy TZ name. And has no relation to CH.
It's just a symlink in a new version of Linux tzdata.
```
# readlink -f /etc/localtime
/usr/share/zoneinfo/UCT
# dpkg -l 'tzdata*'
||/ Name Version Architecture Description
+++-==============-============-============-=========... | 2019-06-30T17:00:08 | 2019-06-30T17:06:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 506,231,945 | 5,761 | blind-oracle | @ramazanpolat
* Do `OPTIMIZE TABLE ... FINAL` to merge all parts together. Usually it decreases space usage
* Use `bytes_on_disk` in `system.parts` to see the on-disk usage | 2019-06-27T07:41:50 | 2019-06-27T07:41:50 | {} | NONE |
ClickHouse | ClickHouse | 506,310,230 | 5,763 | 4ertus2 | You probably use some old version. Current master returns 15.00 | 2019-06-27T11:37:04 | 2019-06-27T11:37:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 506,335,986 | 5,771 | proller | Can you paste all commands with all flags (cmake .....) and all console output with errors? | 2019-06-27T13:01:19 | 2019-06-27T13:01:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 506,650,605 | 5,480 | alesapin | Can you provide table definition and data to reproduce the problem? | 2019-06-28T08:28:16 | 2019-06-28T08:28:16 | {} | MEMBER |
ClickHouse | ClickHouse | 506,950,588 | 5,796 | excitoon | By the way why not use `requirements.txt` for Python dependencies? @alexey-milovidov | 2019-06-29T11:36:08 | 2019-06-29T11:36:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 506,657,303 | 5,480 | kolsys | Broken part was deleted manually with new partition renaming. I can't provide data sample.
```
CREATE TABLE sample (
`date` Date,
`time` DateTime,
`host` String,
`name` String,
`facility` UInt8,
`severity` UInt8,
`pid` UInt32,
`cp` UInt16 CODEC(Delta(2), ZSTD(1)),
`sp` UInt16 CODEC(Delta(... | 2019-06-28T08:49:43 | 2019-06-28T08:54:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 506,693,166 | 5,739 | shodanium | As the original author, I hereby grant the permission to use all the GEODIST() implementation related functions under the MIT license.
Let me know if you need anything even more official :) | 2019-06-28T10:55:05 | 2019-06-28T10:55:05 | {} | NONE |
ClickHouse | ClickHouse | 507,049,948 | 5,804 | simPod | Yup, I know I have to convert it to Int for it to work, thanks.
Just `UCT` seemed like a typo. Do you know a docs/PR where I can learn more about the feature? | 2019-06-30T16:35:52 | 2019-06-30T16:35:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 507,052,081 | 5,804 | den-crane | And CH timezone behavior can be overridden by
Use contents of environment variable TZ as the name for timezone.
It helps to correctly detect default timezone in some cases. https://github.com/yandex/ClickHouse/pull/5443 (Ivan) | 2019-06-30T17:10:16 | 2019-06-30T17:10:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 507,060,322 | 5,384 | alexey-milovidov | Looks like #5387 is the right PR. | 2019-06-30T19:19:45 | 2019-06-30T19:19:45 | {} | MEMBER |
ClickHouse | ClickHouse | 506,467,955 | 5,045 | alexey-milovidov | > Any news concerning this PR ?
We are currently working on this PR, trying to squeeze missing bits of performance. Results are optimistic. Now it's not about 30% slower, but just 5%. But we will try to break even. | 2019-06-27T18:52:11 | 2019-06-27T18:53:11 | {} | MEMBER |
ClickHouse | ClickHouse | 506,651,349 | 5,464 | alesapin | We have turned them back! First testing version (19.11.0.736) can be downloaded here https://repo.yandex.ru/clickhouse/tgz/. Each next testing/prestable/stable release can be downloaded in the same repo. | 2019-06-28T08:30:35 | 2019-06-28T08:30:35 | {} | MEMBER |
ClickHouse | ClickHouse | 506,652,243 | 5,316 | Felixoid | Looks like it was fixed. Maybe in #5368
Is it still broken for you, @excitoon ? | 2019-06-28T08:33:30 | 2019-06-28T08:33:30 | {} | MEMBER |
ClickHouse | ClickHouse | 506,699,075 | 5,480 | alesapin | We had memory corruption for columns less than 32 bits size with Delta codec. In your definition
```
cp UInt16 CODEC(Delta(2), ZSTD(1)),
sp UInt16 CODEC(Delta(2), ZSTD(1)),
```
Probably it was the root cause of your problem. Fix here https://github.com/yandex/ClickHouse/pull/5786. | 2019-06-28T11:19:44 | 2019-06-28T11:19:55 | {} | MEMBER |
ClickHouse | ClickHouse | 506,756,361 | 5,783 | ghost | @den-crane
How can I check whether my current user is RO at leader replica? | 2019-06-28T14:34:14 | 2019-06-28T14:34:14 | {} | NONE |
ClickHouse | ClickHouse | 506,960,156 | 5,784 | alexey-milovidov | Also check overcommit settings: https://clickhouse.yandex/docs/en/operations/tips/#ram | 2019-06-29T14:17:11 | 2019-06-29T14:17:11 | {} | MEMBER |
ClickHouse | ClickHouse | 506,960,453 | 5,782 | alexey-milovidov | The syntax `countIf(DISTINCT(x) < 5)` is not supported - neither in ClickHouse or any other DBMS.
You can write `uniqExactIf(x, x < 5)`. | 2019-06-29T14:20:57 | 2019-06-29T14:20:57 | {
"+1": 3
} | MEMBER |
ClickHouse | ClickHouse | 506,358,724 | 5,761 | ramazanpolat | @blind-oracle Thank you for your suggestions.
* Just checked configs and they are both identical, word by word, letter by letter
* Adding `WHERE active` didn't change the result, it's still 6 GB.
* Found difference between column sizes but this does not explain the reason.
- [x] Identical servers
- [x] Ident... | 2019-06-27T14:00:53 | 2019-06-27T14:00:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 506,651,821 | 5,169 | Felixoid | Looks like it was fixed somewhere in the past... maybe in #5368 | 2019-06-28T08:32:03 | 2019-06-28T08:32:03 | {} | MEMBER |
ClickHouse | ClickHouse | 506,327,726 | 5,761 | ramazanpolat | @blind-oracle
Even after `OPTIMIZE TABLE X FINAL`, the size is **6 GB** vs **0.7 GB**.
```sql
SELECT
table,
sum(bytes_on_disk) / ((1024 * 1024) * 1024) AS GB_Size
FROM system.parts
GROUP BY table
┌─table──┬────────────GB_Size─┐
│ X │ 6.027186650782824 │
└────────┴────────────────────┘
``` | 2019-06-27T12:36:43 | 2019-06-27T12:36:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 506,458,896 | 5,750 | alexey-milovidov | Ok. The change is really insignificant.
About these shell scripts. I'm not sure that we need adopt them.
It can remove multiple similar files, but the code will be slightly less readable and the build process is slightly more complicated - IMO. | 2019-06-27T18:27:09 | 2019-06-27T18:27:09 | {} | MEMBER |
ClickHouse | ClickHouse | 506,959,844 | 5,797 | alexey-milovidov | To solve, introduce `merge_table_num_streams_multiplier` setting with default value `4`.
And multiply `num_streams` by minimum of this value and the number of source tables in `StorageMerge::read`. | 2019-06-29T14:12:53 | 2019-06-29T14:13:34 | {} | MEMBER |
ClickHouse | ClickHouse | 506,988,614 | 1,344 | alexey-milovidov | It will be fixed by this PR: #5042 | 2019-06-29T21:32:55 | 2019-06-29T21:32:55 | {
"hooray": 4
} | MEMBER |
ClickHouse | ClickHouse | 507,064,368 | 5,798 | s-mx | Check it now, please. | 2019-06-30T20:18:13 | 2019-06-30T20:18:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 506,329,801 | 5,761 | blind-oracle | @ramazanpolat
* Take a look at the configs of servers, maybe you have different compression codecs.
* Your query also returns inactive parts, add `WHERE active`
* Take a look inside the part folder on disk after OPTIMIZE to see which columns takes the most space | 2019-06-27T12:43:21 | 2019-06-27T12:43:21 | {} | NONE |
ClickHouse | ClickHouse | 506,342,063 | 5,771 | mfridental | This is what cmake .. says:
-- The C compiler identification is GNU 7.4.0
-- The CXX compiler identification is GNU 7.4.0
-- Check for working C compiler using: Ninja
-- Check for working C compiler using: Ninja -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Detecting C comp... | 2019-06-27T13:18:10 | 2019-06-27T13:18:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 506,355,715 | 5,717 | alexey-milovidov | This issue is not related to gcc-9.
It is related to the newer version of libc in your system.
What is the output of `ldd --version`? | 2019-06-27T13:53:21 | 2019-06-27T13:53:21 | {} | MEMBER |
ClickHouse | ClickHouse | 506,386,402 | 5,772 | thyn | It's ok. I think I was a little confused reading description. Right version is
```
select throwIf(logMessages=0,'no logs') as r0,throwIf (errors>0,'to many errors','') as r1 from logs
```
It's working (without message)
| 2019-06-27T15:07:50 | 2019-06-27T15:08:52 | {} | NONE |
ClickHouse | ClickHouse | 506,471,674 | 5,758 | alexey-milovidov | Perf test changes are definitely unrelated. | 2019-06-27T19:03:16 | 2019-06-27T19:03:16 | {} | MEMBER |
ClickHouse | ClickHouse | 506,705,152 | 5,785 | alexey-milovidov | Ok, thank you! | 2019-06-28T11:45:29 | 2019-06-28T11:45:29 | {} | MEMBER |
ClickHouse | ClickHouse | 506,731,108 | 5,710 | alexey-milovidov | Mostly Ok, only a few changes remain. | 2019-06-28T13:20:03 | 2019-06-28T13:20:03 | {} | MEMBER |
ClickHouse | ClickHouse | 506,783,574 | 5,717 | mariusa | @alexey-milovidov Thanks! How about removing gcc7 from https://clickhouse.yandex/docs/en/development/build/ ? | 2019-06-28T15:51:17 | 2019-06-28T15:51:17 | {} | NONE |
ClickHouse | ClickHouse | 506,785,512 | 5,789 | den-crane | probably appearances of this issue
market.yandex.net : https://t.me/clickhouse_ru/106178 https://t.me/clickhouse_ru/106179
vk.com : https://t.me/clickhouse_ru/103897 https://t.me/clickhouse_ru/104347
DB::Exception: Incorrect mark: (0,0,8192) or (0,0) in data, (0,8,8192) in /var/lib/clickhouse/data/local/li... | 2019-06-28T15:56:59 | 2019-06-28T15:56:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 506,804,570 | 1,469 | cameronbraid | > My experience is xxxState (sumState, countState) will only work across 1-dimension, if you group using more than 1 columns, then the runningAccumulate will get messed up (not in sequence).
I have noticed this too.
Is there any way to make runningAccumulate work with more than 1 group by column ? | 2019-06-28T16:55:35 | 2019-06-28T16:55:35 | {
"+1": 2
} | NONE |
ClickHouse | ClickHouse | 506,810,677 | 5,717 | alexey-milovidov | @mariusa Ok. | 2019-06-28T17:14:23 | 2019-06-28T17:14:23 | {} | MEMBER |
ClickHouse | ClickHouse | 506,960,891 | 5,799 | alexey-milovidov | @Enmk | 2019-06-29T14:26:59 | 2019-06-29T14:26:59 | {} | MEMBER |
ClickHouse | ClickHouse | 506,985,659 | 5,794 | davydovska | > @davydovska а почему это важно? в любом случае если править, то и остальные языки тоже
я заметила, что userid со строчных букв пишется только в этом примере, тогда как везде в документации пишется как и в CH "UserID" и кажется логичным написать именно так, ведь это документация CH, а не другого языка | 2019-06-29T20:39:08 | 2019-06-29T20:39:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 506,362,334 | 5,761 | alexey-milovidov | You can also check the size of individual columns in `system.columns` table
and the size of each column in every data part in `system.parts_columns` table. | 2019-06-27T14:09:40 | 2019-06-27T14:09:40 | {} | MEMBER |
ClickHouse | ClickHouse | 506,377,265 | 5,761 | ramazanpolat | > You can also check the size of individual columns in `system.columns` table
> and the size of each column in every data part in `system.parts_columns` table.
I did check them and they are different. The problem is I don't know why this happened. | 2019-06-27T14:46:19 | 2019-06-27T14:46:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 506,467,043 | 5,601 | alexey-milovidov | I've read the specs briefly. Avro format looks very sound. In simplest case it will be similar to RowBinary format. Schema parsing and processing may be a bit more complicated.
By the way, I'm not familiar with popular use cases of Avro format. Could you please tell, why you are using Avro; where data comes from; wh... | 2019-06-27T18:49:23 | 2019-06-27T18:49:23 | {} | MEMBER |
ClickHouse | ClickHouse | 531,851,773 | 6,947 | alesapin | With gcc we got huge performance degradation, perf top is always something like:
```
52.11% clickhouse [.] _ZN2DB5Dwarf12LineNumberVM4stepERSt17basic_string_viewIcSt11char_traitsIcEE
36.90% clickhouse ... | 2019-09-16T16:23:04 | 2019-09-16T16:23:55 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.