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 | 563,380,678 | 8,097 | alexey-milovidov | Ok. | 2019-12-09T18:57:45 | 2019-12-09T18:57:45 | {} | MEMBER |
ClickHouse | ClickHouse | 563,947,711 | 8,118 | excitoon | Based on #7810 | 2019-12-10T09:35:03 | 2019-12-10T09:35:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 563,971,363 | 7,386 | akuzm | > It looks like it started to work:
Nope, it's the same boring thing that accounts for the most test failures -- implicit reinterpret_cast between signed and unsigned ints. Most instances of it are really OK so maybe I should just allow it. | 2019-12-10T10:33:21 | 2019-12-10T10:33:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 564,142,097 | 8,120 | den-crane | detach table MVname / attach table MVname.
Will be automatically attached on CH restart | 2019-12-10T17:26:34 | 2019-12-10T17:26:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 563,273,380 | 7,812 | Akazz | > I don't think the failed test relates to my change.
No it does not. I've looked into it - looks like a flapping test. | 2019-12-09T14:50:58 | 2019-12-09T14:50:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 563,486,831 | 7,900 | alexey-milovidov | Is it acceptable to keep metric names in CamelCase without any changes? To keep the code more maintenable. | 2019-12-09T23:25:13 | 2019-12-09T23:25:13 | {} | MEMBER |
ClickHouse | ClickHouse | 563,514,252 | 8,091 | onine007 | @ramazanpolat
SQL:
SELECT A, B, sum(C), count() from rollup GROUP BY ROLLUP(A, B) ORDER BY A, B FORMAT PrettyCompact;
+-A---B---sum(C)---count()-+
1 ¦ ¦ 0 ¦ 710 ¦ 8 ¦
2 ¦ a ¦ 0 ¦ 110 ¦ 4 ¦
3 ¦ a ¦ 1 ¦ 10 ¦ 2 ¦
4 ¦ a ¦ 2 ¦ 100 ¦ 2 ¦
5 ¦ b ¦ 0 ¦ 600 ¦ ... | 2019-12-10T01:13:11 | 2019-12-10T01:13:11 | {} | NONE |
ClickHouse | ClickHouse | 564,070,496 | 8,121 | byx313 | > @byx313
> I tried the first option but the problem is still same :(
> Method2: You mean to say single insert will do than multiple inserts ?
Do 'ulimit -a' again to check whether the operation work.
> Method2: You mean to say single insert will do than multiple inserts ?
Yes.10w message a batch a insert is... | 2019-12-10T14:55:41 | 2019-12-10T14:55:53 | {} | NONE |
ClickHouse | ClickHouse | 564,121,574 | 8,125 | den-crane | ????
This an Altinity -- unofficial build . Open a ticket in their github https://github.com/Altinity/clickhouse-rpm/issues
Clickhouse official rpm https://repo.yandex.ru/clickhouse/rpm/ | 2019-12-10T16:41:59 | 2019-12-10T16:43:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 563,925,100 | 8,095 | Ubusubus | Ok.
I think it's good to fix documentation for such things.
https://github.com/ClickHouse/ClickHouse/pull/8114 | 2019-12-10T08:34:36 | 2019-12-10T08:34:36 | {} | NONE |
ClickHouse | ClickHouse | 564,145,109 | 7,737 | amosbird | ncurses usually tries to find `/etc/termcap` as fallback while termcap won't search for the terminfo database. This is the main reason of this pr. However some system's readline lib has good builtin support (termios based) that works well even when termcap database is missing.
Hopefully one day https://github.com/A... | 2019-12-10T17:33:28 | 2019-12-10T17:34:57 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 564,211,465 | 8,126 | alexey-milovidov | Looks like nothing is wrong with `libressl`.
It looks at socket timeout.
But after we set `POCO_BROKEN_TIMEOUTS`, socket timeout is no longer set:
```
void SocketImpl::setReceiveTimeout(const Poco::Timespan& timeout)
{
#ifndef POCO_BROKEN_TIMEOUTS
#if defined(_WIN32)
int value = (int) timeout.totalMillisec... | 2019-12-10T19:40:42 | 2019-12-10T19:40:42 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 563,158,620 | 8,080 | hombit | @alexey-milovidov looks fine! | 2019-12-09T10:08:36 | 2019-12-09T10:08:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 563,214,087 | 8,074 | blinkov | @mgaidamak the only reason this translation does not exist yet is lack of human resources. Also keep in mind that its a bit outdated (at least "processors" are not mentioned). | 2019-12-09T12:22:33 | 2019-12-09T12:22:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 563,215,586 | 8,094 | alexey-milovidov | Looks like the file contains less number of columns than the table. | 2019-12-09T12:27:21 | 2019-12-09T12:27:21 | {} | MEMBER |
ClickHouse | ClickHouse | 563,262,204 | 6,263 | nvartolomei | @hustnn https://github.com/ClickHouse/ClickHouse/blob/master/docs/ru/extended_roadmap.md#16-полиморфные-куски-данных, doesn't seem to be available in english yet. | 2019-12-09T14:26:08 | 2019-12-09T14:26:08 | {
"+1": 1,
"laugh": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 563,372,353 | 8,097 | alexey-milovidov | It's basically Ok (and 1% better overall: https://clickhouse-test-reports.s3.yandex.net/8097/9a40c97e2938cbc7071ca29add80bfde4f74554e/performance_test/comparison_min_time_gcc_9.html), but need to validate two queries manually:
```
SELECT count() FROM system.numbers WHERE NOT ignore(visitParamExtractRaw(materializ... | 2019-12-09T18:37:13 | 2019-12-09T18:37:13 | {} | MEMBER |
ClickHouse | ClickHouse | 564,012,371 | 7,970 | abyss7 | > Sorry. FYI: Clang is another compiler for C++. You can install **clang 8** and then compile without any problems. We will update our documentation ASAP.
Did we do that? | 2019-12-10T12:33:16 | 2019-12-10T12:33:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 564,170,411 | 8,124 | den-crane | OK.
>I added <input_format_import_nested_json>1</input_format_import_nested_json>
>to the default via users.xml
Please show your users.xml (your changes). | 2019-12-10T18:32:29 | 2019-12-10T18:32:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 563,214,964 | 8,094 | Crazylearner30 | @ramazanpolat : Thanks a lot for the quick response , reordering the file could help? | 2019-12-09T12:25:24 | 2019-12-09T12:25:24 | {} | NONE |
ClickHouse | ClickHouse | 563,300,762 | 8,083 | filimonov | You can try `CREATE new_table ... AS (SELECT * from existring table union all select new_data) ` | 2019-12-09T15:48:24 | 2019-12-09T15:51:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 563,373,788 | 8,097 | amosbird | I've been seeing these two queries failing from time to time (with the exact regression rate 85%). | 2019-12-09T18:41:07 | 2019-12-09T18:41:07 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 563,467,511 | 7,916 | azat | >Tests are bad:
Will take a look | 2019-12-09T22:22:32 | 2019-12-09T22:22:32 | {} | MEMBER |
ClickHouse | ClickHouse | 564,108,519 | 7,737 | wangzhen11aaa | 98 -- Using termcap: /usr/lib64/libtinfo.so
I have tried this modification, the cmake log like above, But It does not prevent my clickhouse-client from changing the up arrow into ^[[A. | 2019-12-10T16:14:33 | 2019-12-10T16:14:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 563,216,560 | 8,074 | alexey-milovidov | It will be very appreciated.
I have written it in english to avoid duplicating the efforts. | 2019-12-09T12:30:09 | 2019-12-09T12:30:09 | {} | MEMBER |
ClickHouse | ClickHouse | 563,216,928 | 8,074 | alexey-milovidov | You can translate it regardless to the fact that it's a bit outdated and we will update it. | 2019-12-09T12:31:12 | 2019-12-09T12:31:12 | {} | MEMBER |
ClickHouse | ClickHouse | 563,297,370 | 6,892 | akuzm | > Can you also look on that while we are here?
> [#7736 (comment)](https://github.com/ClickHouse/ClickHouse/issues/7736#issuecomment-552922270)
Not until I finish this update, and my perspective on this is quite grim, given the Mystery Integration Test Failure I'm seeing in CI but not locally. | 2019-12-09T15:40:41 | 2019-12-09T15:40:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 563,297,429 | 8,095 | filimonov | It is by design.
https://clickhouse.yandex/docs/en/query_language/create/#temporary-tables
TEMPORARY TABLES live only in the session. Distributed DDL queries are executed remotely using another session, so anyway you will not be able to see them.
You can just creare Engine=Memory tables with unique name.
... | 2019-12-09T15:40:48 | 2019-12-09T15:40:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 564,023,400 | 8,062 | 4ertus2 | Тикет действительно был слишком глобальный. Сейчас он про более локальную оптимизацию:
Механизм построения set для блока IN и правой таблицы JOIN один и тот же: выполняется подзапрос, собирающий нужные ключи. Их можно строить одновременно (и так даже делается если запросы идентичны). При построении hash-таблицы попу... | 2019-12-10T13:05:18 | 2019-12-10T13:11:41 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 564,028,905 | 8,093 | kaijianding | maybe you can download the clickhouse rpm using `thunder` and transfer the file to your machine and install rpm directly | 2019-12-10T13:20:47 | 2019-12-10T13:20:47 | {} | NONE |
ClickHouse | ClickHouse | 564,072,474 | 8,121 | byx313 | > @byx313
> yes, I did ulimit -a to check and yes the change is reflected.
may be you should try to change you insert frequency.What's the frequency now? | 2019-12-10T15:00:03 | 2019-12-10T15:00:03 | {} | NONE |
ClickHouse | ClickHouse | 564,098,938 | 8,031 | den-crane | >Could you say if this possible to control this behavior?
No.
>implementation specific.
It means this behavior is different for different functions / parsers.
Nothing can be done. Only a validation at client side. | 2019-12-10T15:54:17 | 2019-12-10T15:55:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 564,114,124 | 8,061 | den-crane | It's expected behavior. A feature not a bug.
Nullable types are slower up to 2-3 times that's why CH does not follow common SQL.
```
select max(a) from (select cast(1, 'Nullable(Int64)') a) where 0;
┌─max(a)─┐
│ ᴺᵁᴸᴸ │
└────────┘
select max(a) from (select 1 a) where 0
┌─max(a)─┐
│ 0 │
└────────... | 2019-12-10T16:26:13 | 2019-12-10T16:28:17 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 564,168,632 | 8,124 | samuelblawrence | Thanks - I don't have any issue with the data into the k1 table - it's in the right format and works correctly. My problem is with moving the data over to the t1 table - if I "insert into... select" everything works fine, so why should it be different when running from a materialized view? | 2019-12-10T18:28:26 | 2019-12-10T18:28:26 | {} | NONE |
ClickHouse | ClickHouse | 563,327,595 | 7,799 | sklyar | I have the same problem:
`DB::Exception: The createColumn() method is not implemented for the Set data type (version 19.17.5.18 (official build)). `
```
SELECT links.ID
, links.Type
, count()
, toDateTime(toStartOfInterval(StartTime, INTERVAL 1 week)) AS StartTime
FROM shows
ALL
... | 2019-12-09T16:47:27 | 2019-12-09T16:48:04 | {} | NONE |
ClickHouse | ClickHouse | 563,348,530 | 7,916 | alexey-milovidov | Tests are bad:
https://clickhouse-test-reports.s3.yandex.net/8086/9399b39070296951f03224696edd3819d354ba98/functional_stateless_tests_(unbundled).html
https://clickhouse-test-reports.s3.yandex.net/8086/9399b39070296951f03224696edd3819d354ba98/functional_stateless_tests_(ubsan).html | 2019-12-09T17:37:26 | 2019-12-09T17:37:26 | {} | MEMBER |
ClickHouse | ClickHouse | 563,366,036 | 7,812 | Akazz | Oh, and please, compose the description as you would like it to appear in the changelog. See any other PR for example | 2019-12-09T18:21:38 | 2019-12-09T18:21:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 563,492,691 | 8,086 | alexey-milovidov | https://github.com/ClickHouse/ClickHouse/issues/5739 | 2019-12-09T23:47:15 | 2019-12-09T23:47:15 | {} | MEMBER |
ClickHouse | ClickHouse | 563,520,754 | 6,263 | hustnn | @nvartolomei Thanks. Let me use google translate first to take a look at the basic idea. I am also facing this issue now. | 2019-12-10T01:22:06 | 2019-12-10T01:22:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 563,885,288 | 8,112 | hczhcz | Related: #1211 #7044 #7315 | 2019-12-10T06:27:22 | 2019-12-10T06:27:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 564,060,351 | 8,021 | 4ertus2 | > If I use a real table (has columns a lot), only method 2 will produce the right result.
Could you make a sample to show it? Your sample returns the same result for each method. | 2019-12-10T14:34:47 | 2019-12-10T14:34:47 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 564,180,856 | 8,124 | samuelblawrence | okay - got it working - a server restart seems to have fixed the issue, I guess the users change wasn't picked up by the system process somehow?
Anyway, appreciate your help - it's a great product! | 2019-12-10T18:58:20 | 2019-12-10T18:58:20 | {} | NONE |
ClickHouse | ClickHouse | 563,293,189 | 8,083 | stavrolia | HDFS store immutable files, and the library we choose to integrate with HDFS allow only create file or overwrite it. If I understand correctly the main scenario of writing to HDFS is to add new file. So there is no easy way to append rows to file in HDFS by table in ClickHouse. | 2019-12-09T15:30:53 | 2019-12-09T15:30:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 563,335,978 | 8,102 | alexey-milovidov | `SELECT * FROM table INTO OUTFILE 'file.tsv'` | 2019-12-09T17:06:32 | 2019-12-09T17:06:32 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 563,508,110 | 6,228 | nvartolomei | On first look this almost the same as #5732, speed is throttled only for queries that read QueryProcessingStage::Complete.
I have a branch somewhere which defines new setting `leaf_max ...`. However I'd propose to change the meaning of current settings and apply them always.
https://github.com/ClickHouse/ClickHou... | 2019-12-10T00:47:51 | 2019-12-10T00:47:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 563,555,499 | 6,263 | hustnn | > It is proposed to allow pieces of tables such as MergeTree to arrange data in different formats. Namely: - in RAM; - on a disk with all columns in one file; - on a disk with columns in separate files.
Does one table have only one format or one table can have different formats (part in memory, part in disk) ? | 2019-12-10T01:48:15 | 2019-12-10T01:48:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 563,885,919 | 7,962 | andyyzh | Clickhouse use the C language version of RoaringBitmap/CRoaring, which use 32-bit integer,
The 64 bit version of CPP is to wrap two 32-bit CRoaring into 64 bit implementation instead of long type. The performance is not good. [roaring64map.hh](https://github.com/RoaringBitmap/CRoaring/blob/master/cpp/roaring64map.hh) | 2019-12-10T06:29:46 | 2019-12-10T06:29:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 563,919,664 | 8,083 | vyrodovalexey | Thanks! | 2019-12-10T08:19:20 | 2019-12-10T08:19:20 | {} | NONE |
ClickHouse | ClickHouse | 564,011,329 | 8,085 | abyss7 | I will take a look. Do you build from scratch? | 2019-12-10T12:30:18 | 2019-12-10T12:30:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 564,014,835 | 7,995 | abyss7 | Yes, it looks like improper condition, since it doesn't take into account the compiler name.
Anyway, we want to restrict build of ClickHouse on Mac only to Clang compiler. | 2019-12-10T12:40:18 | 2019-12-10T12:40:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 564,028,281 | 8,120 | filimonov | If your materialized view was created with TO syntax you can just drop materialized view. | 2019-12-10T13:19:07 | 2019-12-10T13:19:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 564,207,962 | 8,126 | alexey-milovidov | To debug, you can strace this thread and check what happens with `read` syscalls - do they retry, with what error code, does timeout get adjusted, etc... | 2019-12-10T19:32:57 | 2019-12-10T19:33:23 | {} | MEMBER |
ClickHouse | ClickHouse | 563,213,826 | 8,094 | ramazanpolat | You CSV file is not in proper format. Luckily it is the first row. You can insert first row manually to see what the problem is. | 2019-12-09T12:21:54 | 2019-12-09T12:21:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 563,355,066 | 7,810 | excitoon | Safe to merge from my side. | 2019-12-09T17:54:30 | 2019-12-09T17:54:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 563,935,706 | 8,117 | hczhcz | For reference: https://medium.com/@sundarstyles89/weight-of-evidence-and-information-value-using-python-6f05072e83eb | 2019-12-10T09:04:42 | 2019-12-10T09:04:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 563,966,803 | 8,108 | alesapin | real trace
```
#0 detail::endsWith (s=..., suffix=0x4466341 ".gz", suffix_size=3) at /home/alesap/code/cpp/ClickHouse/dbms/src/Common/StringUtils/StringUtils.cpp:14
#1 0x000000000bf55ea0 in endsWith (s=..., suffix=0x4466341 ".gz") at /home/alesap/code/cpp/ClickHouse/dbms/src/Common/StringUtils/StringUtils.h:36
#2... | 2019-12-10T10:22:01 | 2019-12-10T10:22:01 | {} | MEMBER |
ClickHouse | ClickHouse | 564,032,949 | 8,096 | filimonov | Reproduces on bit simpler case:
```sql
create database y2018;
create database tmp;
create table y2018.vef_shard(a String,b String) ENGINE=MergeTree()
PARTITION BY a ORDER BY b;
create table tmp.A (a String) ENGINE=MergeTree()
PARTITION BY a ORDER BY a;
create table tmp.B (b String) ENGINE=MergeTree()
PAR... | 2019-12-10T13:31:19 | 2019-12-10T13:31:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 563,281,276 | 7,810 | excitoon | I need to add a test here. | 2019-12-09T15:07:14 | 2019-12-09T15:07:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 563,899,633 | 8,027 | jiugem | @alexey-milovidov | 2019-12-10T07:16:07 | 2019-12-10T07:16:07 | {} | NONE |
ClickHouse | ClickHouse | 563,929,207 | 7,900 | vdimir | > Is it acceptable to keep metric names in CamelCase without any changes? To keep the code more maintenable.
I have not enough experience with Prometheus to decide which solution more acceptable. But now I reverted names to CamelCase and ready to implement other approach for naming if we'll find it.
Other uncompl... | 2019-12-10T08:46:22 | 2019-12-10T08:46:22 | {} | MEMBER |
ClickHouse | ClickHouse | 563,956,096 | 8,098 | alesapin | A simple fix of obvious bugs. All tests passed, so I'll merge it. | 2019-12-10T09:55:25 | 2019-12-10T09:55:25 | {} | MEMBER |
ClickHouse | ClickHouse | 564,094,228 | 8,067 | den-crane | >works by throwing error
Yes it's by design. max_memory_usage -- it is a limiter. It just throws an error to a client. It does NOT reduce memory usage.
>I would like to find a way of make the query run with
>limited RAM (probably at cost of longer runtime,
>but it would be OK). Would this be possible?
No, ... | 2019-12-10T15:44:13 | 2019-12-10T15:48:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 563,217,565 | 8,094 | Crazylearner30 | @alexey-milovidov : Thanks for the response, initially I thought same and manually checked whether the file has fewer columns but, it has the exactly the same number of columns as in the table.
Keeping @ramazanpolat's comment above in mind I tried to reorder the file and it loaded all the records successfully.
Am ... | 2019-12-09T12:32:59 | 2019-12-09T12:32:59 | {} | NONE |
ClickHouse | ClickHouse | 563,484,499 | 7,386 | alexey-milovidov | It looks like it started to work:
```
9.12.09 21:45:23.703176 [ 54 ] {} <Fatal> BaseDaemon: 8. 0x119b5d67 unsigned long& DB::Field::get<unsigned long>() /build/obj-x86_64-linux-gnu/../dbms/src/Core/Field.h:0
2019.12.09 21:45:23.715984 [ 54 ] {} <Fatal> BaseDaemon: 9. 0x1483d3de DB::FlatDictionary::blockToAttributes(... | 2019-12-09T23:17:09 | 2019-12-09T23:17:09 | {} | MEMBER |
ClickHouse | ClickHouse | 563,993,072 | 8,108 | alesapin | I guess problem in this line https://github.com/clickhouse/ClickHouse/blob/master/dbms/src/Storages/StorageFile.cpp#L153, because we return empty vector for non existing path. | 2019-12-10T11:33:47 | 2019-12-10T11:33:47 | {} | MEMBER |
ClickHouse | ClickHouse | 564,034,078 | 8,094 | filimonov | > @alexey-milovidov file had no header and i had specified CSV
In that case you need to list all the needed columns in your insert statement in the proper order.
I.e.
```
INSERT INTO table(col3,col1,col2) FORMAT CSV
``` | 2019-12-10T13:34:13 | 2019-12-10T13:34:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 564,056,161 | 8,121 | byx313 | > Am getting the below error while loading data and only half of the data is being loaded into the table
>
> DB::Exception: Cannot open file /t-3tb-data/clickhouse/data/database/table/tmp_insert_0c87b3bf0c31a7766299a14d202c8da9_648_648_0/TI_verification_status.mrk, errno: 24, strerror: Too many open files.
>
> Ca... | 2019-12-10T14:25:32 | 2019-12-10T14:25:32 | {
"hooray": 1
} | NONE |
ClickHouse | ClickHouse | 564,130,987 | 8,124 | den-crane | You've used not a `JSONEachRow`.
That what CH `JSONEachRow` expects
```
{"bet_id":"1","parts.category_name":["a","b"]}
```
```
CREATE TABLE k1 (
bet_id UInt64, parts Nested (category_name String)) Engine = Memory;
insert into k1 values (1,['a','b']);
select * from k1 format JSONEachRow
SELECT ... | 2019-12-10T17:02:10 | 2019-12-10T17:15:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 564,212,619 | 8,126 | alexey-milovidov | In `libressl`, the timeout is correctly handled: `bss_dgram.c`, `dgram_adjust_rcv_timeout` | 2019-12-10T19:43:19 | 2019-12-10T19:43:19 | {} | MEMBER |
ClickHouse | ClickHouse | 563,241,464 | 6,898 | ilyakarol | Restart helped | 2019-12-09T13:36:47 | 2019-12-09T13:36:47 | {} | NONE |
ClickHouse | ClickHouse | 563,369,805 | 8,097 | amosbird | Perf tests run wild again. | 2019-12-09T18:30:28 | 2019-12-09T18:30:28 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 563,500,403 | 7,916 | alexey-milovidov | I think, it's better to make `SYSTEM RELOAD DICTIONARY` and `DICTIONARIES` always synchronous.
> As a hot-fix, sleep(1)
It's ok for hotfix. | 2019-12-10T00:17:02 | 2019-12-10T00:17:02 | {} | MEMBER |
ClickHouse | ClickHouse | 564,065,836 | 8,061 | ghuname | How can selecting from empty table return something that is not null?
Do you consider it as a bug or not? As I can see you removed bug label few days ago.
So, it is expected behavior or...? | 2019-12-10T14:46:32 | 2019-12-10T14:46:32 | {} | NONE |
ClickHouse | ClickHouse | 564,071,645 | 8,121 | Crazylearner30 | @byx313
yes, I did ulimit -a to check and yes the change is reflected. | 2019-12-10T14:58:12 | 2019-12-10T14:58:12 | {} | NONE |
ClickHouse | ClickHouse | 563,215,028 | 8,091 | ramazanpolat | There is no function named as `GROUPING`. What do you want to achieve here? | 2019-12-09T12:25:38 | 2019-12-09T12:25:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 563,245,618 | 8,094 | alexey-milovidov | @Crazylearner30 No, the order of file does not matter. | 2019-12-09T13:46:58 | 2019-12-09T13:46:58 | {} | MEMBER |
ClickHouse | ClickHouse | 563,385,789 | 8,035 | KochetovNicolai | Part about processors seem to be ok. Haven't read all the changes in formats and kafka yet. | 2019-12-09T19:09:21 | 2019-12-09T19:09:21 | {} | MEMBER |
ClickHouse | ClickHouse | 563,400,089 | 8,086 | alexey-milovidov | Continued in https://github.com/ClickHouse/ClickHouse/pull/8105 | 2019-12-09T19:38:40 | 2019-12-09T19:38:40 | {} | MEMBER |
ClickHouse | ClickHouse | 563,725,443 | 8,091 | den-crane | try Nullable
cast (A, 'Nullable(Int64)') , cast(B,'Nullable(String)'),.... | 2019-12-10T03:57:11 | 2019-12-10T03:57:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 564,042,101 | 8,085 | tisonkun | @abyss7 I'm not sure what "from scratch" means. The output above is the second try I run `ninja` so you can see it started with `[1/1576]` because 2000+ previously done.
Apart from that I checkout source code & run as the document said, i.e., the commands above. | 2019-12-10T13:53:12 | 2019-12-10T13:53:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 563,248,791 | 8,094 | Crazylearner30 | @alexey-milovidov file had no header and i had specified CSV
| 2019-12-09T13:54:58 | 2019-12-09T13:54:58 | {} | NONE |
ClickHouse | ClickHouse | 563,293,900 | 6,892 | filimonov | Can you also look on that while we are here?
https://github.com/ClickHouse/ClickHouse/issues/7736#issuecomment-552922270 | 2019-12-09T15:32:32 | 2019-12-09T15:32:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 564,127,063 | 8,122 | den-crane | First of all -- GRANULARITY 8192 is a nonsense. Your index granula will contain 8192*8192 rows.
Try GRANULARITY 2.
Second. Before optimize 666 is stored in a separate part [insert into test values(45645645, '666');]. Only this Skip index's granula contains this value 666. Other parts don't. After optimize this 666 ... | 2019-12-10T16:53:40 | 2019-12-10T16:53:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 564,212,166 | 8,126 | alexey-milovidov | We can check the values of SO_RCVTIMEO:
https://stackoverflow.com/questions/1824115/display-socket-options | 2019-12-10T19:42:15 | 2019-12-10T19:42:15 | {} | MEMBER |
ClickHouse | ClickHouse | 563,245,813 | 8,094 | alexey-milovidov | Probably you have CSV with header but you specified CSV instead of CSVWithNames? | 2019-12-09T13:47:29 | 2019-12-09T13:47:29 | {} | MEMBER |
ClickHouse | ClickHouse | 563,966,031 | 8,108 | alesapin | In my env situation is even worse:
```
2019.12.10 13:19:19.568145 [ 121 ] {65c15ee9-1b44-4d58-8cb2-aa4ebbebb618} <Debug> executeQuery: (from [::1]:56932) insert into table function file('abc.csv', 'CSV', 'val1 UInt32') values
2019.12.10 13:19:19.575464 [ 123 ] {} <Fatal> BaseDaemon: #################################... | 2019-12-10T10:20:06 | 2019-12-10T10:20:06 | {} | MEMBER |
ClickHouse | ClickHouse | 563,988,888 | 8,100 | 4ertus2 | Duplicate #7556 | 2019-12-10T11:21:26 | 2019-12-10T11:21:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 564,014,560 | 7,970 | nikitamikhaylov | > > Sorry. FYI: Clang is another compiler for C++. You can install **clang 8** and then compile without any problems. We will update our documentation ASAP.
>
> Did we do that?
No :( | 2019-12-10T12:39:34 | 2019-12-10T12:39:34 | {} | MEMBER |
ClickHouse | ClickHouse | 564,051,550 | 8,104 | alesapin | Performance tests failure not related. | 2019-12-10T14:15:03 | 2019-12-10T14:15:03 | {} | MEMBER |
ClickHouse | ClickHouse | 564,055,574 | 7,915 | 4ertus2 | #7556 | 2019-12-10T14:24:15 | 2019-12-10T14:24:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 564,068,302 | 8,121 | Crazylearner30 | @byx313
I tried the first option but the problem is still same :(
Method2: You mean to say single insert will do than multiple inserts ? | 2019-12-10T14:51:36 | 2019-12-10T14:51:36 | {} | NONE |
ClickHouse | ClickHouse | 564,148,264 | 8,126 | alesapin | Actually we have this in our roadmap https://clickhouse.yandex/docs/ru/extended_roadmap/#7-15-zamena-libressl-obratno-na-openssl. | 2019-12-10T17:40:57 | 2019-12-10T17:40:57 | {} | MEMBER |
ClickHouse | ClickHouse | 563,241,054 | 6,898 | ilyakarol | The issue has been appeared again.
The worst part here is that it's impossible to perform any distributed query on the broken machine. | 2019-12-09T13:35:48 | 2019-12-09T13:39:13 | {} | NONE |
ClickHouse | ClickHouse | 563,317,093 | 7,810 | excitoon | I would like to add a test for delayed merge as well. | 2019-12-09T16:24:17 | 2019-12-09T16:24:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 563,475,612 | 7,916 | azat | The problem is that `SYSTEM RELOAD DICTIONARY` is async
```
2019.12.09 05:07:21.862801 [ 36 ] {61df55ee-4a17-4bf3-8cfb-9478490393c3} <Debug> executeQuery: (from [::1]:37422) SYSTEM RELOAD DICTIONARY `foo 1234.dict`
2019.12.09 05:07:21.863289 [ 36 ] {61df55ee-4a17-4bf3-8cfb-9478490393c3} <Debug> MemoryTracker: Peak... | 2019-12-09T22:47:52 | 2019-12-09T22:47:52 | {} | MEMBER |
ClickHouse | ClickHouse | 563,492,702 | 8,105 | alexey-milovidov | https://github.com/ClickHouse/ClickHouse/issues/5739 | 2019-12-09T23:47:18 | 2019-12-09T23:47:18 | {} | MEMBER |
ClickHouse | ClickHouse | 564,069,479 | 8,096 | maqroll | Thanks!!
Just a comment: it works if you take out any of the inner joins.
Either
```
select * from y2018.vef_shard t
inner join tmp.B on t.b = B.b
WHERE t.a in (select a from tmp.A);
```
or
```
select * from y2018.vef_shard t
inner join tmp.A on t.a = A.a
WHERE t.a in (select a from tmp.A);
```
wo... | 2019-12-10T14:53:51 | 2019-12-10T15:05:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 564,075,662 | 8,121 | Crazylearner30 | @byx313
am loading one file after the other , once the first file is loaded starting with the next one. | 2019-12-10T15:06:40 | 2019-12-10T15:19:31 | {} | NONE |
ClickHouse | ClickHouse | 564,143,671 | 7,737 | amosbird | That's because either:
1. your terminfo database is broken
2. your libtinfo.so is built with a non-canonical tc search path
3. your $TERM is non-standard
4. your libreadline doesn't have builtin fallback support when `tgetent` fails | 2019-12-10T17:30:17 | 2019-12-10T17:30:17 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 565,140,521 | 8,169 | 4ertus2 | You may add an alias to subquery and access via it. Doesn't it help?
```
select s1.c1
from (
select t1.c1
from t1
any left join t2 on t1.c1 = t2.c1
) s1;
``` | 2019-12-12T19:04:26 | 2019-12-12T19:04:36 | {} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.