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 | 734,814,026 | 17,485 | alexey-milovidov | Ok, but I'd like to make it automatically determined. | 2020-11-27T12:33:08 | 2020-11-27T12:33:08 | {} | MEMBER |
ClickHouse | ClickHouse | 734,845,797 | 17,461 | den-crane | Because expr. aliases are global in CH, you can use them in any section, where, order by , group by
in your case you make `where true` : ` WHERE '2020-11-01' = '2020-11-01' `
https://clickhouse.tech/docs/en/sql-reference/syntax/#syntax-expression_aliases
WA `events_all.ldate = '2020-11-01'` :
```sql
SELEC... | 2020-11-27T13:53:23 | 2020-11-27T13:54:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 734,881,306 | 16,806 | TCeason | It’s ok now.@tavplubix | 2020-11-27T15:16:34 | 2020-11-27T15:16:34 | {
"+1": 1,
"hooray": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 735,373,365 | 17,453 | alexey-milovidov | Could you please file an issue to RocksDB?
We can also patch it for ourself... | 2020-11-29T10:42:13 | 2020-11-29T10:42:13 | {} | MEMBER |
ClickHouse | ClickHouse | 735,376,244 | 17,453 | azat | >Looks like race in rocksdb, between UnlockFile and LockFile accessing locked_files.
>Yes the mutex should be released later in the code.
One note for the stacktrace above, `DBImpl::CloseHelper` and `DBImpl::Open` these should be different rocksdb instances
>Could you please file an issue to RocksDB?
PR is al... | 2020-11-29T11:09:19 | 2020-11-29T11:09:19 | {} | MEMBER |
ClickHouse | ClickHouse | 734,794,404 | 5,149 | alexey-milovidov | @Amesaru | 2020-11-27T11:41:40 | 2020-11-27T11:41:40 | {} | MEMBER |
ClickHouse | ClickHouse | 734,846,647 | 9,959 | grundrauschen | It would also be great to add a modifier to formatDateTime to print subsecond values when used in conjunction with DateTime64. Together with the timezone offset from above I would then be able to get a full ISO-8601 timestamp from DateTime64 like `2020-11-27T14:54:10.123456Z`. | 2020-11-27T13:55:23 | 2020-11-27T13:55:23 | {} | NONE |
ClickHouse | ClickHouse | 734,969,011 | 17,347 | azat | >Please review (and continue?) this PR #16277
Since the author hadn't responding, I decided to just continue it, see #17459 (all changes was done on top of that PR) | 2020-11-27T20:00:39 | 2020-11-27T20:00:39 | {} | MEMBER |
ClickHouse | ClickHouse | 735,001,089 | 16,457 | den-crane | I guess the problem is in `uniq(arrayJoin( user) )` and it can be solved without arrayJoin. | 2020-11-27T22:16:18 | 2020-11-27T22:16:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 735,036,621 | 17,376 | alexey-milovidov | Yes, it should be checked on type, not by column. | 2020-11-28T04:02:56 | 2020-11-28T04:02:56 | {} | MEMBER |
ClickHouse | ClickHouse | 735,051,908 | 17,496 | alesapin | :( Timeouts in integration tests | 2020-11-28T07:42:52 | 2020-11-28T07:42:52 | {} | MEMBER |
ClickHouse | ClickHouse | 735,220,662 | 17,113 | joshxyzhimself | Hi, is this possible in the HTTP Interface?
https://clickhouse.tech/docs/en/interfaces/http/
For query
```
SET date_time_output_format = 'iso';
```
I've been getting
```
Error: Code: 113, e.displayText() = DB::Exception: There is no session (version 20.11.4.13 (official build))
```
---
I guess ... | 2020-11-28T11:41:08 | 2020-11-28T12:02:37 | {} | NONE |
ClickHouse | ClickHouse | 735,366,953 | 17,463 | alexey-milovidov | You can download the binary and run on CentOS 5 in Docker. | 2020-11-29T09:36:50 | 2020-11-29T09:36:50 | {} | MEMBER |
ClickHouse | ClickHouse | 735,373,063 | 17,453 | azat | >Integration tests (asan) — fail: 0, passed: 961, error: 22
>Integration tests (thread) — fail: 0, passed: 955, error: 22
grpc (hm, seems that the checks just restarted, but will they pass eventually?)
>Functional stateless tests (debug) — fail: 1, passed: 2469, skipped: 16
01317_no_password_in_command_line (... | 2020-11-29T10:39:22 | 2020-11-29T10:39:22 | {} | MEMBER |
ClickHouse | ClickHouse | 735,378,751 | 17,464 | alexey-milovidov | Anyway, you can exclude it from the fast test. | 2020-11-29T11:30:38 | 2020-11-29T11:30:38 | {} | MEMBER |
ClickHouse | ClickHouse | 735,397,723 | 14,448 | alexey-milovidov | Cannot reproduce on master without sample data:
```
ClickHouse client version 20.13.1.1.
Connecting to localhost:9000 as user default.
Connected to ClickHouse server version 20.13.1 revision 54442.
milovidov-desktop :) SELECT uniqCombinedMerge(cast(unhex(state) as AggregateFunction(uniqCombined, String))) as r... | 2020-11-29T13:49:15 | 2020-11-29T13:49:15 | {} | MEMBER |
ClickHouse | ClickHouse | 735,420,673 | 11,714 | alexey-milovidov | Just check that metrics exist and show non-zero value if there are MergeTree tables with some data. | 2020-11-29T16:33:24 | 2020-11-29T16:33:24 | {} | MEMBER |
ClickHouse | ClickHouse | 734,929,463 | 17,485 | alexey-milovidov | > Doing this automatically would be akin to just automatically ignoring all failing functional tests -- convenient in some sense, but defeats their purpose.
It is Ok to have random fluctuations in query time for different builds for small share of queries if there is no consistent decrease of performance. | 2020-11-27T17:23:59 | 2020-11-27T17:23:59 | {} | MEMBER |
ClickHouse | ClickHouse | 734,997,536 | 17,376 | den-crane | An error message is slightly confusing for newbies
SELECT
toUnixTimestamp(toDate('2020-01-01'), 'Europe/Moscow')
DB::Exception: Illegal column UInt16 of first argument of function toUnixTimestamp:
While processing toUnixTimestamp(toDate('2020-01-01')
UInt16, Not Date. | 2020-11-27T21:53:06 | 2020-11-27T21:53:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 735,015,322 | 17,462 | chenziliang | > What do you think about separating it to several settings:
> `asterisk_include_materialized_columns`
> `asterisk_include_alias_columns`
> for more fine grained and explicit control?
Make total sense and revised the patch accordingly. Thanks Alexey. | 2020-11-28T00:09:23 | 2020-11-28T00:09:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 735,033,902 | 17,055 | l1t1 | maybe the root of slow VM is low diskspace
```
SELECT formatReadableSize(filesystemAvailable()) AS `Available space`
Query id: a30b9e6a-bd81-4496-ac34-12ee82d999b3
┌─Available space─┐
│ 2.25 GiB │
└─────────────────┘
``` | 2020-11-28T03:22:05 | 2020-11-28T03:22:05 | {} | NONE |
ClickHouse | ClickHouse | 735,051,994 | 17,055 | filimonov | If you want to see the difference on multiple cpu system you should use `numbers_mt` instead of `numbers` in your test queries.
It seems like the single core on your win7 is few times more performant comparing to single core on your AMD VM. | 2020-11-28T07:44:09 | 2020-11-28T07:44:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 735,064,589 | 17,519 | alexey-milovidov | Old client does not support INDEX. | 2020-11-28T07:59:34 | 2020-11-28T07:59:34 | {} | MEMBER |
ClickHouse | ClickHouse | 735,208,066 | 17,463 | amosbird | How can this PR break centos:5 compatibility... | 2020-11-28T09:46:54 | 2020-11-28T09:46:54 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 735,373,862 | 17,525 | alexey-milovidov | Performance test Ok.
The test `parse_engine_file` maybe slowed down 6%. | 2020-11-29T10:47:09 | 2020-11-29T10:47:09 | {} | MEMBER |
ClickHouse | ClickHouse | 734,854,667 | 17,465 | sanjaypocket52 | Hi Den
when i removed from order (email,username) its works fine.
`CREATE TABLE users_user
(
`id` UInt64,
`created_at` DateTime,
`updated_at` DateTime,
`username` FixedString(100),
`password` Nullable(FixedString(1000)),
`state` Nullable(FixedString(100)),
`mobile_no` FixedStri... | 2020-11-27T14:14:11 | 2020-11-27T14:14:11 | {} | NONE |
ClickHouse | ClickHouse | 735,027,172 | 17,498 | alexey-milovidov | Looks like a kludge. | 2020-11-28T01:59:34 | 2020-11-28T01:59:34 | {} | MEMBER |
ClickHouse | ClickHouse | 735,043,830 | 15,111 | alexey-milovidov | Wow, it works!
Does not look polished though:
```
$ ./clickhouse-grpc-client.py
Connecting to localhost:9100 as user default.
Connected to ClickHouse server version 20.13.1.1 via gRPC protocol in 0.0042569637298583984.
grpc :) SELECT 1
┌─1─┐
│ 1 │
└───┘
Elapsed: 0.002447843551635742 sec.
grpc :) SELE... | 2020-11-28T05:50:19 | 2020-11-28T05:51:42 | {
"hooray": 1
} | MEMBER |
ClickHouse | ClickHouse | 735,398,485 | 16,732 | alexey-milovidov | @Fallenyasha Лимиты берутся из ограничения на запрос, вставляющий данные.
Аллокация данных для таблицы типа Join происходит в контексте запроса на вставку.
Аллокация одной из структур данных требуется не на каждый запрос, а только если места в буфере таблицы недостаточно - в этом случае выделяется в два раза больше... | 2020-11-29T13:54:08 | 2020-11-29T13:54:08 | {} | MEMBER |
ClickHouse | ClickHouse | 734,834,286 | 17,379 | traceon | AppleClang works fine for me. My version is the latest, but 10.0.1 should work too.
@hoverzheng try installing Xcode version of CommandLineTools:
```sh
sudo rm -rf /Library/Developer/CommandLineTools
sudo xcode-select --install
```
Clean everything in build folder and run fresh `cmake ..`. | 2020-11-27T13:26:15 | 2020-11-27T13:26:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 735,000,658 | 16,457 | den-crane | Third time I am looking at this issue and I don't want to retype pictures. Why did you post pictures? | 2020-11-27T22:13:24 | 2020-11-27T22:14:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 735,042,824 | 17,464 | alexey-milovidov | AFAIK Distributed DDLs only work when every server has its own hostname. | 2020-11-28T05:36:41 | 2020-11-28T05:36:47 | {} | MEMBER |
ClickHouse | ClickHouse | 735,126,415 | 17,055 | alexey-milovidov | Cross join can be constrained by memory performance (all that it is doing is copying of memory). | 2020-11-28T08:42:31 | 2020-11-28T08:42:31 | {} | MEMBER |
ClickHouse | ClickHouse | 735,369,663 | 17,463 | amosbird | > You can download the binary and run on CentOS 5 in Docker.
I grabbed the `performance.tgz` from https://clickhouse-builds.s3.yandex.net/17463/c77a4d8379e0b75289607b151514878cf4ca40aa/clickhouse_build_check/report.html and it works fine in centos:5 docker | 2020-11-29T10:05:37 | 2020-11-29T10:05:37 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 735,389,108 | 10,971 | alexey-milovidov | It may happen if you have built ClickHouse without optimizations. | 2020-11-29T12:51:40 | 2020-11-29T12:51:40 | {} | MEMBER |
ClickHouse | ClickHouse | 734,818,561 | 17,465 | sanjaypocket52 | Please find example below on my data. After optimisation merging is not happening.
OPTIMIZE TABLE users_user FINAL DEDUPLICATE
OPTIMIZE TABLE users_user FINAL DEDUPLICATE
Ok.
0 rows in set. Elapsed: 0.129 sec.
select id,email from users_user where id = 101346
SELECT
id,
email
FROM users_u... | 2020-11-27T12:45:22 | 2020-11-27T12:45:22 | {} | NONE |
ClickHouse | ClickHouse | 734,848,643 | 17,465 | den-crane | Please show the output `select id,email, _part from users_user where id = 101346`
and output of `show create table users_user`
and output of
```
set optimize_throw_if_noop=1;
OPTIMIZE TABLE users_user FINAL
```
https://clickhouse.tech/docs/en/engines/table-engines/mergetree-family/replacingmergetree/
... | 2020-11-27T13:59:58 | 2020-11-27T13:59:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 735,042,322 | 17,506 | alexey-milovidov | @akuzm You cannot rely on that something will be done in one second in CI infrastructure. | 2020-11-28T05:29:09 | 2020-11-28T05:29:09 | {} | MEMBER |
ClickHouse | ClickHouse | 735,042,784 | 17,463 | alexey-milovidov | Unfortunately, UNUSED does not work for multiple args. | 2020-11-28T05:36:03 | 2020-11-28T05:36:03 | {} | MEMBER |
ClickHouse | ClickHouse | 735,046,370 | 17,509 | alexey-milovidov | https://github.community/t/github-on-the-ipv6-internet/2794/6
https://github.community/t/cannot-reach-any-github-io-page-via-ipv6/10310
https://stackoverflow.com/questions/62754127/checking-out-from-github-over-ipv6
https://www.reddit.com/r/ipv6/comments/ec8i7y/github_still_doesnt_support_ipv6/ | 2020-11-28T06:25:59 | 2020-11-28T06:25:59 | {} | MEMBER |
ClickHouse | ClickHouse | 735,373,444 | 17,453 | alexey-milovidov | > test_materialize_ttl_in_partition, order of disks
If you know how to fix this test, it will be very appreciated... | 2020-11-29T10:42:54 | 2020-11-29T10:42:54 | {} | MEMBER |
ClickHouse | ClickHouse | 735,388,729 | 8,939 | CurtizJ | Maybe #16988 will fix. But need to check, anyway. | 2020-11-29T12:49:16 | 2020-11-29T12:49:16 | {} | MEMBER |
ClickHouse | ClickHouse | 734,814,332 | 10,708 | traceon | @alexey-milovidov looking... | 2020-11-27T12:34:02 | 2020-11-27T12:34:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 734,844,251 | 17,497 | den-crane | You should use "with csv force quote *;"
```
COPY (select * from crash_test) TO STDOUT with csv force quote *;
COPY (select * from crash_test) TO PROGRAM 'clickhouse-client --query="INSERT INTO crash_test FORMAT CSV"' with csv force quote *;
``` | 2020-11-27T13:49:49 | 2020-11-27T13:49:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 734,859,900 | 17,497 | XavierQuiniou | I'm using CLICKHOUSE4J (cc.blynk.clickhouse)
It's based upon HTTP client.
I proceed like that :
<pre>
String sql = "INSERT INTO <i>my-base.my-table</i> FORMAT CSV settings format_csv_delimiter='|'";
copyManager.copyToDb(sql, <i>my-file</i>);
</pre>
How to place with <tt>csv force quote *;</tt> ? As a setting ?... | 2020-11-27T14:26:11 | 2020-11-27T14:26:11 | {} | NONE |
ClickHouse | ClickHouse | 735,364,924 | 17,550 | l1t1 | create the new table with the row_number
```
VM_0_52_centos :) create table sfz_rn engine=MergeTree order by rn as select * from (select groupArray(code) v,arrayEnumerate(v)rn from sfz group by 1)t array join v,rn;
CREATE TABLE sfz_rn
ENGINE = MergeTree
ORDER BY rn AS
SELECT *
FROM
(
SELECT
grou... | 2020-11-29T09:15:57 | 2020-11-29T09:33:54 | {} | NONE |
ClickHouse | ClickHouse | 735,388,089 | 6,034 | alexey-milovidov | The `distributed_aggregation_memory_efficient` can speed up this class of queries. | 2020-11-29T12:43:55 | 2020-11-29T12:43:55 | {} | MEMBER |
ClickHouse | ClickHouse | 735,422,377 | 17,215 | traceon | What is the content of the original `dmn` and `rpm` tables? How do they change?
| 2020-11-29T16:45:22 | 2020-11-29T16:45:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 735,426,195 | 2,045 | alexey-milovidov | Implemented in #11844. | 2020-11-29T17:14:40 | 2020-11-29T17:14:40 | {} | MEMBER |
ClickHouse | ClickHouse | 734,895,466 | 17,497 | den-crane | @XavierQuiniou it's about PG
```
psql: COPY (select * from crash_test) TO STDOUT with csv force quote *;
```
| 2020-11-27T15:54:52 | 2020-11-27T15:54:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 734,899,655 | 17,465 | den-crane | https://clickhouse.tech/docs/en/engines/table-engines/mergetree-family/replacingmergetree/
_The engine differs from MergeTree in that it removes *duplicate* entries with the same *sorting key* value._
ORDER BY table section defines the KEY which used to IDENTIFY the uniq rows -- sort of PRIMARY KEY.
PRIMARY ... | 2020-11-27T16:06:24 | 2020-11-27T16:08:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 735,302,564 | 17,518 | chenziliang |
> Looks like this should be fixed by #16995
Yes, look like so. @sundy-li , may you help confirm this ? | 2020-11-28T23:09:17 | 2020-11-28T23:09:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 735,359,719 | 17,278 | alexey-milovidov | Yes, Sentry is opt-in.
I also first considered using crypto-hash of FQDN, but it can be brute-forced if reverse DNS is available. | 2020-11-29T08:18:29 | 2020-11-29T08:18:29 | {} | MEMBER |
ClickHouse | ClickHouse | 734,835,296 | 17,379 | traceon | > `$ cmake .. -DCMAKE_CXX_COMPILER=which clang++ -DCMAKE_C_COMPILER=which clang`
Meanwhile, the correct way of substituting the compiler in CMake is this in general:
```sh
$ CC=$(which clang) CXX=$(which clang++) cmake ..
```
...and for brew variants:
```sh
$ CC=$(brew --prefix llvm)/bin/clang CXX=$(brew --p... | 2020-11-27T13:28:35 | 2020-11-27T13:31:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 734,906,441 | 17,497 | den-crane | I'll try to explain. **The problem is not related to CH!!!!!**
CSV format defines that **ROWS**!!! are separated by new_line character. Multiline STRING VALUES MUST BE surrounded by " (double quotes) otherwise it's impossible to distinguish rows.
Here is an example:
```
very
long
row
```
VS
```
"very
lo... | 2020-11-27T16:22:55 | 2020-11-27T16:25:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 735,389,543 | 13,527 | alexey-milovidov | @cmdares ClickHouse does not have builtin feature to calculate cosine distance.
It depends on how do you perform it in your query. | 2020-11-29T12:55:09 | 2020-11-29T12:55:09 | {} | MEMBER |
ClickHouse | ClickHouse | 735,397,857 | 14,448 | alexey-milovidov | @zbitname Could you please send me the data in Telegram `@milovidov_an`? | 2020-11-29T13:49:57 | 2020-11-29T13:49:57 | {} | MEMBER |
ClickHouse | ClickHouse | 734,783,079 | 10,708 | alexey-milovidov | @traceon It has failed: https://github.com/ClickHouse/ClickHouse/issues/17379 | 2020-11-27T11:12:55 | 2020-11-27T11:12:55 | {} | MEMBER |
ClickHouse | ClickHouse | 733,569,659 | 17,379 | hoverzheng | Thanks @filimonov.
The above error come from here:
I build CK like https://clickhouse.tech/docs/en/development/build-osx/：
```
$ cmake .. -DCMAKE_CXX_COMPILER=`which clang++` -DCMAKE_C_COMPILER=`which clang`
CMake Error at cmake/tools.cmake:21 (message):
AppleClang is not supported, you should install clang... | 2020-11-25T09:06:58 | 2020-11-27T18:50:07 | {} | NONE |
ClickHouse | ClickHouse | 735,223,989 | 17,463 | alexey-milovidov | Maybe it's due to Dockerhub rate limiting? (but I don't see it in logs). Let's make another commit. | 2020-11-28T12:14:48 | 2020-11-28T12:14:48 | {} | MEMBER |
ClickHouse | ClickHouse | 735,357,950 | 9,609 | alexey-milovidov | Looks like it's fixed in https://github.com/ClickHouse/ClickHouse/pull/11690. | 2020-11-29T07:59:00 | 2020-11-29T07:59:00 | {} | MEMBER |
ClickHouse | ClickHouse | 735,367,615 | 17,464 | amosbird | > AFAIK Distributed DDLs only work when every server has its own hostname.
But it works when I run fasttest docker image locally. It also works when running the test with hosted clickhouse. | 2020-11-29T09:43:32 | 2020-11-29T09:43:32 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 735,375,494 | 17,542 | alexey-milovidov | > 01415_sticking_mutations
@alesapin | 2020-11-29T11:02:19 | 2020-11-29T11:02:19 | {} | MEMBER |
ClickHouse | ClickHouse | 734,794,330 | 7,203 | alexey-milovidov | @Amesaru | 2020-11-27T11:41:26 | 2020-11-27T11:41:26 | {} | MEMBER |
ClickHouse | ClickHouse | 734,818,513 | 10,708 | traceon | @alexey-milovidov I see you added this:
https://github.com/ClickHouse/ClickHouse/blob/1feebe9d73e76c52eca2ca70332ee7ad547778a0/cmake/tools.cmake#L18-L20
The recent 2 versions worked just fine. Why not allow `AppleClang`? | 2020-11-27T12:45:14 | 2020-11-27T12:45:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 734,874,945 | 17,464 | amosbird | @alesapin ... why would these simple `on cluster` statements time out? | 2020-11-27T15:00:26 | 2020-11-27T15:00:26 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 735,040,042 | 17,465 | sanjaypocket52 | Thanks Dane | 2020-11-28T04:56:29 | 2020-11-28T04:56:29 | {} | NONE |
ClickHouse | ClickHouse | 735,052,236 | 17,519 | Grian | **With more recent client it working:**
ClickHouse client version 20.5.2.7 (official build).
Connecting to database test_gg at click.test.int:9000 as user gg.
Connected to ClickHouse server version 20.10.4 revision 54441.
ClickHouse client version is older than ClickHouse server. It may lack support for new featu... | 2020-11-28T07:47:08 | 2020-11-28T07:53:16 | {} | NONE |
ClickHouse | ClickHouse | 735,089,055 | 17,055 | l1t1 | numbers_mt is better than numbers on AMD VM, but for big join, it's still very slow.
```
VM_0_52_centos :) select count(*) from numbers(100000)a,numbers(10000)b;
SELECT count(*)
FROM numbers(100000) AS a
, numbers(10000) AS b
Query id: d3a1f3cf-7e9e-4bd2-9968-2db5c47e0e7a
┌────count()─┐
│ 1000000000 │
└─... | 2020-11-28T08:16:31 | 2020-11-28T08:16:31 | {} | NONE |
ClickHouse | ClickHouse | 735,121,978 | 17,509 | alexey-milovidov | > Docs check — Docs check passed | 2020-11-28T08:39:23 | 2020-11-28T08:39:23 | {} | MEMBER |
ClickHouse | ClickHouse | 735,356,897 | 15,101 | alexey-milovidov | @KochetovNicolai It does not work. | 2020-11-29T07:47:32 | 2020-11-29T07:47:32 | {} | MEMBER |
ClickHouse | ClickHouse | 735,384,952 | 16,922 | alexey-milovidov | @yard
> 20.6.7.4 x86 on "proxy" server, 20.6.7.4 ARM builds on shards (did not try with x86 shards)
This is the reason.
Different hash function is used on different platforms to split data across buckets.
It means that heterogenous clusters (mixed ARM/x86) are not supported.
But I expect that the hash functi... | 2020-11-29T12:18:40 | 2020-11-29T12:19:23 | {} | MEMBER |
ClickHouse | ClickHouse | 735,388,518 | 7,009 | alexey-milovidov | @dujijun007 Different algorithms are selected for different number and combination of columns.
It's expected that the best algorithm is selected but it may not be the case.
I cannot investigate this case without the dataset. | 2020-11-29T12:47:28 | 2020-11-29T12:47:28 | {} | MEMBER |
ClickHouse | ClickHouse | 735,416,093 | 11,714 | ucasfl | Hi, I implemented this, but I don't know how to test it...... @alexey-milovidov | 2020-11-29T16:01:37 | 2020-11-29T16:01:37 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 734,854,935 | 17,471 | CurtizJ | `01323_redundant_functions_in_order_by` fails. Actually the test answer was wrong.
`EXPLAIN SYNTAX SELECT key, a FROM test ORDER BY key, exp(key + a)` was optimized to
```
SELECT
key,
a
FROM test
ORDER BY key ASC
```
It's incorrect since in this order by expression `a` is a tie-breaker in case of eq... | 2020-11-27T14:14:49 | 2020-11-27T14:14:49 | {} | MEMBER |
ClickHouse | ClickHouse | 735,166,389 | 17,508 | l1t1 | the Chinese document has similar problem, could you mark the machine translation parts? | 2020-11-28T09:10:08 | 2020-11-28T09:10:08 | {} | NONE |
ClickHouse | ClickHouse | 735,172,427 | 17,508 | alexey-milovidov | Chinese is actually the most active after english along with russian.
I personally do not like machine translation at all - I think it was a terrible idea to add it. But for Chinese there is a chance to make the translation normal by editing the articles.
PS. It should be already marked in metadata at the beginni... | 2020-11-28T09:14:16 | 2020-11-28T09:15:32 | {} | MEMBER |
ClickHouse | ClickHouse | 735,310,344 | 17,518 | chenziliang | > > Looks like this should be fixed by #16995
>
> Yes, look like so. @sundy-li , may you help confirm this ?
After discussing with @sundy-li , PR #16995 will fix this problem and more complete. Closing this one for now | 2020-11-29T00:33:27 | 2020-11-29T00:33:27 | {
"heart": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 735,376,516 | 17,453 | azat | > > Integration tests (asan) — fail: 0, passed: 961, error: 22
> > Integration tests (thread) — fail: 0, passed: 955, error: 22
> grpc (hm, seems that the checks just restarted, but will they pass eventually?)
Very unlikely, since those tests fails in upstream too | 2020-11-29T11:11:49 | 2020-11-29T11:11:49 | {} | MEMBER |
ClickHouse | ClickHouse | 735,378,053 | 17,526 | alexey-milovidov | GRPC tests: @vitlibar. | 2020-11-29T11:25:05 | 2020-11-29T11:25:05 | {} | MEMBER |
ClickHouse | ClickHouse | 735,389,036 | 10,971 | alexey-milovidov | The screenshot shows abysmally low performance but there is no info about the hardware and the data structure. | 2020-11-29T12:51:10 | 2020-11-29T12:51:10 | {} | MEMBER |
ClickHouse | ClickHouse | 735,433,428 | 14,414 | alexey-milovidov | Closed because of too many changes. | 2020-11-29T18:11:15 | 2020-11-29T18:11:15 | {} | MEMBER |
ClickHouse | ClickHouse | 734,903,830 | 17,497 | XavierQuiniou | Do I understand that we can not use this feature, when writing sql statement : INSERT INTO ... ? | 2020-11-27T16:17:02 | 2020-11-27T16:17:02 | {} | NONE |
ClickHouse | ClickHouse | 735,385,693 | 6,806 | alexey-milovidov | Fixed by @vdimir in #16804, starting from version 20.13-testing. | 2020-11-29T12:25:08 | 2020-11-29T12:25:08 | {
"+1": 1,
"hooray": 1
} | MEMBER |
ClickHouse | ClickHouse | 735,387,631 | 8,939 | alexey-milovidov | @CurtizJ is it fixed?
@KochetovNicolai what do you think? | 2020-11-29T12:40:21 | 2020-11-29T12:40:21 | {} | MEMBER |
ClickHouse | ClickHouse | 734,860,735 | 17,485 | akuzm | > Ok, but I'd like to make it automatically determined.
What I'm doing here is ignoring the frequent changes in the tests that we didn't manage to explain and eliminate yet. Just so that we don't waste time seeing them in every PR. Doing this automatically would be akin to just automatically ignoring all failing fun... | 2020-11-27T14:28:21 | 2020-11-27T14:28:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 734,902,350 | 17,465 | den-crane | note for myself: Need to emphasize diff order_by VS primary_key in the doc. | 2020-11-27T16:13:11 | 2020-11-27T16:13:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 735,224,386 | 17,518 | azat | Looks like this should be fixed by #16995 | 2020-11-28T12:18:34 | 2020-11-28T12:18:34 | {} | MEMBER |
ClickHouse | ClickHouse | 735,379,049 | 17,144 | alexey-milovidov | It is almost ready. | 2020-11-29T11:33:14 | 2020-11-29T11:33:14 | {} | MEMBER |
ClickHouse | ClickHouse | 735,388,796 | 10,917 | alexey-milovidov | It may simply indicate that 32 GiB is not enough. | 2020-11-29T12:49:42 | 2020-11-29T12:49:42 | {} | MEMBER |
ClickHouse | ClickHouse | 735,422,706 | 17,453 | azat | >If you know how to fix this test, it will be very appreciated...
@alexey-milovidov that was the timing issue, not the order issue - #17568 | 2020-11-29T16:47:58 | 2020-11-29T16:47:58 | {} | MEMBER |
ClickHouse | ClickHouse | 734,807,826 | 17,275 | alesapin | > > Push to dockerhub — Updated yandex/clickhouse-fasttest
>
> Looks like packages are broken, and any fasttest image update will not pass:
>
> ```
> /usr/bin/dpkg-deb: 2: Syntax error: newline unexpected
> dpkg: error processing archive /tmp/apt-dpkg-install-fdxX1t/00-perl-modules-5.30_5.30.0-9ubuntu0.2_all.de... | 2020-11-27T12:16:30 | 2020-11-27T12:16:30 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 734,961,395 | 16,877 | alesapin | Thread sanitizer fixed in https://github.com/ClickHouse/ClickHouse/pull/17498 | 2020-11-27T19:29:47 | 2020-11-27T19:29:47 | {} | MEMBER |
ClickHouse | ClickHouse | 735,356,167 | 17,509 | alexey-milovidov | ```
WITH
arrayMap(x -> toUInt32(x), splitByChar('.', extract(version, 'v(\\d+\\.\\d+\\.\\d+)'))) AS new,
arrayMap(x -> toUInt32(x), splitByChar('.', version())) AS old
SELECT *
FROM url('https://clickhouse.tech/data/version_date.tsv', TSV, 'version String, date Date')
WHERE new > old
``` | 2020-11-29T07:38:42 | 2020-11-29T07:38:42 | {} | MEMBER |
ClickHouse | ClickHouse | 735,363,594 | 17,463 | amosbird | It failed again... | 2020-11-29T09:02:21 | 2020-11-29T09:02:31 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 735,385,724 | 1,205 | alexey-milovidov | Fixed by @vdimir in #16804, starting from version 20.13-testing. | 2020-11-29T12:25:26 | 2020-11-29T12:25:26 | {} | MEMBER |
ClickHouse | ClickHouse | 735,389,879 | 14,099 | alexey-milovidov | @azat
`default_database` was invented for cross-replication when multiple servers may run on the same host.
Now it depends on how localhost optimization is performed. If it takes both host and port into account - everything is ok.
If only host - then it must be disabled.
Most likely the reason why it is disabl... | 2020-11-29T12:57:41 | 2020-11-29T12:57:41 | {} | MEMBER |
ClickHouse | ClickHouse | 751,693,671 | 18,563 | MaryamJFard | **The table**
CREATE TABLE db.mytable
(
`clmn_1` UInt64,
`clmn_2` UInt64,
`clmn_3` Int8,
`clmn_4` Nullable(String),
`clmn_5` DateTime,
`clmn_6` Date,
`clmn_7` DateTime,
`clmn_8` DateTime,
`clmn_9` DateTime
)
ENGINE = MergeTree()
ORDER BY clmn_1
SETTINGS index_granula... | 2020-12-28T12:18:01 | 2020-12-28T12:18:01 | {} | NONE |
ClickHouse | ClickHouse | 751,702,991 | 18,408 | spongedu | @alexey-milovidov Hi, I'd like to try on this if its ok. :) | 2020-12-28T12:54:05 | 2020-12-28T12:54:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 752,049,830 | 18,310 | alexey-milovidov | It's a good idea to do both :) | 2020-12-29T11:55:12 | 2020-12-29T11:55:12 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.