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 | 428,188,187 | 3,322 | dqminh | Indeed this is about partition directories, the `/block_numbers/<partition_id>` znodes themselves are empty as expected.
> simply deleting these nodes can result in data loss if there will be additional inserts into a dropped partition
I havent understood the concern here yet. If the partition is to be dropped a... | 2018-10-09T13:18:39 | 2018-10-09T13:18:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 429,081,294 | 3,347 | blinkov | #3354 | 2018-10-11T19:06:10 | 2018-10-11T19:06:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 429,335,492 | 3,362 | alexey-milovidov | TODO: Integration test. | 2018-10-12T13:59:37 | 2018-10-12T13:59:37 | {} | MEMBER |
ClickHouse | ClickHouse | 429,338,071 | 3,366 | alexey-milovidov | You can also add generic implementation:
- for any pair of identical types;
- with conversion to least supertype. | 2018-10-12T14:07:21 | 2018-10-12T14:07:21 | {} | MEMBER |
ClickHouse | ClickHouse | 429,673,867 | 3,377 | amosbird | how about adding comments here for usability?
https://github.com/yandex/ClickHouse/blob/master/dbms/src/Interpreters/InterpreterDescribeQuery.cpp#L47 | 2018-10-14T23:50:08 | 2018-10-14T23:50:08 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 428,388,854 | 3,210 | alexey-milovidov | We should think how to make this feature available for users with absolute minimum installation steps.
- create a jar with all dependencies inside?
- package it in selfcontained application (is it possible)?
- add it to the `clickhouse-common-static` package? | 2018-10-09T23:36:36 | 2018-10-09T23:36:36 | {} | MEMBER |
ClickHouse | ClickHouse | 428,396,503 | 3,210 | alex-krash | @alexey-milovidov , I've wrapped JDBC bridge into .deb package.
https://github.com/alex-krash/clickhouse-jdbc-bridge#installation
Maybe we should move this repo to Yandex one, and distribute as clickhouse-jdbc-bridge ?
Application consists of single jar with all the dependencies, and requires JDK to start this jar... | 2018-10-10T00:22:27 | 2018-10-10T00:34:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 428,452,743 | 3,223 | veloman-yunkan | @alexey-milovidov Nice work!
Was the original performance fully restored by these changes? | 2018-10-10T06:25:44 | 2018-10-10T06:25:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 428,534,526 | 1,460 | sinlang | @valyala is there any document to introduce how clickhourse using multi disk? how to assign and balance between disk;
> Just use [symbolic links](https://en.wikipedia.org/wiki/Symbolic_link).
>
> Suppose you have `<path>/opt/clickhouse/</path>`. Then create symbolic links in `/opt/clickhouse/data/` for every db y... | 2018-10-10T11:18:20 | 2018-10-10T11:18:20 | {} | NONE |
ClickHouse | ClickHouse | 428,576,683 | 3,339 | alexey-milovidov | Thank you!
PS. Maybe we should get rid of AST range and use something other for prepared sets indexation. | 2018-10-10T13:43:05 | 2018-10-10T13:43:05 | {} | MEMBER |
ClickHouse | ClickHouse | 429,331,442 | 3,336 | ldbl | @alexey-milovidov can you provide some info about the differences how where and prewhere works.
Because I read the docs and there is this info
> This clause has the same meaning as the WHERE clause. The difference is in which data is read from the table.
But I have looked into the patch and there is check fo... | 2018-10-12T13:47:30 | 2018-10-12T13:47:30 | {} | NONE |
ClickHouse | ClickHouse | 429,547,026 | 3,373 | alexey-milovidov | Thank you! I have also fixed the same typos in code. | 2018-10-13T14:34:30 | 2018-10-13T14:34:30 | {} | MEMBER |
ClickHouse | ClickHouse | 428,226,855 | 3,182 | VadimPlh | > lgtm, wait for tests
https://ru.wikipedia.org/wiki/%D0%96%D0%B4%D1%83%D0%BD#/media/File:Beelden_in_Leiden_2016_04_crop.jpg
| 2018-10-09T14:59:34 | 2018-10-09T15:00:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 428,387,159 | 3,210 | alexey-milovidov | > Nullability support abscence. For JDBC bridge I've implemented it (see below), maybe we should implement it for ODBC as well?
We want to add the support for Nullable types in ODBC and MySQL storages.
The logic will be, for `MySQL` and `ODBC` table engine:
- convert NULL to default value, if it is not Nullable ... | 2018-10-09T23:28:06 | 2018-10-09T23:37:14 | {} | MEMBER |
ClickHouse | ClickHouse | 428,947,696 | 3,253 | CurtizJ | Fixed #3340. | 2018-10-11T13:06:20 | 2018-10-11T13:06:20 | {
"hooray": 1
} | MEMBER |
ClickHouse | ClickHouse | 428,220,891 | 3,182 | alesapin | lgtm, wait for tests | 2018-10-09T14:44:59 | 2018-10-09T14:44:59 | {} | MEMBER |
ClickHouse | ClickHouse | 428,322,177 | 3,223 | alexey-milovidov | But this was only related to error messages. | 2018-10-09T19:34:43 | 2018-10-09T19:34:43 | {} | MEMBER |
ClickHouse | ClickHouse | 428,916,856 | 3,284 | EvgenyVinogradov | Same problem with ODBC(Postgres) dictionaries | 2018-10-11T11:15:49 | 2018-10-11T11:15:49 | {} | NONE |
ClickHouse | ClickHouse | 428,209,759 | 3,325 | fleapapa | Since MemSQL is not open sourced, i have no access to it for benchmark. I thought people at Clickhouse would be more curious than me to learn how MemSQL 6.5 performs against Clickhouse.
Nevertheless, thanks for the response. | 2018-10-09T14:17:05 | 2018-10-09T14:17:05 | {} | NONE |
ClickHouse | ClickHouse | 428,291,595 | 3,282 | 4ertus2 | It would be fixed for Decimal32 and Decimal64 in #3329
Decimal128 needs much more work to support. So there would be the same exception for it. | 2018-10-09T18:04:05 | 2018-10-09T18:04:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 428,544,053 | 3,328 | dijkstra33 | Поправили.
Ошибка была в том, что использовали "-" в файле odbc.ini, в названии: gameserver-private.
Заменили на gameserverprivate и всё заработало.
Ошибка которую получали в итоге очень запутывала, только админ смог заподозрить неладное в символе "-". | 2018-10-10T11:57:55 | 2018-10-10T11:57:55 | {
"+1": 5
} | NONE |
ClickHouse | ClickHouse | 428,577,687 | 3,337 | alexey-milovidov | Please add a test in separate PR. | 2018-10-10T13:45:52 | 2018-10-10T13:45:52 | {} | MEMBER |
ClickHouse | ClickHouse | 429,024,997 | 3,284 | Zer0Divis0r | The schema of my dictionary follows. Which of the fields is nullable type? How can they be done not nullable...?
```
CREATE TABLE `sampling_history` (
`id` int(10) UNSIGNED NOT NULL,
`websiteId` int(11) UNSIGNED NOT NULL,
`dateFrom` date NOT NULL DEFAULT '2017-06-01',
`dateTill` date NOT NULL DEFAULT '205... | 2018-10-11T16:32:06 | 2018-10-11T16:32:06 | {} | NONE |
ClickHouse | ClickHouse | 429,086,272 | 3,350 | alexey-milovidov | It failed to build in our CI.
https://clickhouse-builds.s3.yandex.net/3350/c1b8f7cc3106bbc039330343a4822963d11f0a9d/report.html | 2018-10-11T19:22:54 | 2018-10-11T19:22:54 | {} | MEMBER |
ClickHouse | ClickHouse | 429,660,729 | 3,378 | den-crane | workaround: concat with empty string:
select toString(y)||'' z from
(select toDecimal128(5.5,12) y)
group by z | 2018-10-14T20:43:07 | 2018-10-14T20:43:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 428,156,895 | 3,299 | healiseu | Yes, it accepts it this way
Thanks | 2018-10-09T11:27:55 | 2018-10-09T11:27:55 | {} | NONE |
ClickHouse | ClickHouse | 428,384,862 | 3,275 | alexey-milovidov | In my opinion, we should not bloat HTTP headers.
Clients may have troubles with large size of headers or small limits on total size of headers. | 2018-10-09T23:16:23 | 2018-10-09T23:16:23 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 428,387,762 | 3,210 | alexey-milovidov | As I understand, it is disabled right now.
Nevertheless, I've merged this PR to make further development easier.
Things to do:
- enable this feature;
- prepare a manual on how to start using it after installing ClickHouse:
-- how to install JDK;
-- how to install JDBC bridge;
-- how to install example database... | 2018-10-09T23:31:14 | 2018-10-09T23:31:14 | {} | MEMBER |
ClickHouse | ClickHouse | 428,577,724 | 3,339 | alexey-milovidov | Please add a test in separate PR. | 2018-10-10T13:45:59 | 2018-10-10T13:45:59 | {} | MEMBER |
ClickHouse | ClickHouse | 428,775,123 | 2,922 | alexey-milovidov | Testing packages are uploaded automatically when functional tests have passed (for normal release build and for build with sanitizers). This is just a small subset of our [tests](https://clickhouse.yandex/docs/en/development/tests/), but for the last week, even that basic set of tests have not passed due to the bug, th... | 2018-10-11T00:17:25 | 2018-10-11T00:24:40 | {} | MEMBER |
ClickHouse | ClickHouse | 428,776,043 | 2,922 | alexey-milovidov | > On the other hand, we have plans to automatically publish packages for "unstable" builds in each pull request by direct links in GitHub comments.
We made this happen.
Click on th... | 2018-10-11T00:23:13 | 2018-10-11T00:25:08 | {} | MEMBER |
ClickHouse | ClickHouse | 429,209,526 | 3,311 | PhantomPhreak | > We can fix the issue by introducing a method for masking parameters.
> It can be located in `IStorage` and called from `InterpreterShowCreateQuery` and `StorageSystemTables`.
It would be very nice. Thank you in advance! | 2018-10-12T05:34:54 | 2018-10-12T05:34:54 | {} | NONE |
ClickHouse | ClickHouse | 429,337,470 | 3,366 | alexey-milovidov | `test_deduplication_window_in_seconds` - flappy test. | 2018-10-12T14:05:27 | 2018-10-12T14:05:27 | {} | MEMBER |
ClickHouse | ClickHouse | 428,133,078 | 3,299 | filimonov | Try that:
```xml
<attribute>
<name>ndx</name>
<type>UInt16</type>
<null_value>0</null_value>
</attribute>
```
(See my comment above) | 2018-10-09T09:57:12 | 2018-10-09T09:57:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 428,167,438 | 1,522 | sleshJdev | I am experienced the exactly same issue, but in my case I don't have much less load: I pull files from s3 bucket one by one and pipe them to clickhouse thru clickhouse-client. And sometimes I don't now why I see discrepancy of data that seems like some of files was not imported correctly. The only errors I found in ser... | 2018-10-09T12:09:16 | 2018-10-09T12:09:16 | {} | NONE |
ClickHouse | ClickHouse | 428,506,442 | 3,311 | alex-krash | @PhantomPhreak , unfortunately it looks like there is a bug in mysql odbc. It reproduces in ClickHouse integration tests, as well as my experience on local machine (Ubuntu 16.04) | 2018-10-10T09:38:40 | 2018-10-10T09:38:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 428,695,378 | 2,866 | alexey-milovidov | #3341 | 2018-10-10T19:09:57 | 2018-10-10T19:09:57 | {} | MEMBER |
ClickHouse | ClickHouse | 428,978,801 | 3,284 | den-crane | Probably this ability (nullable types for range_min range_max at backend) was removed with https://github.com/yandex/ClickHouse/issues/2093 (it was released in 18.12.17)
I suggest to try workaround and cast to notnull type using view (mysql/pg), though it will break index access.
| 2018-10-11T14:34:06 | 2018-10-11T14:39:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 429,175,409 | 3,284 | alexey-milovidov | Broken after #3123
Steps to reproduce: https://gist.github.com/alexey-milovidov/c31b37f56114b822c7fdcf21d8557c9d
| 2018-10-12T01:38:52 | 2018-10-12T01:38:52 | {} | MEMBER |
ClickHouse | ClickHouse | 429,184,331 | 3,362 | alexey-milovidov | Example:
```
CREATE TABLE mysql_information_schema_tables (
`TABLE_CATALOG` Nullable(String),
`TABLE_SCHEMA` Nullable(String),
`TABLE_NAME` Nullable(String),
`TABLE_TYPE` Nullable(String),
`ENGINE` Nullable(String),
`VERSION` Nullable(UInt64),
`ROW_FORMAT` Nullable(String),
`TABLE_ROWS` Null... | 2018-10-12T02:35:08 | 2018-10-12T02:35:08 | {} | MEMBER |
ClickHouse | ClickHouse | 429,187,586 | 3,339 | alexey-milovidov | Why don't add a test? | 2018-10-12T02:55:39 | 2018-10-12T02:55:39 | {} | MEMBER |
ClickHouse | ClickHouse | 429,393,120 | 3,198 | snorecone | @filimonov no we're not running with docker at all. Kafka, Zookeeper and Clickhouse are all running on ec2 instances - ubuntu 16.04; clickhouse is installed from packages - `18.12.13 revision 54407`. | 2018-10-12T17:01:11 | 2018-10-12T17:01:11 | {} | NONE |
ClickHouse | ClickHouse | 428,180,691 | 3,322 | ztlpn | The problem is not with block number nodes per se, but with partition directories: a block number is allocated by creating an ephemeral sequential node `/<path-to-table>/block_numbers/<partition_id>/block-XXXXXXXXXX` and the directory `/<path-to-table>/block_numbers/<partition_id>` remains even after the partition has ... | 2018-10-09T12:55:34 | 2018-10-09T12:56:28 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 428,315,682 | 3,223 | veloman-yunkan | Thanks! | 2018-10-09T19:14:25 | 2018-10-09T19:14:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 428,316,346 | 3,223 | alexey-milovidov | There is significant performance degradation comparing to version 18.12.17.
```
$ time dbms/programs/clickhouse local --query="SELECT count() FROM table" --input-format=JSONEachRow --structure='...'
user 0m33.576s
$ time ~/clickhouse-18.12.17 local --query="SELECT count() FROM table" --input-format=JSONEac... | 2018-10-09T19:16:13 | 2018-10-09T19:16:13 | {} | MEMBER |
ClickHouse | ClickHouse | 428,582,390 | 3,269 | liuyimin-bytedance | https://github.com/yandex/ClickHouse/pull/3339 | 2018-10-10T13:58:30 | 2018-10-10T13:58:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 429,099,396 | 3,353 | healiseu | Got it, thank you @alexey-milovidov and @blinkov
```sql
SELECT
rowNumberInAllBlocks() AS rowNo,
*
FROM
(
SELECT
trip_pk,
trip_srt_station_id
FROM BicycleTripFlat
ORDER BY trip_srt_station_id ASC
LIMIT 10
)
FORMAT TSVWithNames
```
and output is
```txt
row... | 2018-10-11T20:04:58 | 2018-10-11T20:07:18 | {} | NONE |
ClickHouse | ClickHouse | 429,283,093 | 3,363 | 4ertus2 | First affected revision 94c9a544 (10.10.18)
Fixed in master in 218366e1 (12.10.18) | 2018-10-12T10:38:18 | 2018-10-12T10:39:23 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 429,625,121 | 3,374 | frankie4fingers | It seems it broken starts from 18.14.* I checked today several published versions and only 18.12.17 works fine for me. All shows the same error reported above. | 2018-10-14T13:08:04 | 2018-10-14T13:08:04 | {
"+1": 2
} | NONE |
ClickHouse | ClickHouse | 428,783,143 | 2,922 | alexey-milovidov | > please re-open as last successful build looks to be 18.14.1 on Sep 28, 2018:
Now you can see that 18.14.6 has been published. | 2018-10-11T01:07:27 | 2018-10-11T01:07:27 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 429,604,957 | 3,375 | kshvakov | OK, I'll add a test soon. | 2018-10-14T08:00:53 | 2018-10-14T08:00:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 428,403,584 | 3,326 | SunguckLee | > Just to ensure - you call clickhouse-client with --ask-password, do you?
Yes | 2018-10-10T01:08:38 | 2018-10-10T01:08:49 | {} | NONE |
ClickHouse | ClickHouse | 428,502,241 | 3,311 | PhantomPhreak | @filimonov thanks for your suggestion with ODBC, i tried this way but unfortunately, I couldn't make it work :(
I configured ODBC using this [manual](https://dev.mysql.com/doc/connector-odbc/en/connector-odbc-configuration-dsn-unix.html), checked that it works correctly with
```
# isql -v myodbc8w
```
Then i spaw... | 2018-10-10T09:25:19 | 2018-10-10T09:42:22 | {} | NONE |
ClickHouse | ClickHouse | 428,673,845 | 3,311 | alexey-milovidov | > @PhantomPhreak , unfortunately it looks like there is a bug in mysql odbc. It reproduces in ClickHouse integration tests, as well as my experience on local machine (Ubuntu 16.04)
`myodbc` shared library is incorrectly linked in Ubuntu - it exposes some additional symbols besides the ODBC interface (it should expos... | 2018-10-10T18:08:48 | 2018-10-10T18:08:48 | {} | MEMBER |
ClickHouse | ClickHouse | 428,682,577 | 3,311 | alexey-milovidov | We can fix the issue by introducing a method for masking parameters.
It can be located in `IStorage` and called from `InterpreterShowCreateQuery` and `StorageSystemTables`. | 2018-10-10T18:33:04 | 2018-10-10T18:33:04 | {} | MEMBER |
ClickHouse | ClickHouse | 428,953,468 | 1,784 | PetarLotrean | @ben-motiv
Have you tried OPTIMIZE table FINAL?
https://github.com/yandex/ClickHouse/issues/2086
| 2018-10-11T13:23:38 | 2018-10-11T13:23:38 | {} | NONE |
ClickHouse | ClickHouse | 429,242,827 | 3,362 | alesapin | LGTM | 2018-10-12T08:11:00 | 2018-10-12T08:11:00 | {} | MEMBER |
ClickHouse | ClickHouse | 429,262,552 | 3,363 | 4ertus2 | Hi,
It's known regression that appears a few days ago. You may try previous build or wait a fix. | 2018-10-12T09:21:47 | 2018-10-12T09:21:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 429,276,171 | 3,363 | zahlio | We tried in stable and on test, both had the issue, but awesome you guys are working on a fix. | 2018-10-12T10:10:39 | 2018-10-12T10:10:39 | {} | NONE |
ClickHouse | ClickHouse | 429,617,130 | 3,376 | alesapin | hmmm, very strange... Without abs path in start command CH still can found own executable and run odbc bridge. | 2018-10-14T11:01:03 | 2018-10-14T11:01:03 | {} | MEMBER |
ClickHouse | ClickHouse | 428,321,404 | 3,223 | alexey-milovidov | The bug was present even before this refactoring, when skipping unknown fields.
https://github.com/veloman-yunkan/ClickHouse/blob/fafecb3c25a59777ed691c0a1936d66181d4d093/dbms/src/Formats/JSONEachRowRowInputStream.cpp#L122
name_ref is invalidated at this point. | 2018-10-09T19:32:11 | 2018-10-09T19:32:11 | {} | MEMBER |
ClickHouse | ClickHouse | 428,402,553 | 2,586 | zhang2014 | @alexey-milovidov I will commit the first version within this week. Thank you for waiting. | 2018-10-10T01:01:45 | 2018-10-10T01:01:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 428,763,574 | 2,922 | reflection | @alexey-milovidov please re-open as last successful build looks to be 18.14.1 on Sep 28, 2018: http://repo.yandex.ru/clickhouse/deb/testing/main/
| 2018-10-10T23:11:06 | 2018-10-10T23:11:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 428,979,476 | 3,282 | 4ertus2 | I was wrong. It's not much work to support Decimal128 too #3348 | 2018-10-11T14:35:58 | 2018-10-11T14:35:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 429,098,136 | 3,353 | blinkov | @healiseu try the trick suggested by Alexey | 2018-10-11T20:01:01 | 2018-10-11T20:01:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 429,116,843 | 3,336 | alexey-milovidov | PREWHERE simply cannot work with arrayJoin. | 2018-10-11T21:00:19 | 2018-10-11T21:00:19 | {} | MEMBER |
ClickHouse | ClickHouse | 429,303,484 | 3,198 | filimonov | > Kafkacat and native consume\produce scripts of kafka dist work normally
Kafkacat is more representative for debugging (it uses the same lib for Kafka access as ClickHouse).
> work normally with local kafka.
What do you mean by local? Can you check to consume messages by kafkacat in the same configuration wh... | 2018-10-12T12:05:56 | 2018-10-12T12:05:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 428,282,520 | 3,322 | ztlpn | > I havent understood the concern here yet. If the partition is to be dropped and we have very late arriving data as you said, i would expect such data to also be dropped.
The problem is that there no way to guarantee consistency between replicas. One replica can execute DROP PARTITION before the new block arrives (... | 2018-10-09T17:36:39 | 2018-10-09T17:36:39 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 428,322,451 | 3,223 | alexey-milovidov | > One way of fixing the problem while preserving the original optimization intact would be to defer the call to skipColonDelimiter() until the variable name_ref was fully consumed, however that would result in worse code (skipColonDelimiter() would need to be called in three different places where it doesn't really bel... | 2018-10-09T19:35:39 | 2018-10-09T19:35:39 | {} | MEMBER |
ClickHouse | ClickHouse | 428,667,554 | 3,271 | rajadayalan | There is difference between counts/aggregations of source table and materialized views also. Please find below the counts/aggregations.
Counts from Source table (unreplicated)
----------------------------------------
srv002 :) select count(), sum(hits) from logs.access_logs_local where request_date='2018-10-09'
... | 2018-10-10T17:50:53 | 2018-10-10T17:59:12 | {} | NONE |
ClickHouse | ClickHouse | 429,162,218 | 3,284 | den-crane | @Enmk
check this, range_hashed does not work with any type
Received exception from server (version 18.14.6):
Code: 50. DB::Exception: Received from localhost:9000, ::1. DB::Exception: Unsupported type Nullable(UInt64).
Code: 50. DB::Exception: Received from localhost:9000, ::1. DB::Exception: Unsupported ty... | 2018-10-12T00:10:41 | 2018-10-12T00:25:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 429,254,424 | 3,360 | k0st1an | ```
ii libiodbc2:amd64 3.52.9-2 amd64 iODBC Driver Manager
ii libodbc1:amd64 2.3.1-4.1 amd64 ODBC library for Unix
ii odbc-postgresql:amd64 1:10.03.0000-1.pgdg16.04+1 ... | 2018-10-12T08:53:31 | 2018-10-12T08:53:31 | {} | NONE |
ClickHouse | ClickHouse | 429,547,373 | 3,143 | alexey-milovidov | Yes. It is fixed in 18.14.8 (this version will be released today). | 2018-10-13T14:40:03 | 2018-10-13T14:40:03 | {
"+1": 2
} | MEMBER |
ClickHouse | ClickHouse | 429,565,584 | 3,198 | simross | Sorry for misunderstanding. I have only one implementation of kafka on local environment - 3-node docker based cluster. Kafkacat successfully produce and consume messages from this cluster. If clickhouse server is running on host machine - bug takes place. If clickhouse added as a service to docker network with kafka -... | 2018-10-13T18:42:36 | 2018-10-13T18:42:36 | {} | NONE |
ClickHouse | ClickHouse | 429,651,693 | 2,595 | alexey-milovidov | There was an issue with proper serialization of expressions in columns definition in ReplicatedMergeTree.
It was present in all versions before 1.1.54387 and was fixed in https://github.com/yandex/ClickHouse/commit/8901414a
It is possible to execute an ALTER query on ReplicatedMergeTree with proper escaping: `spl... | 2018-10-14T18:50:30 | 2018-10-14T18:50:30 | {} | MEMBER |
ClickHouse | ClickHouse | 428,205,474 | 3,243 | vpanfilov | It is a duplicate of https://github.com/yandex/ClickHouse/issues/146 | 2018-10-09T14:05:58 | 2018-10-09T14:05:58 | {} | NONE |
ClickHouse | ClickHouse | 428,389,271 | 2,586 | alexey-milovidov | How is it going? | 2018-10-09T23:38:54 | 2018-10-09T23:38:54 | {} | MEMBER |
ClickHouse | ClickHouse | 429,078,977 | 3,353 | alexey-milovidov | There is `rowNumberInAllBlocks` function.
It returns the row numbers for data it processes, and the output depends on a context this function is executed. To ensure proper context, you can wrap your query in a subquery:
```
SELECT *, rowNumberInAllBlocks() FROM (SELECT ...)
``` | 2018-10-11T18:58:29 | 2018-10-11T18:58:29 | {
"+1": 10
} | MEMBER |
ClickHouse | ClickHouse | 429,187,153 | 3,210 | alexey-milovidov | > We want to add the support for Nullable types in ODBC and MySQL storages.
The logic will be, for MySQL and ODBC table engine:
convert NULL to default value, if it is not Nullable in ClickHouse table;
read NULL value if it is Nullable in ClickHouse table.
This is implemented in #3362
> For mysql and ... | 2018-10-12T02:53:22 | 2018-10-12T02:53:22 | {} | MEMBER |
ClickHouse | ClickHouse | 429,249,645 | 3,350 | alex-krash | @alexey-milovidov , regarding failure on Yandex CI, I see error in each build:
```
CMake Error at contrib/base64-cmake/CMakeLists.txt:16 (add_library):
Cannot find source file:
/build/contrib/base64/lib/lib.c
Tried extensions .c .C .c++ .cc .cpp .cxx .m .M .mm .h .hh .h++ .hm .hpp
.hxx .in .txx
```... | 2018-10-12T08:36:26 | 2018-10-12T08:36:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 429,425,397 | 3,269 | 4ertus2 | Test for issue #3370 | 2018-10-12T18:54:01 | 2018-10-12T18:54:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 429,517,026 | 3,143 | sivasamyk | We are running ClickHouse version 18.12.14 revision 54407. In our case MemoryTracking metrics is behaving like a counter (it keeps increasing). Currently it is reporting around `170GB`.
```
select metric,value from system.metrics where metric='MemoryTracking';
SELECT
metric,
value
FROM system.metrics... | 2018-10-13T06:57:13 | 2018-10-13T06:57:13 | {} | NONE |
ClickHouse | ClickHouse | 428,224,218 | 3,223 | alexey-milovidov | Now I will test for performance regression and merge. | 2018-10-09T14:53:15 | 2018-10-09T14:53:15 | {} | MEMBER |
ClickHouse | ClickHouse | 429,078,408 | 3,353 | blinkov | Try https://clickhouse.yandex/docs/en/query_language/functions/other_functions/#rownumberinallblocks , maybe it'll work fine for your usecase. | 2018-10-11T18:56:41 | 2018-10-11T18:56:41 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 429,105,738 | 3,336 | alexey-milovidov | Confirmed. It is related to arrayJoin inside PREWHERE. | 2018-10-11T20:25:12 | 2018-10-11T20:25:12 | {} | MEMBER |
ClickHouse | ClickHouse | 429,341,697 | 3,284 | alexey-milovidov | Fixed in master. | 2018-10-12T14:18:31 | 2018-10-12T14:18:31 | {} | MEMBER |
ClickHouse | ClickHouse | 429,559,792 | 3,371 | vpanfilov | You should just change the alias for `sum(allTransCount)` column. You are using the column `allTransCount` for `response Rate` calculation. | 2018-10-13T17:25:06 | 2018-10-13T17:25:06 | {
"+1": 6
} | NONE |
ClickHouse | ClickHouse | 428,169,055 | 3,321 | alesapin | Failed:
stateless.00411_long_accurate_number_comparison
stateless.00730_unicode_terminal_format
| 2018-10-09T12:15:12 | 2018-10-09T12:15:12 | {} | MEMBER |
ClickHouse | ClickHouse | 428,353,037 | 3,223 | alexey-milovidov | Simply returning back the optimization gives just 3 of 8 seconds. | 2018-10-09T21:08:59 | 2018-10-09T21:08:59 | {} | MEMBER |
ClickHouse | ClickHouse | 428,362,911 | 3,309 | alexey-milovidov | Fixed in master. | 2018-10-09T21:41:43 | 2018-10-09T21:41:43 | {} | MEMBER |
ClickHouse | ClickHouse | 428,736,557 | 3,292 | alexey-milovidov | TODO:
- style errors;
- rename `FinishMergeSortingBlockInputStream` to `FinishSortingBlockInputStream` because it doesn't merge anything. | 2018-10-10T21:21:56 | 2018-10-10T21:21:56 | {} | MEMBER |
ClickHouse | ClickHouse | 429,039,639 | 2,922 | reflection | Thanks for the detailed replies. Makes sense that packages are only uploaded to repo after functional tests pass, and unstable builds from pull requests are pretty cool. | 2018-10-11T17:10:31 | 2018-10-11T17:10:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 429,259,022 | 3,198 | simross | 1. Kafkacat and native consume\produce scripts of kafka dist work normally with local kafka.
2. Clickhouse was installed from package
3. Nothing suspicious in either log.
4. 3 nodes of kafka.
Kafka is running via following docker-compose:
```
version: '2'
services:
zookeeper:
image: wurstmeister/zoo... | 2018-10-12T09:09:31 | 2018-10-12T09:09:31 | {} | NONE |
ClickHouse | ClickHouse | 429,583,710 | 3,375 | alexey-milovidov | Ok.
PS. Could you try to add integration test?
These tests are located in `dbms/tests/integration`.
More info: https://clickhouse.yandex/docs/en/development/tests/ | 2018-10-13T23:31:59 | 2018-10-13T23:31:59 | {} | MEMBER |
ClickHouse | ClickHouse | 428,143,671 | 3,324 | 4ertus2 | Make check - github tumeout (gcc-7, others are success)
Autotest - timeout (stateless.00411_long_accurate_number_comparison) | 2018-10-09T10:35:46 | 2018-10-09T10:35:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 428,384,343 | 3,275 | alexey-milovidov | Questionable.
```
Alexey Milovidov, [04.10.18 21:33]
Можно, но три вопроса:
- почему разрешение - один процент; в типичном случае отображения прогресс-бара будет дёргаться;
- почему величина в процентах, а не в доле; для машинно-читаемых значений лучше использовать долю от единицы;
- зачем это нужно, ведь это -... | 2018-10-09T23:13:35 | 2018-10-09T23:13:35 | {} | MEMBER |
ClickHouse | ClickHouse | 428,641,935 | 3,223 | alexey-milovidov | Yes, now performance is even a little better, than in 18.12.17.
```
18.12.17 (baseline):
user 0m25.540s
This pull request:
user 0m33.576s
Restored optimization (direct reference to the input buffer without a copy):
user 0m29.848s
Added optimization (caching of field order), that was mentioned i... | 2018-10-10T16:34:47 | 2018-10-10T16:34:47 | {} | MEMBER |
ClickHouse | ClickHouse | 428,779,355 | 3,344 | vavrusa | This was broken in https://github.com/yandex/ClickHouse/commit/dd88c092b772aeea882a564b6edc5303deaf98f8
It's reproducible if this block is commented out: https://github.com/yandex/ClickHouse/blob/master/dbms/src/Interpreters/ProcessList.cpp#L270 | 2018-10-11T00:44:03 | 2018-10-11T00:44:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 429,125,533 | 3,342 | alexey-milovidov | ```SELECT materialize(1) ? materialize(toDate('0000-00-00')) : toDate('0000-00-00')```
doesn't work. | 2018-10-11T21:29:53 | 2018-10-11T21:29:53 | {} | MEMBER |
ClickHouse | ClickHouse | 429,343,717 | 3,350 | alesapin | @alex-krash The failure is bug in CI system. Submodule was not cloned. | 2018-10-12T14:24:28 | 2018-10-12T14:24:28 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.