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 | 559,878,108 | 7,972 | den-crane | just stop CH and save folder with old table in the same filesystem
mkdir /var/tmp/
`mv /data/../ clicks_sharded /var/tmp/` or `mv /data/../clicks_sharded_old /var/tmp/`
save metadata `mv /metadata/../clicks_sharded_old.sql /var/tmp/` | 2019-11-29T20:43:02 | 2019-11-29T20:43:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 560,358,861 | 7,464 | blinkov | @Felixoid at least there are some file conflicts | 2019-12-02T11:36:06 | 2019-12-02T11:36:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 559,679,035 | 7,747 | hczhcz | @tavplubix
`...->getValue<UInt8>()` is used in many high-order functions (`arrayFilter`, `arrayCount`, `arrayAll`, etc).
It might be good to make developers aware the suspicious thing by adding some comments & docs. | 2019-11-29T06:48:24 | 2019-11-29T06:48:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 559,847,208 | 7,902 | den-crane | >What if I don't need to start it as service..
Sure. Not a problem at all.
And you can run as any user just change folders' owner.
```
$ sudo service clickhouse-server stop
$ sudo chown -R denis.zhuravlev:Domain.Users /var/lib/clickhouse /var/log/clickhouse-server /etc/clickhouse-server /etc/clickhouse-client... | 2019-11-29T17:10:11 | 2019-11-29T17:12:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 559,873,687 | 7,952 | den-crane | @kaijianding
https://clickhouse.yandex/docs/en/operations/table_engines/replacingmergetree/
Data deduplication occurs only during a merge. Merging occurs in the background at an unknown time, so you can't plan for it.
**Some of the data may remain unprocessed.** Although you can run an unscheduled merge using t... | 2019-11-29T20:05:39 | 2019-11-29T20:05:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 559,899,268 | 7,972 | bzon | @den-crane hmm what I did is the following for the server 2:
Fixing the startup problem.
* stop the server.
* move the clicks_sharded table and clicks_sharded metadata to a temporary folder.
* start-up the server again and now it's running fine again.
* recreate the tables. zookeeper data paths are now created... | 2019-11-30T00:29:13 | 2019-11-30T00:29:13 | {} | NONE |
ClickHouse | ClickHouse | 560,348,928 | 7,970 | svetlyak40wt | @alexey-milovidov why did you mark this issue as "invalid"?
Pages https://clickhouse.yandex/docs/en/development/build/ https://clickhouse.yandex/docs/en/development/build_osx/ say nothing about Clang. Build fails with this obscure error and there is no information about solving the issue. | 2019-12-02T11:07:43 | 2019-12-02T11:07:43 | {} | NONE |
ClickHouse | ClickHouse | 560,356,719 | 7,981 | alesapin | Fix is trivial, but tests for it is quite complex. I tested it locally and it works as expected.
Before fix:
```
:) select countDistinct(BannerID) from file('log_tsv*.tsv', 'TSV', 'EventTime DateTime, BannerID UInt64, Cost UInt64, CounterType UInt8');
2019.12.02 14:23:45.962774 [ 132 ] {8debc8cb-b8e8-4cf9-84c5-52... | 2019-12-02T11:29:14 | 2019-12-02T11:29:14 | {} | MEMBER |
ClickHouse | ClickHouse | 560,396,605 | 7,911 | nikitamikhaylov | Thanks!
| 2019-12-02T13:28:43 | 2019-12-02T13:28:43 | {} | MEMBER |
ClickHouse | ClickHouse | 560,396,170 | 7,911 | KochetovNicolai | done for 19.16, 19.15 and 19.14 | 2019-12-02T13:27:34 | 2019-12-02T13:29:31 | {} | MEMBER |
ClickHouse | ClickHouse | 559,495,252 | 7,957 | filimonov | If you will set move_factor to 0.8 it will try to preserve 80% of disk space free. So once the disk will be filled to 20.5% it will start moving parts (largest first) to the next volume. | 2019-11-28T13:29:59 | 2019-11-28T13:29:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 559,546,326 | 7,957 | filimonov | > Well, This means the process will not move any data from first disk to the next one if the available disk space is >= 80% ? I'm asking this question because i see its moved the data and the available disk space was 87%
It pushes data with move_factor between volumes, inside one volume it uses round-robin.
> So... | 2019-11-28T16:05:20 | 2019-11-28T16:05:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 559,587,389 | 7,935 | alexey-milovidov | The results are very promising! | 2019-11-28T19:16:13 | 2019-11-28T19:16:13 | {} | MEMBER |
ClickHouse | ClickHouse | 559,879,981 | 7,972 | den-crane | What CH version do you use?
Please show the end of **.sql files** from old and new tables with Engine and ... order by partition by | 2019-11-29T21:00:28 | 2019-11-29T21:00:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 559,888,938 | 7,972 | den-crane | Need to set enable_mixed_granularity_parts=1 in config.xml <merge_tree> section | 2019-11-29T22:28:38 | 2019-11-29T22:28:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 560,301,400 | 4,486 | KochetovNicolai | Yes, fixed by #4486 | 2019-12-02T09:06:20 | 2019-12-02T09:06:20 | {
"confused": 1
} | MEMBER |
ClickHouse | ClickHouse | 560,385,778 | 7,911 | KochetovNicolai | I can try to resolve conflict if you want | 2019-12-02T12:58:42 | 2019-12-02T12:58:42 | {} | MEMBER |
ClickHouse | ClickHouse | 559,586,949 | 7,935 | filimonov | intermediate results:
```
┌─version()────┬─record_count─┬─consume_time─┬──────consume_speed─┐
│ 19.14.7.15 │ 33554432 │ 203 │ 165292.7684729064 │
│ 19.15.3.6 │ 33554432 │ 623 │ 53859.4414125200 │
│ 19.17.1.1557 │ 33554432 │ 499 │ 67243.3507014028 │
│ 19.18.1.1745... | 2019-11-28T19:13:35 | 2019-11-29T07:54:45 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 559,820,696 | 7,902 | alex-rsk | May you point me, how to downgrade to any stable version, please?
I managed to run it as clickhouse, for a while
`
sudo -u clickhouse clickhouse-server --config=/etc/clickhouse-server/config.xml
` | 2019-11-29T15:16:31 | 2019-11-29T15:21:36 | {} | NONE |
ClickHouse | ClickHouse | 559,876,639 | 7,972 | bzon | @den-crane I'm about to do the rename and I got this error from ZK.
```
Code: 999. DB::Exception: Received from localhost:9000. Coordination::Exception. Coordination::Exception: No node, path: /clickhouse/tables/01/default/clicks_sharded/replicas/clickhouse-01-0/host.
```
I don't have clicks_sharded there, I on... | 2019-11-29T20:29:53 | 2019-11-29T20:31:25 | {} | NONE |
ClickHouse | ClickHouse | 559,878,560 | 7,972 | den-crane | Restart for what? What is the current state?
Do you have data in the new table? Do you see the old table in CH ? | 2019-11-29T20:47:11 | 2019-11-29T20:47:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 559,889,647 | 7,972 | den-crane | yes. and restart | 2019-11-29T22:35:36 | 2019-11-29T22:35:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 560,338,768 | 7,978 | Felixoid | This PR is a documentation patch for #4213 | 2019-12-02T10:40:38 | 2019-12-02T10:40:38 | {} | MEMBER |
ClickHouse | ClickHouse | 560,411,327 | 7,440 | alesapin | @SaltTan
This behavior seems to be expected and didn't change for a long time https://github.com/ClickHouse/ClickHouse/blame/master/dbms/src/Dictionaries/FlatDictionary.cpp#L391-L408. Are you sure, that before 19.10 everything works as you describe? | 2019-12-02T14:08:38 | 2019-12-02T14:08:38 | {} | MEMBER |
ClickHouse | ClickHouse | 559,780,728 | 7,935 | filimonov | OK. Now we are back with the original performance in the case when each message contains one row (tested on TSV). In cases when one message contains more rows new versions should be a bit faster than the older one.
```
# ┌─version()────┬─record_count─┬─consume_time─┬──────consume_speed─┐
# │ 19.14.7.15 │ 335... | 2019-11-29T12:53:29 | 2019-11-29T12:53:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 559,793,265 | 7,717 | MaxFedotov | Hi @yurriy, sorry for stupid mistakes - hadn't written on C++ since university :) I've applied all code review notes, can you please take a look?
Thank,
Max | 2019-11-29T13:39:47 | 2019-11-29T13:39:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 559,873,321 | 7,972 | bzon | @den-crane, I have 2 nodes currently, by **new** do you mean new replicated with a different name? | 2019-11-29T20:02:52 | 2019-11-29T20:02:59 | {} | NONE |
ClickHouse | ClickHouse | 559,874,968 | 7,972 | bzon | @den-crane I only see this `increment.txt` under the shadow folder.
```
root@clickhouse-01-0:/# cat /var/lib/clickhouse/shadow/increment.txt
1
```
Can you elaborate step 3 further, please? Thanks. | 2019-11-29T20:15:49 | 2019-11-29T20:16:58 | {} | NONE |
ClickHouse | ClickHouse | 559,960,432 | 7,812 | yuzhichang | I don't think the failed test relates to my change. | 2019-11-30T12:30:43 | 2019-11-30T12:30:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 560,371,146 | 32 | blinkov | @Rostish afaik nothing new here | 2019-12-02T12:14:57 | 2019-12-02T12:14:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 560,388,488 | 7,758 | nikitamikhaylov | Conflicts with 19.16
| 2019-12-02T13:06:40 | 2019-12-02T13:06:40 | {} | MEMBER |
ClickHouse | ClickHouse | 560,400,187 | 7,973 | alexey-milovidov | > And strings are not quoted
It means it cannot be based on CSV (because in CSV we quote all strings in output).
Maybe our new `CustomSeparated` format will suite better for this case? | 2019-12-02T13:38:31 | 2019-12-02T13:38:31 | {} | MEMBER |
ClickHouse | ClickHouse | 559,667,708 | 7,951 | zhang2014 | 👍 LGTM | 2019-11-29T05:50:49 | 2019-11-29T05:50:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 559,771,845 | 7,916 | alesapin | lgtm | 2019-11-29T12:17:08 | 2019-11-29T12:17:08 | {} | MEMBER |
ClickHouse | ClickHouse | 559,805,659 | 7,969 | chukynax | kafka logs
```[2019-11-29 14:17:15,188] INFO [GroupCoordinator 1002]: Stabilized group clickhouse generation 6 (__consumer_offsets-4) (kafka.coordinator.group.GroupCoordinator)
[2019-11-29 14:17:15,201] INFO [GroupCoordinator 1002]: Assignment received from leader for group clickhouse for generation 6 (kafka.coordi... | 2019-11-29T14:22:36 | 2019-11-29T14:22:36 | {} | NONE |
ClickHouse | ClickHouse | 559,824,180 | 7,902 | den-crane | > May you point me, how to downgrade to any stable version, please?
apt-get install clickhouse-server=19.17.4.11 clickhouse-client=19.17.4.11 clickhouse-common-static=19.17.4.11
Why do you start CH not with `sudo service clickhouse-server start` ?
| 2019-11-29T15:30:20 | 2019-11-29T15:30:20 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 559,874,423 | 7,972 | bzon | @den-crane thanks for the info. I will try it out. | 2019-11-29T20:11:37 | 2019-11-29T20:11:37 | {} | NONE |
ClickHouse | ClickHouse | 560,047,748 | 7,973 | sundy-li | > BTW, it is possible to write SET format_csv_delimiter = '\x01'
I just noticed that. It will be very helpful.
> Do you know how the character \x01 is escaped inside strings in TextFile format?
How the strings are quoted? How the NULLs are written?
'\x01' will not be escaped inside strings in TextFile format ... | 2019-12-01T03:26:42 | 2019-12-01T03:26:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 560,071,348 | 7,957 | MahmoudGoda0 | Thanks for the details, I will try to re-test this case again to make sure "move_factor" working fine.
So, Considering the following example, If I have 4 disks (2 SSD, 2 HDD) and "move_factor" = 0.8 and I would like to have the fresh data in SSDs disks and the old data to be moved to HDDs disk based on the move factor... | 2019-12-01T07:51:19 | 2019-12-01T07:51:19 | {} | NONE |
ClickHouse | ClickHouse | 560,363,241 | 7,464 | Felixoid | @blinkov conflicts are resolved | 2019-12-02T11:50:35 | 2019-12-02T11:50:35 | {} | MEMBER |
ClickHouse | ClickHouse | 559,493,745 | 6,991 | alex-krash | Or, even more crazy idea - "ExternalUDFStream". Write incoming data into RowBinary format into socket/pipe/mmaped files/MQ/stdin for "worker", that can be implemented in any programming language. | 2019-11-28T13:25:12 | 2019-11-28T13:25:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 559,496,682 | 7,952 | filimonov | ClickHouse do background merges all the time. There is background merge pool, It periodically wakes up to check if something should be merged. If it has nothing to do it can sleep a bit longer.
It uses some heuristics to peek the parts to be merged, it depends on part sizes, currently executed merges, number of the... | 2019-11-28T13:34:30 | 2019-11-28T13:34:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 559,498,077 | 7,957 | MahmoudGoda0 | > So once the disk will be filled to 20.5% it will start moving parts (largest first) to the next volume.
So it will move parts = 0.5 % of the disk space ? | 2019-11-28T13:38:56 | 2019-11-28T13:38:56 | {} | NONE |
ClickHouse | ClickHouse | 559,877,536 | 7,972 | bzon | The detach does not work. I think it was because of the previous command did not work properly.
```
clickhouse-01-0.clickhouses.clickshield.svc.cluster.local :) detach table clicks_sharded;
DETACH TABLE clicks_sharded
Received exception from server (version 19.14.3):
Code: 60. DB::Exception: Received from lo... | 2019-11-29T20:38:13 | 2019-11-29T20:39:27 | {} | NONE |
ClickHouse | ClickHouse | 559,896,361 | 7,972 | bzon | @den-crane I managed to get the first replica for each of my shards up and running!
When running the second replica of the shard I get this error.
```
<Error> Application: DB::Exception: Cannot create object 'clicks_sharded' from query ....
....
....
error: Coordination::Exception: No node, path: /clickhouse/... | 2019-11-29T23:50:07 | 2019-11-29T23:52:13 | {} | NONE |
ClickHouse | ClickHouse | 560,304,627 | 5,670 | alesapin | Bug fixed in all versions after 19.9. | 2019-12-02T09:14:53 | 2019-12-02T09:14:53 | {} | MEMBER |
ClickHouse | ClickHouse | 560,398,345 | 7,535 | alesapin | Cannot reproduce:
```
CREATE TABLE default.dict_invalidate
( ... | 2019-12-02T13:33:19 | 2019-12-02T13:33:19 | {} | MEMBER |
ClickHouse | ClickHouse | 559,492,387 | 7,955 | filimonov | Maybe related #7915 | 2019-11-28T13:20:52 | 2019-11-28T13:20:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 559,634,824 | 7,952 | JongMinP | It periodically wakes up to check if something should be merged
=>I wonder something example) insert?, event? | 2019-11-29T01:59:52 | 2019-11-29T01:59:52 | {} | NONE |
ClickHouse | ClickHouse | 559,738,397 | 7,935 | filimonov | intermediate results
```
# ┌─version()────┬─record_count─┬─consume_time─┬──────consume_speed─┐
# │ 19.14.7.15 │ 33554432 │ 203 │ 165292.7684729064 │
# │ 19.15.3.6 │ 33554432 │ 623 │ 53859.4414125200 │
# │ 19.17.1.1557 │ 33554432 │ 499 │ 67243.3507014028 │
# │ 19.... | 2019-11-29T10:16:03 | 2019-11-29T10:48:11 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 559,876,939 | 7,972 | den-crane | detach table table_name
mv /metadata/../ table_name.sql /var/tmp/
mv /data/../table_name /var/tmp/ | 2019-11-29T20:32:29 | 2019-11-29T20:33:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 559,993,922 | 7,952 | den-crane | @kaijianding
>but it doesn't happen for a very long time(maybe for one day)
it could happen never for that partition. Especially if no more inserts in that partition.
>any idea to make it happen more frequently
No such feature. Try run optimize. | 2019-11-30T16:46:00 | 2019-11-30T16:46:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 560,019,817 | 7,973 | alexey-milovidov | Do you know how the character \x01 is escaped inside strings in TextFile format?
How the strings are quoted? How the NULLs are written?
What about Date and DateTime data types? | 2019-11-30T20:45:21 | 2019-11-30T20:45:21 | {} | MEMBER |
ClickHouse | ClickHouse | 559,535,795 | 7,949 | den-crane | >Bro what if i use mysql as my main database for managing other parts, and
>just for hits and reporting on hits, i use clickhouse on same site same time ? what you think?
That's OK.
>instead of update, i will insert data in another table and on
>reporting i will use join to merge all records
Try https://c... | 2019-11-28T15:32:51 | 2019-11-28T15:32:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 559,875,306 | 7,972 | bzon | @den-crane
> What a response you got with ALTER TABLE table_name FREEZE ?
```
clickhouse-01-0.clickhouses.clickshield.svc.cluster.local :) ALTER TABLE clicks_sharded FREEZE
ALTER TABLE clicks_sharded
FREEZE
Ok.
``` | 2019-11-29T20:18:26 | 2019-11-29T20:18:26 | {} | NONE |
ClickHouse | ClickHouse | 559,903,406 | 7,972 | bzon | Thanks for all the help @den-crane! I really appreciate it. Closing this issue now! 🍻 | 2019-11-30T01:44:41 | 2019-11-30T01:44:41 | {} | NONE |
ClickHouse | ClickHouse | 560,348,007 | 7,970 | svetlyak40wt | And what should I do with this information? I don't want to dig into what Clang is. Just want to built a command-line client to ClickHouse. | 2019-12-02T11:05:12 | 2019-12-02T11:05:12 | {} | NONE |
ClickHouse | ClickHouse | 560,385,240 | 7,911 | nikitamikhaylov | Conflicts with 19.16
| 2019-12-02T12:57:09 | 2019-12-02T12:57:09 | {} | MEMBER |
ClickHouse | ClickHouse | 559,562,439 | 7,955 | 4ertus2 | Most probable engine Join related issue #7556 | 2019-11-28T17:04:34 | 2019-11-28T17:04:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 559,819,600 | 7,902 | alex-rsk | Ok, i added myself through usermod -a -G
```
groups alexrsk
alexrsk : alexrsk adm cdrom sudo dip plugdev lpadmin sambashare vboxusers clickhouse
```
And still got the same error...
Logging trace to console
2019.11.29 18:10:30.968014 [ 1 ] {} <Information> : Starting ClickHouse 19.17.4.11 with revision 54428
... | 2019-11-29T15:12:33 | 2019-11-29T15:14:04 | {} | NONE |
ClickHouse | ClickHouse | 559,844,912 | 7,970 | filimonov | I see the message: "Darwin build is supported only for Clang" in the output. :) | 2019-11-29T16:58:53 | 2019-11-29T16:58:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 559,999,072 | 7,973 | alexey-milovidov | > Currently set format_csv_delimiter = char(1) is not supported and select '\001' will translated to String 01.
BTW, it is possible to write `SET format_csv_delimiter = '\x01'` | 2019-11-30T17:21:22 | 2019-11-30T17:21:22 | {} | MEMBER |
ClickHouse | ClickHouse | 560,005,713 | 7,973 | alexey-milovidov | The existence of a special name for CSV with a specific separator is Ok. | 2019-11-30T18:00:47 | 2019-11-30T18:00:47 | {} | MEMBER |
ClickHouse | ClickHouse | 560,027,084 | 4,486 | orloffv | @KochetovNicolai fixed yet? | 2019-11-30T22:41:18 | 2019-11-30T22:41:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 560,384,740 | 7,916 | nikitamikhaylov | Conflicts with 19.16.
| 2019-12-02T12:55:55 | 2019-12-02T12:55:55 | {} | MEMBER |
ClickHouse | ClickHouse | 560,387,827 | 7,762 | nikitamikhaylov | Conflicts with 19.16
| 2019-12-02T13:04:36 | 2019-12-02T13:04:36 | {} | MEMBER |
ClickHouse | ClickHouse | 559,875,195 | 7,972 | den-crane | What a response you got with `ALTER TABLE table_name FREEZE` ? | 2019-11-29T20:17:37 | 2019-11-29T20:17:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 559,886,919 | 7,972 | bzon | After restarting the node I'm getting this error..
```
DB::Exception: Cannot create table from metadata file /var/lib/clickhouse/metadata/default//clicks_sharded.sql, error: DB::Exception: Table contains parts with adaptive and non adaptive marks, but `setting enable_mixed_granularity_parts` is disabled, stack trac... | 2019-11-29T22:08:02 | 2019-11-29T22:08:02 | {} | NONE |
ClickHouse | ClickHouse | 559,585,530 | 7,958 | alexey-milovidov | You have to download submodules:
```
git submodule update --init --recursive
```
as suggested. | 2019-11-28T19:05:14 | 2019-11-28T19:05:14 | {} | MEMBER |
ClickHouse | ClickHouse | 559,781,643 | 7,935 | filimonov | TODO:
- [ ] review
- [ ] cleanups and after-review fixes
- [ ] tests of inserts in different formats (probably will do that outside of your CI) | 2019-11-29T12:57:00 | 2019-11-29T12:57:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 559,872,992 | 7,972 | den-crane | https://clickhouse.yandex/docs/en/operations/table_engines/replication/#recovery-when-metadata-in-the-zookeeper-cluster-is-lost-or-damaged
You have to create new replicated tables and attach partitions.
To avoid data download try `ATTACH PARTITION ... FROM` https://github.com/ClickHouse/ClickHouse/issues/5312#issue... | 2019-11-29T20:00:18 | 2019-11-29T20:00:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 559,949,080 | 7,952 | kaijianding | @den-crane I get that merging occurs at an unknown time, but it doesn't happen for a very long time(maybe for one day), any idea to make it happen more frequently? | 2019-11-30T11:09:19 | 2019-11-30T11:09:19 | {} | NONE |
ClickHouse | ClickHouse | 560,404,362 | 7,971 | nikitamikhaylov | Which version of ClickHouse did you test? | 2019-12-02T13:50:03 | 2019-12-02T13:50:03 | {} | MEMBER |
ClickHouse | ClickHouse | 559,842,335 | 7,902 | alex-rsk | >
> Why do you start CH not with `sudo service clickhouse-server start` ?
In official documentation
What if I don't need to start it as service.. | 2019-11-29T16:46:49 | 2019-11-29T16:47:42 | {} | NONE |
ClickHouse | ClickHouse | 559,878,222 | 7,972 | bzon | @den-crane restart is required? I did that. | 2019-11-29T20:44:14 | 2019-11-29T20:44:14 | {} | NONE |
ClickHouse | ClickHouse | 559,897,164 | 7,972 | den-crane | >>create a new replicated table table_name at replica (it will download the data)
you need to execute create table at replicas with the same zk path as zk path at the first replicas | 2019-11-30T00:00:13 | 2019-11-30T00:00:13 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 559,557,022 | 7,943 | tavplubix | > You need to write comma separator and this query will be Ok.
But it should not be OK. It works only because `readIntTextUnsafe(...)` in `DataTypeNumberBase<T>::deserializeText(...)` doesn't throw an exception and doesn't change position in buffer when it cannot parse int, so zero is inserted. I think such behaviou... | 2019-11-28T16:42:51 | 2019-11-28T16:42:51 | {} | MEMBER |
ClickHouse | ClickHouse | 559,667,482 | 7,572 | zhang2014 | > I like it! The suggested design looks legit. Thank you for adding a number of tests. Please fix typos.
Afaiu, @abyss7 is willing to do an additional review which is great.
Done. Thank you for your review suggestion : ) | 2019-11-29T05:49:26 | 2019-11-29T05:49:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 559,715,933 | 7,952 | kaijianding | @filimonov I found duplicates in my ReplacingMergeTree table are not removed after 1 day, I have to do OPTIMIZE FINAL to remove duplicates. Any idea why background merge is not happening? My version is 19.7.3 | 2019-11-29T09:09:16 | 2019-11-29T09:09:16 | {} | NONE |
ClickHouse | ClickHouse | 559,927,147 | 3,484 | v1rusw0rm | I tried it on version 19.17.2 and it seems working. Is this issue not closed because it's not tested enough? | 2019-11-30T08:35:48 | 2019-11-30T08:35:48 | {
"confused": 1
} | NONE |
ClickHouse | ClickHouse | 560,005,618 | 7,973 | alexey-milovidov | TextFile is too good name for it. Maybe better to name it `HiveTextFile`? | 2019-11-30T18:00:10 | 2019-11-30T18:00:10 | {} | MEMBER |
ClickHouse | ClickHouse | 560,387,144 | 7,911 | nikitamikhaylov | There are many references to this PR. So, I think your changes will be useful. | 2019-12-02T13:02:22 | 2019-12-02T13:02:22 | {} | MEMBER |
ClickHouse | ClickHouse | 560,396,493 | 7,982 | alexey-milovidov | @excitoon build on master is run on every commit and you can check the results here:
https://github.com/ClickHouse/ClickHouse/commits/master | 2019-12-02T13:28:28 | 2019-12-02T13:28:28 | {} | MEMBER |
ClickHouse | ClickHouse | 560,405,069 | 7,935 | abyss7 | Looks good - let's make some cleanup and merge. | 2019-12-02T13:51:57 | 2019-12-02T13:51:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 559,497,525 | 7,957 | MahmoudGoda0 | Well, This means the process will not move any data from first disk to the next one if the available disk space is >= 80% ? I'm asking this question because i see its moved the data and the available disk space was 87% | 2019-11-28T13:37:03 | 2019-11-28T13:37:03 | {} | NONE |
ClickHouse | ClickHouse | 559,587,043 | 6,991 | alexey-milovidov | > Or, even more crazy idea - "ExternalUDFStream". Write incoming data into RowBinary format into socket/pipe/mmaped files/MQ/stdin for "worker", that can be implemented in any programming language.
It is one of the working ideas. Probably the most safe but least performant solution. BTW, it's already available via e... | 2019-11-28T19:14:13 | 2019-11-28T19:14:26 | {} | MEMBER |
ClickHouse | ClickHouse | 559,833,272 | 7,961 | den-crane | >when I pass where timezone=something
What do you mean pass?
```
TZ=Europe/Moscow cl --use_client_time_zone=1
create table TT(D DateTime) Engine=Memory;
insert into TT select now();
SELECT * FROM TT
┌───────────────────D─┐
│ 2019-11-29 19:02:46 │
└─────────────────────┘
SELECT * FROM TT
WHERE D =... | 2019-11-29T16:05:42 | 2019-11-29T16:37:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 559,879,024 | 7,972 | bzon | @den-crane you mentioned just stop CH. I have data in the temp table (new). I restarted clickhouse and it won't start, it's getting an error `DB::Exception: Cannot create table from metadata file /var/lib/clickhouse/metadata/default//clicks_sharded_tmp.sql, error: DB::Exception: Existing table metadata in ZooKeeper di... | 2019-11-29T20:51:23 | 2019-11-29T20:52:12 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 559,889,371 | 7,972 | bzon | @den-crane like this?
```
<yandex>
<merge_tree>
<enable_mixed_granularity_parts>1</enable_mixed_granularity_parts>
</merge_tree>
</yandex>
``` | 2019-11-29T22:32:51 | 2019-11-29T22:32:51 | {} | NONE |
ClickHouse | ClickHouse | 560,048,053 | 7,973 | sundy-li | And there are some other formats in Hive like `serd2`, it allows to set some custom properties like `quoteChar `, `separatorChar `
```
CREATE TABLE test(a string, b string,..)
ROW FORMAT SERDE 'org.apache.hadoop.hive.serde2.OpenCSVSerde'
WITH SERDEPROPERTIES (
"separatorChar" = ",",
"quoteChar" = "\""... | 2019-12-01T03:29:54 | 2019-12-01T03:29:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 559,843,405 | 7,961 | feribg | This works `SELECT * FROM TT WHERE D = toDateTime('2019-11-29 19:02:46', timezone the column is in)`
This casts the timezone to UTC first and then compares which is weird if the column has a non UTC timezone. Just not my expectation, but I can always cast them.
`SELECT * FROM TT WHERE D = '2019-11-29 19:02:46'... | 2019-11-29T16:51:56 | 2019-11-29T16:51:56 | {} | NONE |
ClickHouse | ClickHouse | 559,845,136 | 7,961 | den-crane | > the column has a non UTC timezone
this column attribute just a hint to clients how to render this values to strings.
>Do you usually recommend all UTC
yes. It makes less confusion. | 2019-11-29T16:59:56 | 2019-11-29T17:00:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 559,875,633 | 7,972 | bzon | @den-crane nevermind there is a directory called `1`. I did not notice it. 😆 | 2019-11-29T20:21:11 | 2019-11-29T20:21:11 | {} | NONE |
ClickHouse | ClickHouse | 559,876,310 | 7,972 | den-crane | better to use another ZK path for TEMP.
because in case of drop old replicated RO table it will try to remove data from ZK
or you can just `detach table table_name_old` and remove /data/../table_name_old /metadata/.../table_name_old.sql | 2019-11-29T20:26:57 | 2019-11-29T20:26:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 559,880,005 | 7,972 | bzon | @den-crane
I recreated steps 1 to 6 but in step 1, as I said, I recreated the replicated TEMP table using the ZK path of the ORIGINAL table name. So instead of clicks_sharded_tmp, I used clicks_sharded.
```
CREATE TABLE clicks_sharded_tmp
....
ENGINE = ReplicatedMergeTree('/clickhouse/tables/{shard}/default... | 2019-11-29T21:00:41 | 2019-11-29T21:08:22 | {} | NONE |
ClickHouse | ClickHouse | 560,339,755 | 7,980 | hieunguyen1809 | replace ' with " and it works | 2019-12-02T10:43:04 | 2019-12-02T10:43:04 | {} | NONE |
ClickHouse | ClickHouse | 560,360,145 | 5,789 | alesapin | Fixed | 2019-12-02T11:40:20 | 2019-12-02T11:40:20 | {} | MEMBER |
ClickHouse | ClickHouse | 560,398,744 | 7,981 | alexey-milovidov | @stavrolia Do we support globs in S3 storage? | 2019-12-02T13:34:27 | 2019-12-02T13:34:27 | {} | MEMBER |
ClickHouse | ClickHouse | 560,411,071 | 7,871 | blinkov | also there's a file conflict | 2019-12-02T14:07:58 | 2019-12-02T14:07:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 563,248,098 | 7,948 | victor-perov | It would be nice to get an update on the issue. | 2019-12-09T13:53:13 | 2019-12-09T13:53:13 | {} | NONE |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.