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 | 404,874,257 | 205 | alexey-milovidov | @blinkov We can collect the links to public statements about ClickHouse usage in production.
There are plenty of them. But we just not collecting. | 2018-07-13T15:52:00 | 2018-07-13T15:52:00 | {} | MEMBER |
ClickHouse | ClickHouse | 404,925,615 | 2,643 | alexey-milovidov | Could you please also update `ru` version? | 2018-07-13T19:07:08 | 2018-07-13T19:07:08 | {} | MEMBER |
ClickHouse | ClickHouse | 404,929,467 | 2,643 | ilyabreev | Sure. Here it is #2647 | 2018-07-13T19:22:30 | 2018-07-13T19:22:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,009,218 | 46 | blinkov | If you need daily data purge, nowadays you should consider using custom partitioning:
https://clickhouse.yandex/docs/en/table_engines/custom_partitioning_key/ | 2018-07-14T08:42:39 | 2018-07-14T08:42:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,196,685 | 522 | blinkov | @mathewsp79 do you need any more help here? | 2018-07-16T09:46:05 | 2018-07-16T09:46:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,277,645 | 780 | SaltTan | I can't remember when the last time it happened. I wouldn't say we get a lot of crashes though :) | 2018-07-16T15:02:03 | 2018-07-16T15:02:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,396,896 | 2,325 | bobrik | OpenJDK is using both shares and determine CPU limit:
* http://hg.openjdk.java.net/jdk/jdk/rev/7f22774a5f42 | 2018-07-16T22:08:03 | 2018-07-16T22:08:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,549,023 | 2,642 | alex-zaitsev | Currently: read-merge-write. Don't forget that merge has to decompress and compress the data that is CPU intensive.
With fetch only: network-write
Less CPU and I/O. | 2018-07-17T11:24:53 | 2018-07-17T11:24:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,703,791 | 867 | silviucpp | @BayoNet was documentation added ? In case yes you can close the bug | 2018-07-17T19:42:17 | 2018-07-17T19:42:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,705,006 | 867 | alexey-milovidov | It wasn't. | 2018-07-17T19:46:45 | 2018-07-17T19:46:45 | {} | MEMBER |
ClickHouse | ClickHouse | 405,730,963 | 2,665 | alexey-milovidov | Looks Ok. Thank you! | 2018-07-17T21:17:17 | 2018-07-17T21:17:17 | {} | MEMBER |
ClickHouse | ClickHouse | 404,869,984 | 205 | blinkov | Unfortunately, it's not that easy from legal standpoint. Though if company you work for is ok about stating publicly that's it's using ClickHouse in production — feel free to do so. | 2018-07-13T15:37:09 | 2018-07-13T15:37:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 404,972,508 | 2,640 | macdonsp | I'm able to reproduce this using a simple example, and it seems related to the use of 'insert_distributed_sync'. I provisioned a single node using a local Zookeeper installation and the following remote_servers configuration:
```
<remote_servers>
<!-- Test only shard config for testing distributed stor... | 2018-07-13T22:48:51 | 2018-07-13T22:48:51 | {} | NONE |
ClickHouse | ClickHouse | 405,054,247 | 2,608 | alexey-milovidov | 664 мегабайта - это нормально для исполняемого файла ClickHouse с отладочной информацией.
Убрать отладочную информацию можно так: `strip clickhouse`, но это не обязательно. Наличие отладочной информации не влияет ни на производительность ни на время старта сервера. | 2018-07-14T22:35:42 | 2018-07-14T22:35:42 | {} | MEMBER |
ClickHouse | ClickHouse | 405,144,432 | 2,652 | alexey-milovidov | The documentation is not quite correct.
The first row should contain correct number of columns.
We would like to add an option to use header for reordering columns in the future. | 2018-07-16T03:59:55 | 2018-07-16T03:59:55 | {} | MEMBER |
ClickHouse | ClickHouse | 405,190,458 | 911 | blinkov | Is this issue still relevant? | 2018-07-16T09:19:36 | 2018-07-16T09:19:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,201,087 | 654 | blinkov | Most of it has been implemented as `ALTER TABLE ... DELETE ...`, see: https://clickhouse.yandex/docs/en/query_language/queries/#mutations | 2018-07-16T10:04:49 | 2018-07-16T10:04:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,202,547 | 720 | blinkov | What exactly this integration is supposed to allow to do? | 2018-07-16T10:11:47 | 2018-07-16T10:11:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,313,026 | 2,657 | furmur | thank you ) i had to check default sort behavior first | 2018-07-16T16:52:27 | 2018-07-16T16:52:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,648,350 | 2,664 | alexey-milovidov | Thank you! It looks clearly better. | 2018-07-17T16:41:37 | 2018-07-17T16:41:37 | {} | MEMBER |
ClickHouse | ClickHouse | 405,816,743 | 2,671 | alexey-milovidov | I doubt that is makes anything better. | 2018-07-18T05:49:52 | 2018-07-18T05:49:52 | {} | MEMBER |
ClickHouse | ClickHouse | 406,139,938 | 2,670 | alexey-milovidov | Missing a test for changes in `arrayEnumerateUniq` (that it started to work in generic case). | 2018-07-19T03:04:15 | 2018-07-19T03:04:15 | {} | MEMBER |
ClickHouse | ClickHouse | 406,291,634 | 2,668 | champtar | If I split the computation per hour it works just fine
```
SELECT quantilesTDigestMerge(0.5, 0.9, 0.99)(s)
FROM
(
SELECT
toStartOfHour(datetime),
quantilesTDigestState(0.5, 0.9, 0.99)(response_time) AS s
FROM vme4
GROUP BY toStartOfHour(datetime)
)
┌─quantilesTDigestMerge(0... | 2018-07-19T14:16:21 | 2018-07-19T14:16:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,416,822 | 2,628 | alexey-milovidov | Server will clean old tmp files on startup. So, the issue will be solved by server restart.
Why we don't remove tmp file if it is already exist before creating new file? This is left as an additional protection against concurrent creates (we open file with O_EXCL; though concurrent creates are already protected by `... | 2018-07-17T00:00:54 | 2018-07-17T00:01:29 | {} | MEMBER |
ClickHouse | ClickHouse | 405,444,717 | 2,662 | mattrobenolt | Ok, after investigating a bit more, this seems directly tied to the `Array(Nullable(String))` type. If I use `Array(String)` all of this behaves as I'd expect and `[1]` is silently returned as an empty string (in this case) without a segfault. | 2018-07-17T03:03:31 | 2018-07-17T03:03:31 | {} | NONE |
ClickHouse | ClickHouse | 405,640,105 | 2,666 | alexey-milovidov | We also have in order of 10..20 million of znodes in our ZK cluster. This still work Ok.
At this moment `use_minimalistic_checksums_in_zookeeper` helps for our case. | 2018-07-17T16:15:29 | 2018-07-17T16:15:29 | {} | MEMBER |
ClickHouse | ClickHouse | 405,642,489 | 2,606 | shawel | @filimonov thanks for the suggestion! I have about 5000 categories in total. I will try and see if creating columns on the prefix of the categories might help to cut it down to about ~500, or maybe convert to columns only heavily used categories.
Thanks! | 2018-07-17T16:22:38 | 2018-07-17T16:22:38 | {} | NONE |
ClickHouse | ClickHouse | 405,703,554 | 2,026 | silviucpp | @alexey-milovidov I tested this on master and it's fine. You can close it !
Thanks,
Silviu | 2018-07-17T19:41:28 | 2018-07-17T19:41:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 406,037,880 | 2,673 | dmitrydrynov | the same problem
I try
<listen_host>::</listen_host>
In previous version its worked. I updated clickhouse and has this problem now | 2018-07-18T18:55:47 | 2018-07-18T18:55:47 | {} | NONE |
ClickHouse | ClickHouse | 405,222,072 | 2,630 | proller | https://github.com/yandex/ClickHouse/pull/2656/files | 2018-07-16T11:48:50 | 2018-07-16T11:48:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,387,234 | 2,655 | VadimPlh | This is expected result, because `transform()` and `CASE` use `DefaultImplementationForNulls` which returns NULL. It will be fixed.
You can find it [here](https://github.com/yandex/ClickHouse/blob/0fde78584f359435371dd91e39ed942d95ec9c0a/dbms/src/Functions/IFunction.h#L62) | 2018-07-16T21:26:18 | 2018-07-16T21:52:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,715,130 | 2,669 | alexey-milovidov | Probably it will be better to log inside `readSuffix` method? | 2018-07-17T20:22:26 | 2018-07-17T20:23:19 | {} | MEMBER |
ClickHouse | ClickHouse | 405,809,120 | 2,663 | chenxing-xc | ok, simple test is added into 0_stateless directory | 2018-07-18T04:58:27 | 2018-07-18T04:58:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,160,257 | 1,952 | lamberken | +1, @alexey-milovidov @ztlpn hi, we also meet the message. version 1.1.54388
how to solve it, thank you very match
| 2018-07-16T06:43:12 | 2018-07-16T06:47:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,187,651 | 974 | blinkov | Is this issue still relevant? If so, is there any way to reproduce it? | 2018-07-16T09:08:37 | 2018-07-16T09:08:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,573,648 | 2,606 | filimonov | >I was using something like arrayExists(x-> x='keyword', category) on the category column and filter out user id for specific categories. and this does a full scan (i have about 500mil rows)
>My hope is to cut down on the number of rows by leveraging the index.
If number of categories is limited - you can just add... | 2018-07-17T13:02:15 | 2018-07-17T13:04:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,739,232 | 2,669 | leonardoce | Fixed. Thanks! | 2018-07-17T21:48:01 | 2018-07-17T21:48:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 406,141,012 | 2,670 | alexey-milovidov | Missing a test for tuple arguments and some other complex types (like multidimensional arrays). | 2018-07-19T03:12:04 | 2018-07-19T03:12:04 | {} | MEMBER |
ClickHouse | ClickHouse | 406,219,627 | 2,635 | wowtous | good job | 2018-07-19T09:45:52 | 2018-07-19T09:45:52 | {} | NONE |
ClickHouse | ClickHouse | 406,284,515 | 2,682 | VadimPlh | You can use `--echo` | 2018-07-19T13:55:08 | 2018-07-19T13:55:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 406,343,648 | 2,674 | Amper | I have the same problem.
My output for current version from master (with "-j1" option): [build.log](https://github.com/yandex/ClickHouse/files/2210896/build.log)
MacOS Sierra 10.13.5
gcc version 8.1.0 (Homebrew GCC 8.1.0)
| 2018-07-19T16:51:11 | 2018-07-19T16:51:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,502,405 | 1,770 | VladislavPershin | @deonwise, @alexm93, @simPod, есть такое костыль :
SELECT uuid
FROM test.test
WHERE uuid IN ('5b2f98bd-272b-020c-6004-145100001003')
┌─────────────────────────────────uuid─┐
│ 5b2f98bd-272b-020c-6004-145100001003 │
└────────────────────────────────── ──┘
1 rows in set. Elapsed: 0.0... | 2018-07-17T08:28:03 | 2018-07-17T08:29:40 | {
"heart": 6
} | NONE |
ClickHouse | ClickHouse | 405,718,999 | 2,669 | leonardoce | > Probably it will be better to log inside readSuffix method?
Sure. I'll fix the patch. | 2018-07-17T20:35:57 | 2018-07-17T20:35:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,964,047 | 2,668 | champtar | Hi @alexey-milovidov,
Thanks for the tip, I'm just starting to fully evaluate ClickHouse and haven't tested `quantilesTiming`
I also have the following 2 requests that I expect will have the same issue:
```
SELECT quantilesTDigest(0.5,0.9,0.99)(response_size) FROM vme4;
SELECT quantilesTDigest(0.5,0.9,0.99)(respon... | 2018-07-18T15:05:14 | 2018-07-18T15:05:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,190,121 | 868 | blinkov | To give answer on why this happened someone has to spend some time reproducing it, and since you have already found the workaround it doesn't make much sense to do so, especially given this issue has been abandoned for a while and that ClickHouse version is pretty old by now. | 2018-07-16T09:18:17 | 2018-07-16T09:18:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,650,412 | 2,663 | alexey-milovidov | Thank you!
Could you please also add a test (into `dbms/tests/queries/0_stateless` directory)? | 2018-07-17T16:48:35 | 2018-07-17T16:48:35 | {} | MEMBER |
ClickHouse | ClickHouse | 406,015,677 | 2,674 | proller | 1. please show full cmake output.
2. please find and show error (maybe easier via run ninja -j1 ) your details have only warnings and nothing about actual error.
3. what version you try to compile? some issues was fixed in current master
| 2018-07-18T17:42:32 | 2018-07-18T17:42:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 406,250,665 | 2,612 | Joozty | Anyway, this is not related to the original question I asked. Therefore I am closing it now. @filimonov thanks again for your help. :) | 2018-07-19T11:56:19 | 2018-07-19T11:56:19 | {} | NONE |
ClickHouse | ClickHouse | 404,871,393 | 198 | blinkov | Yep, nowadays the recommended approach for resharding is clickhouse-copier. | 2018-07-13T15:41:58 | 2018-07-13T15:41:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 404,877,436 | 86 | blinkov | There's more or less complete support of NULLs in ClickHouse already implemented, the documentation is coming. | 2018-07-13T16:02:55 | 2018-07-13T16:02:55 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,396,716 | 2,325 | bobrik | If you want to use throttling, be aware that it's broken:
* https://bugzilla.kernel.org/show_bug.cgi?id=198197 | 2018-07-16T22:07:06 | 2018-07-16T22:07:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,417,338 | 2,606 | alexey-milovidov | > Is there any way to achieve this outside of exploding each category per row.
No, this is the only way. | 2018-07-17T00:04:01 | 2018-07-17T00:04:01 | {} | MEMBER |
ClickHouse | ClickHouse | 405,687,998 | 2,668 | alexey-milovidov | Just in case: if you want to calculate quantiles from response time in milliseconds, better to use `quantilesTiming` function, which is designed specifically for this purpose. It is:
- faster;
- more precise;
- deterministic. | 2018-07-17T18:48:27 | 2018-07-17T18:48:27 | {} | MEMBER |
ClickHouse | ClickHouse | 404,966,814 | 2,452 | blinkov | I'm grouping up issues about updates in #167 | 2018-07-13T22:12:28 | 2018-07-13T22:12:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,198,073 | 887 | blinkov | @bobrik could you please provide some instructions on how to reproduce this? | 2018-07-16T09:51:58 | 2018-07-16T09:51:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 406,154,245 | 1,952 | tust05102405 | +1, version 1.1.54390
any idea for solve it？ @alexey-milovidov @ztlpn
thank you very match | 2018-07-19T04:52:49 | 2018-07-19T04:52:49 | {} | NONE |
ClickHouse | ClickHouse | 405,306,680 | 2,657 | alexey-milovidov | Missing test. | 2018-07-16T16:31:14 | 2018-07-16T16:31:14 | {} | MEMBER |
ClickHouse | ClickHouse | 405,469,440 | 2,655 | siberex | At least it would be nice to have it documented.
And it is unexpected because it have completely different implementation in any other SQL-flavoured DB:
```
mysql> SELECT CASE 1
-> WHEN 0 THEN 'ZERO'
-> WHEN 1 THEN 'ONE'
-> ELSE 'NONE'
-> END AS result;
+--------+
| result |
+--------+... | 2018-07-17T06:03:20 | 2018-07-17T06:04:17 | {} | NONE |
ClickHouse | ClickHouse | 405,577,663 | 2,263 | AntonSaykovsky | > does --format_csv_delimiter= support in v1.1.54385-stable?
> Errors:
> Bad arguments: unrecognised option '--format_csv_delimiter= '
1.1.54390 version has the same error. =((( | 2018-07-17T13:15:43 | 2018-07-17T13:15:43 | {} | NONE |
ClickHouse | ClickHouse | 405,796,714 | 2,666 | alexey-milovidov | > Question: if we enable this setting, is it compatible with residual data in zk and local host?
Yes.
> is there any command to refresh the checksum, which already exists in zk, using this minimal format ?
No. Old checksums will be replaced during background merges. | 2018-07-18T03:22:38 | 2018-07-18T03:22:38 | {} | MEMBER |
ClickHouse | ClickHouse | 405,352,864 | 887 | bobrik | I don't think I can a year after. Feel free to close it, but I don't think saying it's not a Clickhouse for a segfault is fair. | 2018-07-16T19:22:57 | 2018-07-16T19:22:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,823,270 | 2,671 | amosbird | @alexey-milovidov It's targeted for the TODO item. It might generate better code when SSE2 is not available. | 2018-07-18T06:26:56 | 2018-07-18T06:27:18 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 406,142,645 | 2,671 | amosbird | > For better example of Duff devices look at
I'm impressed by the imperative efforts in those examples
> We need any proof that it makes at least something better.
Yeah, I agree. Lemme try something to make it at least reasonable... | 2018-07-19T03:24:03 | 2018-07-19T03:24:03 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 406,205,757 | 2,676 | zhang2014 | 👍
But integration testing doesn't work well on the mac. 😞 It took me a lot of time to give up. should plan to support it ? | 2018-07-19T08:57:13 | 2018-07-19T08:57:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 406,226,685 | 2,608 | arduanov | Проблема решена в версии 1.1.54394, и размер бинарника уменьшился в 2 раза.
Не смог раньше предоставить дополнительную инфу. | 2018-07-19T10:11:50 | 2018-07-19T10:11:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 404,966,597 | 2,452 | blinkov | > by the way, why the code responsible for removing of rows is in AlterQuery (ALTER DELETE vs DELETE)
To make it clear that those are not cheap operations that can be done with high rate (unlike normal DELETEs in RDBMS).
> if you give me some tips how to deal with that?
Point ALTER towards local tables, not th... | 2018-07-13T22:11:14 | 2018-07-13T22:11:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,054,367 | 2,635 | alexey-milovidov | Confirmed. We should quote all identifiers in a query for external database. | 2018-07-14T22:38:44 | 2018-07-14T22:38:44 | {} | MEMBER |
ClickHouse | ClickHouse | 405,160,833 | 9 | alexey-milovidov | #2521 | 2018-07-16T06:47:09 | 2018-07-16T06:47:09 | {} | MEMBER |
ClickHouse | ClickHouse | 405,791,454 | 2,666 | chenxing-xc | Question: if we enable this setting, is it compatible with residual data in zk and local host? and is there any command to refresh the checksum, which already exists in zk, using this minimal format ? | 2018-07-18T02:46:47 | 2018-07-18T02:46:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 406,138,452 | 2,671 | alexey-milovidov | 1. Every x86_64 machine is guaranteed to have SSE2.
2. For example, SSE2 is unavailable on AArch64 machines. But for them your optimization shouldn't be the first - it's much better to use NEON instructions in main loop.
3. Using Duff device or unrolling loops doesn't mean that code will be more efficient. Your l... | 2018-07-19T02:54:33 | 2018-07-19T02:54:33 | {} | MEMBER |
ClickHouse | ClickHouse | 406,180,820 | 1,287 | filimonov | Some related talks from Telegram channel:
>Alessandro Rizzo:
>Ok. During ReplicatedMergeTree tables creation, I need to set ZK path and replica name.
>But with 3 nodes sharded and 2 replicas for each one, I will have a structure like that:
>```
>node1: shard1replica1, shard3replica2
>node2: shard2replica1, sha... | 2018-07-19T07:24:43 | 2018-07-19T07:24:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,185,710 | 81 | blinkov | Nowadays it's mentioned in docs:
```
operations/server_settings/settings.md:313:- level – Logging level. Acceptable values: ``trace``, ``debug``, ``information``, ``warning``, ``error``.
operations/server_settings/settings.md:323: <level>trace</level>
``` | 2018-07-16T09:00:19 | 2018-07-16T09:00:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,492,089 | 2,645 | champtar | I did some more tests, no problem using a `MergeTree` or `AggregatingMergeTree` | 2018-07-17T07:48:55 | 2018-07-17T07:48:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 406,081,567 | 1,287 | ajardan | We also have multiple shards and host a replica of each shard on every node, which basically makes ON CLUSTER requests unusable...
| 2018-07-18T21:34:10 | 2018-07-18T21:34:10 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 405,371,173 | 887 | alexey-milovidov | The relevant code was rewritten and now we have more checks when reading arrays. | 2018-07-16T20:28:11 | 2018-07-16T20:28:11 | {} | MEMBER |
ClickHouse | ClickHouse | 405,576,463 | 2,665 | kvap | cc @champtar, @alexey-milovidov, @bocharov, @vavrusa, @twalwyn, @Dorokhov | 2018-07-17T13:11:52 | 2018-07-17T13:11:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,639,381 | 2,666 | alexey-milovidov | Recently we have introduced an option to lower the size of znodes.
(This doesn't change amount of different znodes but total size of ZK memory and snapshot is much lower.)
To enable, add
```
<merge_tree>
<use_minimalistic_checksums_in_zookeeper>1</use_minimalistic_checksums_in_zookeeper>
</merge_tree>
```
... | 2018-07-17T16:13:09 | 2018-07-17T16:13:09 | {} | MEMBER |
ClickHouse | ClickHouse | 405,761,982 | 1,770 | simPod | Thanks for sharing! A bit non-standard tho | 2018-07-17T23:40:32 | 2018-07-17T23:40:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,798,096 | 2,632 | thanhthai3010 | I just found the solution to this problem. I create a new account for MongoDB and then re-connect with clickhouse.
```
db.createUser(
{
user: "userReadTest",
pwd: "passwordUserRead",
roles: [ { role: "read", db: "test" } ]
}
)
``` | 2018-07-18T03:32:58 | 2018-07-18T03:33:19 | {} | NONE |
ClickHouse | ClickHouse | 406,013,102 | 2,673 | filimonov | Please check [listen_host](https://clickhouse.yandex/docs/en/operations/server_settings/settings/#listen_host) directive and your firewall/network settings. | 2018-07-18T17:34:02 | 2018-07-18T17:34:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 406,225,906 | 2,684 | filimonov | 1) there were few issues with IN operator in recent versions of Clickhouse
2) try to replace WHERE to PREWHERE in your query.
3) Isn't your kkt_sn is some kind of autoincrement / other unique id for accessing single record? If so - please remember that Clickhouse is not good for point selects. It's main scenario is d... | 2018-07-19T10:08:46 | 2018-07-19T10:08:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 406,228,143 | 2,682 | VadimPlh | Yes, it is normal. If your query fell, you would see error message. | 2018-07-19T10:17:30 | 2018-07-19T10:17:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 404,853,737 | 248 | blinkov | @bookin , do you have any more plans/questions regarding this issue? | 2018-07-13T14:41:35 | 2018-07-13T14:41:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 404,935,333 | 2,642 | filimonov | Does it have sence? What are advantages of that scenario?
I.e. currently merge process time = read data from disk time + write data to disk time. Bottleneck is disk io, not CPU, as merge sort in case if source parts are already sorted (they are) should be executed in linear time.
After change it would be:
merge ... | 2018-07-13T19:47:25 | 2018-07-13T20:06:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,008,578 | 2,648 | blinkov | It's not that critical, but free args usually go after named args, so it'd better be like `echo 'CREATE TABLE t (a UInt8) ENGINE = Memory' | curl --data-binary @- 'http://localhost:8123/'`
Also we might add `-v` arg, because most of those command just quietly exit and it's not clear whether it was successful or not. | 2018-07-14T08:29:34 | 2018-07-14T08:29:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,144,760 | 2,645 | alexey-milovidov | Looks like that the bug was introduced in #2566 | 2018-07-16T04:03:17 | 2018-07-16T04:03:17 | {} | MEMBER |
ClickHouse | ClickHouse | 405,204,492 | 780 | blinkov | @SaltTan is this issue still relevant? | 2018-07-16T10:20:41 | 2018-07-16T10:20:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,834,085 | 1,861 | PangKuo | @ludv1x hi, I found the non-leader replica send query to the leader using default user, too. However, we don't use default database and default user in most case. Do you know any way to customize user and password? Thank you very much! | 2018-07-18T07:17:15 | 2018-07-18T07:17:15 | {} | NONE |
ClickHouse | ClickHouse | 404,876,770 | 205 | blinkov | I'd still not put it anywhere near official website without their written approval. Just collecting the list of what's already out there is possible of course. | 2018-07-13T16:00:33 | 2018-07-13T16:00:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 404,927,059 | 2,546 | filimonov | `--help` works for clickhouse-client. | 2018-07-13T19:12:55 | 2018-07-13T19:12:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 404,965,711 | 2,179 | blinkov | Batch deletes has been released few weeks ago, batch updates are in progress. | 2018-07-13T22:05:33 | 2018-07-13T22:05:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,161,900 | 2,595 | perez1987 | Sorry for the delay in answering @filimonov
Version: 1.1.54385
Exactly query:
ALTER TABLE BBDD.Table ADD COLUMN Filename DEFAULT splitByChar('\\', Path)[ length (splitByChar('\\', Path)) ] AFTER Path
If I execute a select instead of an alter I execute the split without problems.
However, if I execute alte... | 2018-07-16T06:54:37 | 2018-07-16T06:54:37 | {} | NONE |
ClickHouse | ClickHouse | 404,875,246 | 134 | blinkov | This has been implemented back in 2017, see input_format_allow_errors_ratio and input_format_allow_errors_num settings. | 2018-07-13T15:55:20 | 2018-07-16T06:56:20 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,187,976 | 973 | blinkov | Have you tried inserting directly to ReplicatedMergeTree tables instead of Distributed one? | 2018-07-16T09:09:53 | 2018-07-16T09:09:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,465,383 | 1,567 | champtar | Hi @KochetovNicolai, do you have an estimation when this feature is going to be ready ?
Is the branch already testable ? | 2018-07-17T05:38:56 | 2018-07-17T05:38:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,564,184 | 2,645 | kvap | That's my bad, the fix is on the way. | 2018-07-17T12:26:25 | 2018-07-17T12:26:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,635,524 | 2,667 | champtar | Hi @TH-HA
you might be interested in this PR #1567 | 2018-07-17T16:01:23 | 2018-07-17T16:01:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 405,731,498 | 2,636 | alexey-milovidov | Changelog for both 1.1.54390 and 1.1.54394 is going to be prepared this week. | 2018-07-17T21:19:03 | 2018-07-17T21:19:03 | {} | MEMBER |
ClickHouse | ClickHouse | 405,762,184 | 2,611 | simPod | https://github.com/yandex/ClickHouse/issues/1770#issuecomment-405502405 | 2018-07-17T23:41:47 | 2018-07-17T23:41:47 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 406,114,957 | 2,673 | proller | what clickhouse prints about listen on start? | 2018-07-19T00:25:07 | 2018-07-19T00:25:07 | {} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.