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 | 574,356,449 | 7,878 | filimonov | Related: #7533 #1226
| 2020-01-14T20:23:23 | 2020-01-14T20:23:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 574,565,882 | 8,657 | alexey-milovidov | Ok, but actually it doesn't matter. | 2020-01-15T09:12:39 | 2020-01-15T09:12:39 | {} | MEMBER |
ClickHouse | ClickHouse | 574,662,911 | 8,664 | den-crane | Duplicate https://github.com/ClickHouse/ClickHouse/issues/8553 | 2020-01-15T13:37:38 | 2020-01-15T13:37:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 575,023,509 | 8,679 | 865357384 | I see the status of my delete operation statement is “is_done=1” in this table. but The amount of data has doubled. but this problem is hard to reproduce. | 2020-01-16T07:38:14 | 2020-01-16T07:39:01 | {} | NONE |
ClickHouse | ClickHouse | 575,499,045 | 2,278 | winter7 | > Now data parts are loaded in parallel.
can you tell me which version starts to support parallel loading? | 2020-01-17T06:52:36 | 2020-01-17T06:52:36 | {} | NONE |
ClickHouse | ClickHouse | 574,385,187 | 8,168 | den-crane | But actually maybe it's kinda rare use-case and this 'workaround'? works:
```
CREATE TABLE mskipi
(
`A` Int64,
`I1` Int64,
`I2` Int64,
INDEX i1i (I1, I2) TYPE minmax GRANULARITY 1
)
ENGINE = MergeTree
ORDER BY A
insert into mskipi select number, 0,0 from numbers(10000000);
insert into m... | 2020-01-14T21:36:08 | 2020-01-14T21:36:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 574,565,849 | 7,646 | ivanprostran | > users.xml (select * from system.settings)
>
> ```
> <?xml version="1.0"?>
> <yandex>
> <profiles>
> <default>
> <background_schedule_pool_size>25</background_schedule_pool_size>
> ```
>
> also users
>
> ```
> <max_insert_block_size>10485760</max_insert_b... | 2020-01-15T09:12:35 | 2020-01-15T09:14:04 | {} | NONE |
ClickHouse | ClickHouse | 574,603,269 | 5,371 | alexey-milovidov | Thank you!
The only remaining part is to implement full support for numeric ranges. | 2020-01-15T10:44:07 | 2020-01-15T10:44:07 | {} | MEMBER |
ClickHouse | ClickHouse | 574,615,185 | 7,646 | ivanprostran | 'background_schedule_pool_size' has been increased to 50
```
SELECT *
FROM system.settings
WHERE name = 'background_schedule_pool_size'
┌─name──────────────────────────┬─value─┬─changed─┬─description──────────────────────────────────────────────────────────────────────────────────────────────┐
│ background_sc... | 2020-01-15T11:16:46 | 2020-01-15T11:16:46 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 574,751,139 | 7,799 | v-isitor | Unfortunately we had the same issue, when we upgraded our cluster to the stable version 19.17.4.11.
I analyzed the exception's stack trace and the source code and it seems to be cause by a performance improvement, which was introduced with version 19.17.4.11:
"Run another pass of syntax/expression analysis to get pot... | 2020-01-15T16:51:45 | 2020-01-15T16:51:45 | {} | NONE |
ClickHouse | ClickHouse | 575,001,101 | 8,676 | alex-zaitsev | I like that better. But to make it totally usable we need to make sure that the same replicated cluster can be used both for replicated and distributed tables. See this https://github.com/ClickHouse/ClickHouse/issues/3268 | 2020-01-16T06:15:20 | 2020-01-16T06:15:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 575,007,612 | 8,422 | excitoon | The source of that error is https://github.com/ClickHouse/ClickHouse/blob/1a87da948143724af267adfac6cd0dcb3fe89534/dbms/src/Storages/MergeTree/MergeTreePartsMover.cpp#L224 . Destination directory is being removed during move but it could contain part which is to be deleted but still active. | 2020-01-16T06:40:49 | 2020-01-16T06:40:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 575,039,008 | 8,680 | excitoon | That may be not a solution yet. Possibly part moved from disk A to disk B then from disk B to disk A erasing existing directory, then `DeleteOnDestroy` logic removed existing part. | 2020-01-16T08:25:00 | 2020-01-16T08:25:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 575,072,928 | 8,679 | 865357384 | My use case is actually very simple, which is to delete data first and then insert data, similar to simple ETL work. | 2020-01-16T09:53:40 | 2020-01-16T09:53:40 | {} | NONE |
ClickHouse | ClickHouse | 574,325,397 | 5,912 | alexey-milovidov | Looks like no one needs it. | 2020-01-14T19:07:51 | 2020-01-14T19:07:51 | {} | MEMBER |
ClickHouse | ClickHouse | 574,647,998 | 3,473 | alexey-milovidov | We can support compression for INTO OUTFILE in the same way as for File-related storages (with a switch: `auto`, `none`, `gz`, `br` where auto means "detect by file extension" and it is by default).
The question is where to place this switch: in settings (easy) or in special syntax clause.
I prefer first. (Most of ... | 2020-01-15T12:55:47 | 2020-01-15T12:55:47 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 575,089,321 | 6,256 | YiuRULE | @Akazz Hi Akazz, did you had time to check a bit about this PR ? does I need to change something on my side ? | 2020-01-16T10:34:17 | 2020-01-16T10:34:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 575,089,511 | 8,511 | YiuRULE | @Akazz Hi Akazz, did you had time to check a bit about this PR ? does the solution seem ok ? | 2020-01-16T10:34:45 | 2020-01-16T10:34:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 575,227,911 | 8,511 | Akazz | > @Akazz Hi Akazz, did you had time to check a bit about this PR ? does the solution seem ok ?
I've identified the problem and currently am working on a solution. Please wait for a soon update, I'll post all of the details | 2020-01-16T16:18:36 | 2020-01-16T16:18:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 575,264,393 | 8,688 | alexey-milovidov | At least Stable functions should be included in documentation and I think that changing the type for intermediate calculations to be always Float64 is also reasonable (it will make serialization of aggregation state backwards incompatible but it still ok to do). | 2020-01-16T17:40:44 | 2020-01-16T17:40:44 | {} | MEMBER |
ClickHouse | ClickHouse | 574,455,597 | 8,634 | cliicy | thanks very much for your explanation. they are very helpful.
| 2020-01-15T01:39:21 | 2020-01-15T01:39:21 | {} | NONE |
ClickHouse | ClickHouse | 574,684,330 | 7,019 | alexey-milovidov | > @alexey-milovidov is this expected?
No. The query should be cancelled if the server can detect that the client is disconnected. | 2020-01-15T14:30:00 | 2020-01-15T14:30:00 | {} | MEMBER |
ClickHouse | ClickHouse | 574,798,724 | 8,674 | orloffv | Can u check my case? https://github.com/ClickHouse/ClickHouse/issues/7799#issuecomment-560025636
```Method createColumn() is not implemented for data type Function((String, String) -> String)```
```
SELECT
arrayJoin AS kv_key
FROM system.query_log
ARRAY JOIN ProfileEvents.Names AS arrayJoin
PREWHERE has(ar... | 2020-01-15T18:49:16 | 2020-01-15T18:49:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 574,808,715 | 7,019 | silviucpp | Problem is related to https://github.com/ClickHouse/ClickHouse/issues/7736 and I described there from where it comes. | 2020-01-15T19:06:52 | 2020-01-15T19:06:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 575,016,323 | 8,680 | excitoon | This fix shall be cherry-picked everywhere when possible. | 2020-01-16T07:13:10 | 2020-01-16T07:13:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 575,104,834 | 8,681 | maskshell | > Is this a distributed table? It seems that distributed tables cannot be updated or deleted ，Can only be inserted
as u can see,
> ENGINE = ReplicatedReplacingMergeTree
not Distributed engine. | 2020-01-16T11:15:00 | 2020-01-16T11:15:00 | {} | NONE |
ClickHouse | ClickHouse | 575,125,644 | 8,637 | 4ertus2 | In progress. Build with release + processors fails:
01062_pm_all_join_with_block_continuation: [ FAIL ] | 2020-01-16T12:15:50 | 2020-01-16T14:35:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 574,761,134 | 777 | Slach | @mnach
you can try https://github.com/Slach/clickhouse-zetcd
but without any guarantee
also zetcd it's a proxy to etcd
so one request to zetcd fire one and more request to etcd
so performance of zetcd is more slow than Zookeeper | 2020-01-15T17:14:26 | 2020-01-15T17:14:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 575,210,871 | 8,686 | Taxel | Thanks, that fixed it! | 2020-01-16T15:42:11 | 2020-01-16T15:42:11 | {} | NONE |
ClickHouse | ClickHouse | 575,296,509 | 8,674 | orloffv | @alexey-milovidov i'v created new issue with my problem, because current issue - closed
https://github.com/ClickHouse/ClickHouse/issues/8689 | 2020-01-16T19:02:21 | 2020-01-16T19:02:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 574,335,773 | 8,168 | alexey-milovidov | @akonyaev90 reported this issue but then did not provide any relevant info. | 2020-01-14T19:33:11 | 2020-01-14T19:33:11 | {} | MEMBER |
ClickHouse | ClickHouse | 574,545,156 | 8,559 | maskshell | > Could you please take a look to `ClickHouse build check — 0/14 builds are OK`?
My fault. Already fixed.
Could u pls tell me how to fix the failed of "Style check" ? | 2020-01-15T08:12:28 | 2020-01-15T08:12:28 | {} | NONE |
ClickHouse | ClickHouse | 574,566,108 | 5,601 | Sugaroverdose | @oandrew just to make it clear what i mean when saying 'multiple url's':
Multiple urls are used just to support failover like confluent client does:
in config client gets list of strings(maybe choose first node from list randomly, to spread load) and if chosen schema registry node responded with status 5xx or timed o... | 2020-01-15T09:13:10 | 2020-01-15T09:46:04 | {} | NONE |
ClickHouse | ClickHouse | 574,833,132 | 8,677 | alexey-milovidov | You can provide DCO instead of CLA: https://developercertificate.org/ | 2020-01-15T20:03:48 | 2020-01-15T20:03:48 | {} | MEMBER |
ClickHouse | ClickHouse | 575,036,435 | 8,679 | 865357384 | How can I wait for it to complete this operation? Do I need to delay for the next task? | 2020-01-16T08:18:07 | 2020-01-16T08:18:07 | {} | NONE |
ClickHouse | ClickHouse | 575,040,985 | 8,680 | excitoon | If we don't allow to move part back until `DeleteOnDestroy` happen then this is a correct solution. | 2020-01-16T08:30:30 | 2020-01-16T08:30:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 575,079,543 | 8,674 | alexey-milovidov | > I'll try to add the fix to this PR.
In next PR. | 2020-01-16T10:10:35 | 2020-01-16T10:10:35 | {} | MEMBER |
ClickHouse | ClickHouse | 575,543,259 | 8,635 | bgranvea | Thanks I didn't know about the start without ZK.
I could store somewhere the knowledge that I'm trying to restore a replica from scratch (so need to attach tables instead of create them) but I need some control logic when I execute the SQL script. My point was to have the same SQL statement which can do both (thus ... | 2020-01-17T09:18:58 | 2020-01-17T09:18:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 574,334,432 | 7,302 | alexey-milovidov | I've tested. It works perfectly on Huawei Taishan. | 2020-01-14T19:30:01 | 2020-01-14T19:30:01 | {} | MEMBER |
ClickHouse | ClickHouse | 574,349,294 | 6,762 | vitlibar | https://github.com/ClickHouse/ClickHouse/pull/8292 | 2020-01-14T20:06:01 | 2020-01-14T20:06:01 | {} | MEMBER |
ClickHouse | ClickHouse | 574,360,223 | 8,168 | den-crane | I did a quick test.
It seems works with `and`
```
set allow_experimental_data_skipping_indices=1;
CREATE TABLE mskipi
(
`A` Int64,
`I1` Int64,
`I2` Int64,
INDEX i1i I1 TYPE minmax GRANULARITY 1,
INDEX i2i I2 TYPE minmax GRANULARITY 1
)
ENGINE = MergeTree
ORDER BY A;
system stop m... | 2020-01-14T20:32:41 | 2020-01-14T20:33:08 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 574,461,900 | 8,636 | den-crane | >2020.01.13 23:43:56.376109 [ 969 ] {} BaseDaemon:
>2020.01.13 23:43:56.376119 [ 970 ] {} BaseDaemon:
It's kinda unusual. Two CH servers process? | 2020-01-15T02:08:10 | 2020-01-15T02:08:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 574,557,184 | 7,646 | ivanprostran | Settings:
```
FROM system.settings
WHERE name = 'background_schedule_pool_size'
┌─name──────────────────────────┬─value─┬─changed─┬─description──────────────────────────────────────────────────────────────────────────────────────────────┐
│ background_schedule_pool_size │ 16 │ 0 │ Number of threads pe... | 2020-01-15T08:48:26 | 2020-01-15T08:49:04 | {} | NONE |
ClickHouse | ClickHouse | 574,596,424 | 5,371 | stavrolia | Current status:
- support for wildcards have been implemented;
- numeric ranges work without leading zeros like {12..123} ({012..123} doesn't work at the moment);
- virtual columns _file and _path are implemented in #8489. | 2020-01-15T10:26:53 | 2020-01-15T10:27:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 574,800,178 | 8,674 | alexey-milovidov | @orloffv It isn't fixed :( I'll try to add the fix to this PR. | 2020-01-15T18:53:10 | 2020-01-15T18:53:10 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 575,002,675 | 8,422 | excitoon | This error looks like it was there for a long time. | 2020-01-16T06:21:50 | 2020-01-16T06:21:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 575,016,521 | 8,422 | excitoon | That fixed in #8680 . | 2020-01-16T07:13:48 | 2020-01-16T07:13:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 575,091,474 | 8,681 | 865357384 | Is this a distributed table? It seems that distributed tables cannot be updated or deleted ，Can only be inserted | 2020-01-16T10:40:05 | 2020-01-16T10:40:05 | {} | NONE |
ClickHouse | ClickHouse | 575,379,278 | 7,646 | filimonov | @ivanprostran how many Kafka tables you have? Did it solve the issue with timeouts? | 2020-01-16T22:36:38 | 2020-01-16T22:36:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 574,350,278 | 8,248 | alexey-milovidov | @alex-zaitsev said that 19.16 is Ok and backport is not worth the efforts. | 2020-01-14T20:08:31 | 2020-01-14T20:08:31 | {} | MEMBER |
ClickHouse | ClickHouse | 575,432,545 | 8,656 | leoones | @den-crane | 2020-01-17T02:09:29 | 2020-01-17T02:09:29 | {} | NONE |
ClickHouse | ClickHouse | 574,326,562 | 6,440 | alexey-milovidov | Looks easy, maybe reassign to @millb or @achulkov2 ? | 2020-01-14T19:10:50 | 2020-01-14T19:10:50 | {} | MEMBER |
ClickHouse | ClickHouse | 574,631,458 | 8,655 | alex-zaitsev | Right side does not respect the FINAL, I guess. | 2020-01-15T12:06:56 | 2020-01-15T12:06:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 574,717,597 | 8,668 | se-teryoshkin | > Какой в этом смысл? Да КХ не отдает память, это сделано специально чтобы не тратить время на аллокацию.
Из-за этого за ним иногда приходит oom-killer | 2020-01-15T15:41:00 | 2020-01-15T15:41:00 | {} | NONE |
ClickHouse | ClickHouse | 575,033,971 | 8,679 | MeteHanC | You might have executed the 'insert' query before the deletion took place. I assume you are inserting from the same table that you want to delete data from. You should wait for the mutation to be completed. As I said, it does not work real time, it is rather an asynchronous job. | 2020-01-16T08:11:26 | 2020-01-16T08:11:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 575,122,209 | 8,684 | m1khal3v | On version 19.17.4, in the first case, 0 rows are returned. And I forgot to add that both tables are distributed. If you make a request to the table on the node, then the response is not empty, but still not the same
| 2020-01-16T12:05:31 | 2020-01-16T12:27:24 | {} | NONE |
ClickHouse | ClickHouse | 575,207,103 | 8,686 | den-crane | On install CH asked for a default password for default user and placed it to /etc/clickhouse-server/users.d/default-password.xml
You can change this password or remove this file to empty password. | 2020-01-16T15:34:21 | 2020-01-16T15:34:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 574,673,036 | 8,655 | filimonov | I think the most natural is to parse FINAL as a part of table name.
The query in original request was like:
```sql
select cols from from kafka_table inner join customer_data final using (customer_id, tenant_id) where tenant_id != 0;
```
So if we have comma join i would expect repeating final for every table wh... | 2020-01-15T14:03:29 | 2020-01-15T14:03:29 | {
"+1": 8
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 574,716,981 | 8,668 | den-crane | >Возможно ли как-то влиять с помощью настроек на кеширование в аллокаторе?
Какой в этом смысл? Да КХ не отдает память, это сделано специально чтобы не тратить время на аллокацию. | 2020-01-15T15:39:42 | 2020-01-15T15:39:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 574,720,033 | 8,668 | den-crane | > > Какой в этом смысл? Да КХ не отдает память, это сделано специально чтобы не тратить время на аллокацию.
>
> Из-за этого за ним иногда приходит oom-killer
Не вижу связи. oom-killer приходит потому что КХ аллоцирует память, а не потому что он не деаллоцирует.
Лимитируйте максимальное потребление в районе 75% ... | 2020-01-15T15:46:14 | 2020-01-15T15:46:14 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 574,721,155 | 8,422 | excitoon |
[clickhouse-server.log](https://github.com/ClickHouse/ClickHouse/files/4069617/clickhouse-server.log)
```
self = <multiprocessing.pool.ApplyResult object at 0x7fcedb36fed0>, timeout = 60
def get(self, timeout=None):
self.wait(timeout)
if not self._ready:
raise TimeoutError
... | 2020-01-15T15:48:39 | 2020-01-16T06:21:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 575,016,622 | 8,680 | excitoon | See https://github.com/ClickHouse/ClickHouse/pull/8422#issuecomment-574721155 | 2020-01-16T07:14:10 | 2020-01-16T07:14:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 575,039,736 | 8,679 | MeteHanC | Delaying is not wise since you don't have any clue about how long it will take. You can wait until your task is completed programmatically. Though, I am struggling to understand your purpose, there might be better ways but I need to understand your use case first. | 2020-01-16T08:26:51 | 2020-01-16T08:26:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 575,428,141 | 8,022 | ghost | @alexey-milovidov @alesapin
Hello.
Can you help me to verify this PR?
I am waiting for about one month.
If these features aren't aligned with ClickHouse, wouldn't be worth approve just only `local query cache`?
I hope you look these features positively.
Thank you. | 2020-01-17T01:48:18 | 2020-01-17T01:48:18 | {} | NONE |
ClickHouse | ClickHouse | 574,336,251 | 8,448 | alexey-milovidov | @millb It's a good thing to do. | 2020-01-14T19:34:20 | 2020-01-14T19:34:20 | {} | MEMBER |
ClickHouse | ClickHouse | 574,609,415 | 8,416 | abyss7 | > std::unique_ptr and forward declaration works fine for pImpl.
Yes `void *` is a crooked implementation - will do in a better way. | 2020-01-15T11:00:46 | 2020-01-15T11:00:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 574,706,383 | 8,670 | vzakaznikov | Same issue if an MV is created as
```sql
CREATE MATERIALIZED VIEW mv
ENGINE = Memory() AS
SELECT sum(value) / 10
FROM
(
SELECT
ProfileEvent_DiskReadElapsedMicroseconds AS value,
event_time
FROM system.metric_log
ORDER BY event_time DESC
LIMIT 10
)
``` | 2020-01-15T15:17:14 | 2020-01-15T15:17:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 575,077,139 | 8,646 | kushagra391 | Thanks @den-crane and @alexey-milovidov !
Would check out `clickhouse-local`
I was looking more of a java / python clickhouse library that would generate native format data and use it to build final clickhouse data with spark, and then bulk upload it. It might still be an expensive operation to do with clickhouse.... | 2020-01-16T10:04:14 | 2020-01-16T10:11:36 | {} | NONE |
ClickHouse | ClickHouse | 574,331,450 | 6,762 | alexey-milovidov | Fixed (?). No longer reproduces in CI. | 2020-01-14T19:23:09 | 2020-01-14T19:23:09 | {} | MEMBER |
ClickHouse | ClickHouse | 574,332,122 | 6,496 | alexey-milovidov | @ValBaturin | 2020-01-14T19:24:42 | 2020-01-14T19:24:42 | {} | MEMBER |
ClickHouse | ClickHouse | 574,581,849 | 8,592 | edonin | Thanks for the answer. For me, changing the max_thread to 1 is working. I am able to run my big query on a very small machine. | 2020-01-15T09:51:58 | 2020-01-15T09:51:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 574,622,250 | 7,948 | hagen1778 | I'm hitting the same issue. Is there any updates? | 2020-01-15T11:38:24 | 2020-01-15T11:38:24 | {} | NONE |
ClickHouse | ClickHouse | 574,738,441 | 777 | mnacharov | What's the status? | 2020-01-15T16:26:07 | 2020-01-15T16:26:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 574,938,613 | 8,676 | zhang2014 | In fact, this logic has been around for a long time in my repo(https://github.com/zhang2014/ClickHouse/tree/add_default_on_cluster), but it's a little different(If it is acceptable , I can immediately provide a pull request ：） ):
first:
```
<default_on_cluster_name>cluster_name</default_on_cluster_name>
```
th... | 2020-01-16T01:32:14 | 2020-01-16T01:39:02 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 575,077,716 | 8,679 | MeteHanC | You might want to take a look at [CollapsingMergeTree](https://clickhouse.yandex/docs/en/operations/table_engines/collapsingmergetree/).
Basically, you can add a Sign column to your table and rather than actually deleting the data you can insert the exact same rows with -1 Sign. And execute your query like this ;
... | 2020-01-16T10:05:43 | 2020-01-16T10:06:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 575,078,216 | 8,679 | 865357384 | OK，Thank you very much. | 2020-01-16T10:07:00 | 2020-01-16T10:07:00 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 575,161,463 | 8,684 | 4ertus2 | It's very hard to say anything without data. Could you make a reproducable test with sample data? | 2020-01-16T13:54:29 | 2020-01-16T13:54:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 575,173,943 | 8,681 | den-crane | duplicate: https://github.com/ClickHouse/ClickHouse/issues/2712
https://github.com/ClickHouse/ClickHouse/issues/3526
https://github.com/ClickHouse/ClickHouse/issues/8277
| 2020-01-16T14:23:36 | 2020-01-16T14:23:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 574,509,143 | 8,468 | cliicy | I encountered the same error. | 2020-01-15T05:59:22 | 2020-01-15T05:59:22 | {} | NONE |
ClickHouse | ClickHouse | 574,563,654 | 8,658 | excitoon | @onine007 You will need newer version of ClickHouse, 19.19+. | 2020-01-15T09:06:24 | 2020-01-15T09:06:24 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 574,695,682 | 8,667 | alexey-milovidov | @Avogar | 2020-01-15T14:55:17 | 2020-01-15T14:55:17 | {} | MEMBER |
ClickHouse | ClickHouse | 574,695,971 | 1,604 | alexey-milovidov | @Avogar | 2020-01-15T14:55:55 | 2020-01-15T14:55:55 | {} | MEMBER |
ClickHouse | ClickHouse | 574,709,576 | 8,668 | den-crane | КХ может не освобождать и после успешного запроса (кеширование в аллокаторе).
1. наблюдается ли рост если этот запрос с ошибкой повторить многократно?
2. наблюдается ли рост если этот запрос без ошибки (модифицированный или с другими max_memory_usage) повторить многократно?
| 2020-01-15T15:23:54 | 2020-01-15T15:25:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 575,200,080 | 8,567 | JamalSale | Confirm, have the same issue with
```
SELECT version()
┌─version()──┐
│ 19.17.6.36 │
└────────────┘
``` | 2020-01-16T15:19:35 | 2020-01-16T15:19:35 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 575,447,366 | 8,568 | huang--yijian | > You had only rpms from 19.17 in that folder before :) Now you are trying to install both 19.16 and 19.17 at the same time. But you need to install v19.17, right? :)
>
> Please list all rpms which need to be installed explicitly, or use appropriate wildcard.
>
> ```
> rpm -Fvh *19.17.6*.rpm
> ```
>
> or
> ... | 2020-01-17T03:23:24 | 2020-01-17T03:23:24 | {} | NONE |
ClickHouse | ClickHouse | 574,349,600 | 7,119 | alexey-milovidov | @Potya | 2020-01-14T20:06:51 | 2020-01-14T20:06:51 | {} | MEMBER |
ClickHouse | ClickHouse | 574,716,222 | 8,668 | se-teryoshkin | > КХ может не освобождать и после успешного запроса (кеширование в аллокаторе).
>
> 1. наблюдается ли рост если этот запрос с ошибкой повторить многократно?
> 2. наблюдается ли рост если этот запрос без ошибки (модифицированный или с другими max_memory_usage) повторить многократно?
В обоих случаях рост не наблюд... | 2020-01-15T15:38:06 | 2020-01-15T15:38:06 | {} | NONE |
ClickHouse | ClickHouse | 574,769,140 | 777 | alexey-milovidov | This task is in progress, it is assigned to external contributor @levysh
Currently it depends on https://github.com/ClickHouse/ClickHouse/pull/8435 | 2020-01-15T17:33:25 | 2020-01-15T17:33:25 | {} | MEMBER |
ClickHouse | ClickHouse | 574,768,476 | 777 | alexey-milovidov | @Slach zetcd doesn't work and that's guaranteed. | 2020-01-15T17:31:49 | 2020-01-15T18:26:09 | {} | MEMBER |
ClickHouse | ClickHouse | 575,180,753 | 8,060 | filimonov | Duplicate #6556 | 2020-01-16T14:38:39 | 2020-01-16T14:38:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 575,382,565 | 8,571 | filimonov | > AvroConfluent only caches schemas per an instance of an InputFormat. This means SchemaRegistry will be queried each time a batch from Kafka is processed.
That sounds suboptimal. Especially for Kafka - it would be stupid to refetch schema for each block of data. Did you consider storing cached schemas somewhere els... | 2020-01-16T22:46:17 | 2020-01-16T22:46:17 | {
"+1": 4
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 575,528,430 | 7,646 | ivanprostran | Hi,
There are 5 tables.
Input topics have 10 partitions.
Each node consumes from a single partition (10 nodes 5+5 repl.factor 1)
Unfortunately the errors are still present (possible reduced a little bit..)
```
2020.01.16 15:23:27.393222 [ 69 ] {} <Error> void DB::StorageKafka::threadFunc(): std::exception. ... | 2020-01-17T08:33:15 | 2020-01-17T08:33:15 | {} | NONE |
ClickHouse | ClickHouse | 574,325,946 | 6,020 | alexey-milovidov | Looks not relevant anymore. | 2020-01-14T19:09:13 | 2020-01-14T19:09:13 | {} | MEMBER |
ClickHouse | ClickHouse | 574,566,853 | 8,636 | alexey-milovidov | @den-crane It's Ok, we output messages from different segfaults in different threads (in recent versions). | 2020-01-15T09:14:52 | 2020-01-15T09:14:52 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 574,630,685 | 8,655 | 4ertus2 | Could you attach a sample and expected result, please? It's not clear for which of joining tables FINAL should be applyed. | 2020-01-15T12:04:37 | 2020-01-15T12:04:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 574,635,150 | 8,655 | 4ertus2 | I.e. we have multiple or comma JOIN with FINAL. Which table we should FINAL to apply? What if the query is big? It's very easy to make a mistake changing tables order.
I think the solution is to make special syntax: `FINAL table1, table2` and forbid FINAL without table specification for JOINs. Even then we could hav... | 2020-01-15T12:18:12 | 2020-01-15T12:23:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 574,788,627 | 8,674 | alexey-milovidov | @4ertus2 has the hypothesis that the mechanics is something like that:
- we create prepared set (DataTypeSet) for IN;
- condition with set is pushed down to subquery;
- we analyze data types in subquery, but it's unexpected that one of the data types is Set. | 2020-01-15T18:23:28 | 2020-01-15T18:23:28 | {} | MEMBER |
ClickHouse | ClickHouse | 575,019,062 | 8,679 | MeteHanC | If I understood you right, alter table mutations are executed asynchronously. So when you ran the mutation query, you just scheduled the mutation job. Data is not updated/deleted right away. You can check the mutation status by using 'system.mutations' table. | 2020-01-16T07:23:07 | 2020-01-16T07:24:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 575,181,324 | 6,556 | filimonov | Workaround: create table with Engine=MySQL where that field will be listed as String instead of FixedString. | 2020-01-16T14:39:49 | 2020-01-16T14:39:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 590,182,430 | 6,168 | YiuRULE | @alexey-milovidov could I have some details of the bug and in which condition ? will try to apply a fix | 2020-02-24T06:24:19 | 2020-02-24T06:24:19 | {} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.