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 | 800,352,793 | 21,794 | akuzm | > Any ideas how to handle it?
No way, probably. Hash join uses binary comparison for keys, so it requires a supertype. Logical comparison with `equals` can work for different types w/o conversion, so a merge join would work. By the way, does it work if you enable experimental bigint types? | 2021-03-16T15:16:39 | 2021-03-16T15:16:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 800,548,664 | 21,782 | den-crane | блин, вы думаете никто не в курсе что-ли?
388-й раз никто это не будет обсуждать.
| 2021-03-16T19:36:04 | 2021-03-16T19:36:04 | {
"-1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 800,779,535 | 21,774 | filimonov | Do you copy & paste your long query in clickhouse-client interactive session? It may be for example some issue with tabs.
Workarounds:
- you can pipe in the file
- use queries-file command line argument.
- remove newlines and tabs from the query before pasting it into interactive session | 2021-03-17T04:12:46 | 2021-03-17T04:12:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 799,876,135 | 21,747 | phamtai97 | Ah, I try to perform SQL:
```sql
SELECT field4, COUNT(*) FROM test WHERE time between 1615783324200 and 1615783324200 group by field4;
```
The result has 4 records. | 2021-03-16T01:34:27 | 2021-03-16T01:34:27 | {} | NONE |
ClickHouse | ClickHouse | 799,982,058 | 16,541 | filimonov | Wondering why 'Not for changelog'... | 2021-03-16T06:10:48 | 2021-03-16T06:10:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 799,995,135 | 21,613 | dmalkr | Yes, you can implement product of elements via `arrayFold`: `SELECT arrayFold(x,acc -> x*acc, [1,2,3,4,5], toInt64(1));`
PR #21589 (`arrayFold` function) on review now. | 2021-03-16T06:36:38 | 2021-03-16T06:36:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 800,035,939 | 21,080 | taylor12805 | > Try:
>
> ```
> CREATE TABLE default.event_table
> (
> `id` String,
> `extended_field` Array(Nullable(String))
> ) ENGINE = MergeTree()
> ORDER BY tuple()
> ```
>
> How the task is going? Any progress? Do you need any help?
I think i missed this scenario in the implementation, i will try to fix... | 2021-03-16T07:49:42 | 2021-03-16T07:49:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 800,056,251 | 20,477 | azat | >I am confused
Sorry, let me try to summarize what I want to solve, I want to add an ability to use separate resource tracking/limits/logging/process for:
- distributed queries with prefer_localhost_replica=1
- queries for flushing the data from StorageBuffer/StorageKafka/e.t.c.
- PushingToViewsBlockOutputStream
... | 2021-03-16T08:21:58 | 2021-03-16T08:21:58 | {} | MEMBER |
ClickHouse | ClickHouse | 800,087,598 | 21,529 | KochetovNicolai | > Any clue? Links are under nda.
Also timeout issues. I have increased timeout for msan. | 2021-03-16T09:11:33 | 2021-03-16T09:11:33 | {} | MEMBER |
ClickHouse | ClickHouse | 800,104,042 | 9,404 | alexey-milovidov | @KrishnaPG I also have an idea that we can make a LUT of 16777216 second intervals (it is about 194 days) and then assume that there is at most one time transition during this interval (daylight saving time or global change). In the LUT cell we will record an offset to the beginning of the year from beginning of the in... | 2021-03-16T09:35:59 | 2021-03-16T09:38:53 | {} | MEMBER |
ClickHouse | ClickHouse | 800,148,280 | 21,624 | alexey-milovidov | @azat FYI another idea is to make std::terminate in debug and logging in release. | 2021-03-16T10:39:23 | 2021-03-16T10:39:23 | {} | MEMBER |
ClickHouse | ClickHouse | 799,564,669 | 21,376 | Jokser | @pumelo
You can try to set 'max_connections' options for DiskS3 to a big number like 10000.
Probably it should help with S3 hanging in versions 21.2+ | 2021-03-15T16:36:55 | 2021-03-16T10:54:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 800,184,768 | 21,630 | nikitamikhaylov | AST Fuzzer failure doesn't reproduce | 2021-03-16T11:37:24 | 2021-03-16T11:37:24 | {} | MEMBER |
ClickHouse | ClickHouse | 800,204,233 | 21,759 | alexey-milovidov | > test_materialize_mysql_database
Flaky test, need help from @zhang2014 | 2021-03-16T12:06:15 | 2021-03-16T12:06:15 | {} | MEMBER |
ClickHouse | ClickHouse | 800,546,379 | 21,569 | den-crane | @lehasm I have a concern about 'some cases', because it's like 99.999% cases. | 2021-03-16T19:32:17 | 2021-03-16T19:32:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 799,812,269 | 21,588 | filimonov | @KochetovNicolai @azat any ideas? | 2021-03-15T23:03:25 | 2021-03-15T23:03:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 799,856,259 | 21,747 | phamtai97 | Version Clickhouse: 20.1.4.14. And the above error I see happens randomly. | 2021-03-16T00:39:32 | 2021-03-16T00:39:32 | {} | NONE |
ClickHouse | ClickHouse | 800,221,465 | 18,583 | alexey-milovidov | ```
CREATE TABLE memcpy_benchmark
(
`machine` LowCardinality(String),
`name` LowCardinality(String),
`size` UInt64,
`iterations` UInt64,
`threads` UInt16,
`generator` UInt8,
`memcpy` UInt8,
`elapsed` UInt64
)
ENGINE = MergeTree
ORDER BY (machine, size, threads, generator, me... | 2021-03-16T12:34:51 | 2021-03-16T12:34:51 | {} | MEMBER |
ClickHouse | ClickHouse | 799,652,402 | 21,757 | asadkovich | @kssenii,
Yes, I'm using name of exisiting queue for ```rabbitmq_queue_base``` setting.
All settings:
```SQL
rabbitmq_host_port = '...',
rabbitmq_exchange_name = 'exchange-name',
rabbitmq_queue_base = 'logs-clickhouse',
rabbitmq_exchange_type = 'topic',
rabbitmq_routing_key_list = 'logs',
rabbitmq_format = ... | 2021-03-15T18:31:50 | 2021-03-16T12:54:22 | {} | NONE |
ClickHouse | ClickHouse | 800,406,292 | 16,669 | alexey-milovidov | Let's make MySQL connection timeout controlled by query-level settings. | 2021-03-16T16:21:29 | 2021-03-16T16:21:29 | {} | MEMBER |
ClickHouse | ClickHouse | 800,409,990 | 16,669 | alexey-milovidov | Also if there are no multiple replicas specified, we can make connection timeout higher (up to max_execution_time). | 2021-03-16T16:26:34 | 2021-03-16T16:26:34 | {} | MEMBER |
ClickHouse | ClickHouse | 799,968,371 | 21,763 | filimonov | Sound a bit strange. From RBAC you usually expect dererminism w/o too many 'automatic' changes.
Can you explain why exactly is it needed? Playground currently uses single user. | 2021-03-16T05:41:14 | 2021-03-16T05:41:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 800,108,047 | 9,404 | alexey-milovidov | Another idea is to place a pointer to virtual table in the LUT cells. Basically the idea is to make LUTCell a polymorphic class and allocate it in LUT with placement new. It will make simple cases very cheap in cost of one indirect function call. | 2021-03-16T09:42:31 | 2021-03-16T09:42:50 | {} | MEMBER |
ClickHouse | ClickHouse | 800,244,124 | 21,767 | den-crane | here 21.3 https://clickhouse-builds.s3.yandex.net/21495/f0c97a2b9643fda8cf32dec46248a8d098fb4af6/clickhouse_build_check/report.html | 2021-03-16T13:09:07 | 2021-03-16T13:09:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 800,256,625 | 8,687 | alexey-milovidov | `ARRAY JOIN x, y` - zip (working)
`ARRAY JOIN x ARRAY JOIN y` - product (the missing feature, only available with `arrayJoin` function) | 2021-03-16T13:27:00 | 2021-03-16T13:27:00 | {} | MEMBER |
ClickHouse | ClickHouse | 800,440,783 | 16,798 | zhicwu | @litaoyx, thanks for sharing your solution. I read through your blog and was able to add RoaringBitmap support. One thing I want to clarify is that we don't need a base64 encoded string, since we can insert serialized value into AggregateFunction column directly. I guess the error you met earlier is related to the `Tab... | 2021-03-16T17:00:49 | 2021-03-16T17:00:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 800,456,467 | 21,782 | den-crane | @proggga I took it from https://clickhouse.tech/docs/en/sql-reference/functions/date-time-functions/#tostartofmonth
And yes, I think I will add the same snippet to toDate / toDateTime .... functions.
```
Attention
The behavior of parsing incorrect dates is implementation specific. ClickHouse may return zero dat... | 2021-03-16T17:21:02 | 2021-03-16T17:21:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 799,748,372 | 21,438 | filimonov | > How can I make description check happy?
Just edit PR description to leave single category, and fill in changelog entry. | 2021-03-15T21:00:06 | 2021-03-15T21:00:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 799,927,702 | 21,747 | phamtai97 | I will close this issue. Thank you for your support. | 2021-03-16T04:01:07 | 2021-03-16T04:01:07 | {} | NONE |
ClickHouse | ClickHouse | 799,987,199 | 21,765 | aopetrov86 | @filimonov, thanks. Not reproduced in version 20.11.6 | 2021-03-16T06:21:06 | 2021-03-16T06:21:06 | {} | NONE |
ClickHouse | ClickHouse | 800,046,787 | 21,701 | nshah14285 | May, I know, when this fix will release? | 2021-03-16T08:06:30 | 2021-03-16T08:06:30 | {} | NONE |
ClickHouse | ClickHouse | 800,193,091 | 8,687 | akuzm | The title is always confusing me. Do I get it right that we are talking about per-row cartesian product of arrays?
Now we can either zip (`ARRAY JOIN x, y`) or do a cartesian product for all rows (`arrayJoin(x), arrayJoin(y)`). | 2021-03-16T11:52:03 | 2021-03-16T11:52:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 800,230,104 | 21,589 | akuzm | So basically it's `arrayReduce`, but it takes a lambda instead of an aggregate function, right? The interface of normal functions is not really suitable for this, that's why you have to do the `cut()`s. Maybe it's OK for the first try. The performance is going to be bad, but I can't see an easy way around this now.
... | 2021-03-16T12:48:58 | 2021-03-16T12:48:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 800,449,748 | 21,782 | proggga | > my quote is a copy/paste from the documentation.
>
> I can add the same sentence to the documentation for toDate function if you like.
Your quote is copy/paste from documentation, but that sentence not in the documentation and you can add this to docs?
Am I right? | 2021-03-16T17:12:18 | 2021-03-16T17:12:18 | {} | NONE |
ClickHouse | ClickHouse | 800,522,945 | 21,782 | hrissan | Can clickhouse at least do it consistently?
When using toDate in selects, it is very hard to notice errors, when garbage is accepted as valid dates.
Behavior is implemenation-specific, true, but implementation should strive to make it as useful as possible.
Parsing "nice-06-10" as "1970-01-01", but throwing er... | 2021-03-16T18:55:39 | 2021-03-16T18:55:39 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 800,443,250 | 20,244 | Felixoid | Another thing I can bring for debugging.
I use [POST.txt](https://github.com/ClickHouse/ClickHouse/files/6150866/POST.txt) file to make requests to ClickHouse servers over the network:
`cat POST | nc some.long.host.name 8123`
And as results mostly [200.txt](https://github.com/ClickHouse/ClickHouse/files/615088... | 2021-03-16T17:03:46 | 2021-03-16T19:30:51 | {} | MEMBER |
ClickHouse | ClickHouse | 799,839,394 | 21,588 | den-crane | >SELECT ORDER BY timestamp DESC LIMIT 100 FORMAT JSONCompact
@SaltTan ? also has HTTP `SELECT ORDER BY ... LIMIT ... FORMAT TabSeparatedWithNamesAndTypes`
Maybe it's related to LIMIT | 2021-03-15T23:56:24 | 2021-03-15T23:56:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 800,144,209 | 20,179 | pumelo | This one caught me too. It was not pointed out but let me add: **This leads to a corrupted database and lost data!**
If using `alter table xxx.yyyy move partition '2020-10' to volume 's3'` the files uploaded to s3 are incomplete (due to memory error during upload), but the part is removed from local disk.
As @alex-... | 2021-03-16T10:32:41 | 2021-03-16T10:32:41 | {} | NONE |
ClickHouse | ClickHouse | 800,383,531 | 21,794 | UnamedRus | > For query SELECT a FROM t1 FULL JOIN t2 USING (a) we will have column that contains all values conatenated from t1.a and t2.a, so a in result table have to hold all values
~~CH would say a is ambiguous column, so you actually should have t1.a and t2.a colums with Int64 and UInt64 type.~~
I am wrong.
```
W... | 2021-03-16T15:53:20 | 2021-03-16T16:03:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 799,805,375 | 21,239 | filimonov | There are too many different cases there, and quite a lot of invariants would be broken.
Clickhouse is the guy from big data world.
Imagine you need to sync the dictionary between 1000 or 10000 nodes, and the dictionary size tends to grow to gigabytes. Some old data need to be cleaned, etc. Some replicas may go ... | 2021-03-15T22:47:03 | 2021-03-15T22:48:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 799,807,899 | 21,588 | SaltTan | It's not a duplicate. There was a crash before, now it's just an exception.
I get it sometimes on 21.1.6 (which I installed to get rid of the crash).
Here is an example:
```
type | event_time_microseconds | result_rows | exception_code | exception ... | 2021-03-15T22:53:29 | 2021-03-15T22:53:29 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 800,082,550 | 21,194 | vdimir | Ref PR https://github.com/ClickHouse/ClickHouse/pull/21624 https://github.com/ClickHouse/ClickHouse/pull/21325 | 2021-03-16T09:03:38 | 2021-03-16T09:03:38 | {} | MEMBER |
ClickHouse | ClickHouse | 800,150,107 | 21,701 | nshah14285 | @kssenii
Thank You. | 2021-03-16T10:42:26 | 2021-03-16T10:42:26 | {} | NONE |
ClickHouse | ClickHouse | 800,160,709 | 21,438 | ilejn | > > How can I make description check happy?
>
> Just edit PR description to leave single category, and fill in changelog entry.
Done.
I think that we should treat this issue as a bugfix. | 2021-03-16T10:59:29 | 2021-03-16T10:59:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 800,363,729 | 21,794 | akuzm | > So we actually don't care for any Int64 keys < 0, because we already know that we can't find corresponding UInt64 key for them.
> In that case can we safely convert it to UInt64 then?
You're right, we don't need a supertype, but only a type that can hold any common value of the key types. For UInt64 and Int64, th... | 2021-03-16T15:29:20 | 2021-03-16T15:30:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 800,535,018 | 21,569 | lehasm | > In limited circumstances, this can make a query faster by executing it on more servers. ... There are cases where performance will not improve or even worsen:
It changed to _A query may be processed faster if it is executed on several servers in parallel. But the query performance may degrade in some cases:_ in _Add... | 2021-03-16T19:13:51 | 2021-03-16T19:13:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 799,773,080 | 21,588 | filimonov | Duplicate of #21167 ? | 2021-03-15T21:44:34 | 2021-03-15T21:44:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 799,775,327 | 20,218 | filimonov | @gersmann any news here? | 2021-03-15T21:49:23 | 2021-03-15T21:49:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 799,857,523 | 21,747 | den-crane | 20.1 is out of support.
Check/reproduce the issue against 21.3. | 2021-03-16T00:42:36 | 2021-03-16T00:42:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 800,019,298 | 20,387 | alesapin | UB reported https://github.com/ClickHouse/ClickHouse/issues/21769 | 2021-03-16T07:20:31 | 2021-03-16T07:20:31 | {} | MEMBER |
ClickHouse | ClickHouse | 800,137,143 | 21,701 | nshah14285 | @kssenii
so when it will be release on stable version? As I am using clickhouse stable version. | 2021-03-16T10:23:53 | 2021-03-16T10:23:53 | {} | NONE |
ClickHouse | ClickHouse | 800,250,613 | 21,569 | den-crane | I see it removes these sentences
**In limited circumstances, this can make a query faster by executing it on more servers. ... There are cases where performance will not improve or even worsen:**
It's actually quite important. max_parallel_replicas is mostly unusable and in 99% cases makes performance worse. | 2021-03-16T13:18:34 | 2021-03-16T13:18:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 800,370,608 | 21,781 | akuzm | GitHub has a reactions feature for expressing your interest in the issue, please use it. Posting ten different comments with a dot is harmful, because it hides the discussion and generates meaningless notifications. I have temporarily locked the discussion on this issue. | 2021-03-16T15:37:02 | 2021-03-16T15:37:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 800,560,120 | 21,701 | vialorn | > it is not a bug fix, so it will not be backported to previos versions
Do I get it right - the 21.3 LTS release (mentioned at the online-meetup a few days ago) won't get the fix? | 2021-03-16T19:55:01 | 2021-03-16T19:55:01 | {} | NONE |
ClickHouse | ClickHouse | 799,668,771 | 21,757 | kssenii | Actually, in the list of usages of `rabbitmq_queue_base` settings it is `not` written that it can be used as a queue defined by user, from which to receive messages. It was anyway supposed that publishing is done only to `exchange`, not to queue and `exchange and queue are not created by user`, because if settings mak... | 2021-03-15T18:52:19 | 2021-03-15T21:09:52 | {} | MEMBER |
ClickHouse | ClickHouse | 799,981,712 | 21,765 | filimonov | Most probably fixed in https://github.com/ClickHouse/ClickHouse/pull/16541 (20.11+) | 2021-03-16T06:10:13 | 2021-03-16T06:10:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 799,982,648 | 21,569 | lehasm | There are some conflicts. I will try to resolve them a bit later. | 2021-03-16T06:11:59 | 2021-03-16T06:11:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 800,100,113 | 9,404 | alexey-milovidov | Now performance is almost break even. | 2021-03-16T09:30:17 | 2021-03-16T09:30:17 | {} | MEMBER |
ClickHouse | ClickHouse | 800,228,694 | 21,757 | asadkovich | @kssenii,
Yep, publishing starts before table creation.
I have two similar queues binded to general exchange. One for clickhouse service and another for different service. The point is I'd like to avoid data inconsistency between both services if something goes wrong with clickhouse and I restart it. So i need to rea... | 2021-03-16T12:46:48 | 2021-03-16T13:02:03 | {} | NONE |
ClickHouse | ClickHouse | 800,306,545 | 21,268 | olgarev | > @olgarev, PR is fixed and merged. Can you, please, merge the changes?
@kssenii Great! Will do my best )
> Also, there are two PR's which also need PostgreSQL doc :)
> #21711 (additional parameter `table_schema`)
> #21710 (same as now supported for mysql)
If you tag them with "docs-alert", we will pick them... | 2021-03-16T14:28:20 | 2021-03-16T14:28:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 800,347,687 | 21,794 | UnamedRus | > Any ideas how to handle it?
Can we just convert minus to plus?
```
SELECT *
FROM
(
SELECT number
FROM numbers(18446744073709551605, 10)
) AS a
,
(
SELECT number
FROM numbers(18446744073709551605, 10)
) AS b
WHERE (a.number + 1) = b.number
Query id: 15a48d3d-a6c2-4c15-b386-f1bf26... | 2021-03-16T15:10:49 | 2021-03-16T15:10:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 800,379,110 | 21,794 | vdimir | For query `SELECT a FROM t1 FULL JOIN t2 USING (a)` we will have column that contains all values conatenated from `t1.a` and `t2.a`, so `a` in result table have to hold all values | 2021-03-16T15:47:56 | 2021-03-16T15:47:56 | {} | MEMBER |
ClickHouse | ClickHouse | 799,793,073 | 20,387 | filimonov | https://github.com/ClickHouse/ClickHouse/issues/20304 | 2021-03-15T22:19:42 | 2021-03-15T22:19:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 799,858,688 | 21,747 | phamtai97 | Oh, Currently, we can not upgrade version. So, if version 21.3 still has this error? | 2021-03-16T00:45:11 | 2021-03-16T00:45:11 | {} | NONE |
ClickHouse | ClickHouse | 800,177,761 | 20,179 | Jokser | @pumelo @vmihailenco
Decreasing these options
s3_min_upload_part_size
s3_max_single_part_upload_size
to 32Mb should help with memory consumption if you use CH version 21.1
They should be placed under 'default' profile section in users.xml config file. | 2021-03-16T11:26:10 | 2021-03-16T11:26:10 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 800,729,226 | 21,774 | yangyongyongyong | when i install clickhouse on docker, show this error; but when i install clickhouse in centos,than The constructional sentences pass. so it is docker issue. | 2021-03-17T01:43:28 | 2021-03-17T01:43:28 | {} | NONE |
ClickHouse | ClickHouse | 800,118,158 | 21,701 | kssenii | @nshah14285, till the end of this day will be in master. | 2021-03-16T09:57:53 | 2021-03-16T09:57:53 | {
"rocket": 1
} | MEMBER |
ClickHouse | ClickHouse | 800,236,309 | 21,782 | den-crane | The behavior of parsing incorrect dates is implementation specific. ClickHouse may return zero date, throw an exception or do “natural” overflow. | 2021-03-16T12:58:07 | 2021-03-16T12:58:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 800,341,297 | 21,794 | UnamedRus | Well, it works in regular WHERE condition:
```
│ Filter (WHERE) │
│ Header: number UInt64 │
│ b.number UI... | 2021-03-16T15:06:11 | 2021-03-16T15:06:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 799,910,236 | 21,747 | phamtai97 | I find the issues: https://github.com/ClickHouse/ClickHouse/issues/11634, So the issue is the same as my issue. | 2021-03-16T03:12:36 | 2021-03-16T03:12:36 | {} | NONE |
ClickHouse | ClickHouse | 799,917,967 | 21,588 | azat | Should be fixed by #21643 | 2021-03-16T03:33:17 | 2021-03-16T03:33:17 | {} | MEMBER |
ClickHouse | ClickHouse | 799,481,156 | 21,715 | alexey-milovidov | [result_c5metal.txt](https://github.com/ClickHouse/ClickHouse/files/6142203/result_c5metal.txt) - baremetal, some modern Intel
[result_e5_2650v2.txt](https://github.com/ClickHouse/ClickHouse/files/6142205/result_e5_2650v2.txt) - baremetal, very old Intel
[result_yandex_cascade_lake.txt](https://github.com/ClickHouse/... | 2021-03-15T14:50:03 | 2021-03-16T12:07:14 | {} | MEMBER |
ClickHouse | ClickHouse | 800,283,246 | 8,687 | akuzm | Yes, looks like a workaround with `arrayJoin` is also possible:
```
*2* :) SELECT arrayJoin(arrayZip(person.name, person.surname)), arrayJoin(arrayZip(properties.key, properties.value)) FROM test_multiple_array_join
┌─arrayJoin(arrayZip(person.name, person.surname))─┬─arrayJoin(arrayZip(properties.key, propert... | 2021-03-16T14:01:56 | 2021-03-16T14:01:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 800,330,701 | 21,794 | vdimir | It's expected, because `UInt64` and `Int64` is too big and we will need `Int128` to fit all possible range of data.
Any ideas how to handle it? | 2021-03-16T14:55:45 | 2021-03-16T14:55:45 | {} | MEMBER |
ClickHouse | ClickHouse | 800,358,056 | 21,794 | UnamedRus | > By the way, does it work if you enable experimental bigint types?
No.
> No way, probably. Hash join uses binary comparison for keys, so it requires a supertype.
Ok, but for example we have Int64 and UInt64 as datatypes for keys, and have equals condition on them.
So we actually don't care for any Int64 ke... | 2021-03-16T15:22:33 | 2021-03-16T15:22:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 799,861,090 | 21,747 | den-crane | >So, if version 21.3 still has this error?
I don't know.
I would install a test server with 20.1.4.14. Copy freezed parts from a production. Checked that the issue reproduces. Then I would upgrade a test server to 21.3 and checked there.
| 2021-03-16T00:51:45 | 2021-03-16T00:53:13 | {
"heart": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 799,894,490 | 21,764 | den-crane | 20.8.3.18 is obsolete.
check with 20.8.14.4 | 2021-03-16T02:26:35 | 2021-03-16T02:26:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 800,017,862 | 20,387 | alesapin | Hung check -- not related to changes. | 2021-03-16T07:17:50 | 2021-03-16T07:17:50 | {} | MEMBER |
ClickHouse | ClickHouse | 800,140,595 | 21,701 | kssenii | @nshah14285, it is not a bug fix, so it will not be backported to previos versions. It will be only in master and all futher releases. | 2021-03-16T10:27:24 | 2021-03-16T10:27:24 | {} | MEMBER |
ClickHouse | ClickHouse | 800,332,197 | 14,612 | traceon | https://github.com/ClickHouse/ClickHouse/blob/master/docs/en/operations/external-authenticators/ldap.md | 2021-03-16T14:57:13 | 2021-03-16T14:57:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 800,469,381 | 14,648 | UsedC | We've also encountered this exception with fromUnixTimestamp64Milli. The number was quite smaller, as you can expect with milliseconds, and it took us some time before we figured out what's going on.
This behavior was somewhat unexpected, and we weren't even sure if it was a bug or not until stumbling upon this issu... | 2021-03-16T17:38:14 | 2021-03-16T17:38:14 | {} | NONE |
ClickHouse | ClickHouse | 800,501,892 | 21,724 | azat | >Functional stateless tests flaky check (address) — fail: 1, passed: 169
Failures are unrelated, will be investigated separately. | 2021-03-16T18:24:06 | 2021-03-16T18:24:06 | {} | MEMBER |
ClickHouse | ClickHouse | 799,975,404 | 21,694 | filimonov | I'm afraid that tree may be not enough.
UA parsing is often state machine not a simple tree traversal - it can have negative regular expressions which can reset the state and switch to another branch. | 2021-03-16T05:57:25 | 2021-03-16T05:57:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 800,145,861 | 21,701 | alexey-milovidov | > so when it will be release on stable version?
They are released once a month - the next release 21.4 will be in April. | 2021-03-16T10:35:23 | 2021-03-16T10:35:23 | {} | MEMBER |
ClickHouse | ClickHouse | 800,318,116 | 21,794 | UnamedRus | @vdimir | 2021-03-16T14:41:38 | 2021-03-16T14:41:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 800,651,622 | 21,804 | den-crane | HASHED -- layout supports only UInt64 as PRIMARY KEY
You need to use -- COMPLEX_KEY_HASHED
```sql
CREATE DICTIONARY test
(
id String,
foo String
)
PRIMARY KEY id
SOURCE(FILE(path '/var/lib/clickhouse/user_files/test.json' format 'JSONEachRow'))
LAYOUT(COMPLEX_KEY_HASHED())
LIFETIME(MIN 5 ... | 2021-03-16T22:20:26 | 2021-03-16T23:15:17 | {
"laugh": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 799,861,736 | 21,747 | phamtai97 | oh, thank you so much
| 2021-03-16T00:53:31 | 2021-03-16T00:53:31 | {} | NONE |
ClickHouse | ClickHouse | 800,136,647 | 8,687 | zhuyoucui0909 | Is there any plan that ClickHouse will support multiple array join officially?
The table has multiple Array type columns. The array type columns have different cardinality. We need a way to do multiple array join like follows.
`SELECT *
FROM multiple_arrayType_column_table
ARRAY JOIN arrayTypeColumn1
ARRAY JOIN a... | 2021-03-16T10:23:34 | 2021-03-16T10:23:34 | {} | NONE |
ClickHouse | ClickHouse | 800,206,228 | 18,583 | alexey-milovidov | Raw data for memcpy benchmark:
[result_c5metal.txt](https://github.com/ClickHouse/ClickHouse/files/6142203/result_c5metal.txt) - baremetal, some modern Intel
[result_e5_2650v2.txt](https://github.com/ClickHouse/ClickHouse/files/6142205/result_e5_2650v2.txt) - baremetal, very old Intel
[result_yandex_cascade_lake.t... | 2021-03-16T12:09:52 | 2021-03-16T12:10:36 | {} | MEMBER |
ClickHouse | ClickHouse | 800,272,530 | 21,782 | den-crane | I am not a CH developer. I am a member of CH community and I am doing triage.
My opinion: CH developers decided that it's a responsibility of a client for the sake of performance.
Nothing to discuss here. I closed tons of issues like yours. | 2021-03-16T13:48:30 | 2021-03-16T13:48:36 | {
"-1": 3
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 800,351,544 | 21,794 | UnamedRus | ```
SELECT
18446744073709551605 - 1 AS x,
CAST(x, 'UInt64') AS y
Query id: d77d5d86-d32f-4f87-ade6-6f547e6426c6
┌───x─┬────────────────────y─┐
│ -12 │ 18446744073709551604 │
└─────┴──────────────────────┘
``` | 2021-03-16T15:15:10 | 2021-03-16T15:15:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 800,672,574 | 21,711 | tavplubix | AST fuzzer (UBSan) - #21769
Stress test (address) — Hung check failed - `OPTIMIZE TABLE alter_table FINAL` - looks like #10368
Integration tests (thread) - `test_storage_postgresql` passed | 2021-03-16T23:10:21 | 2021-03-16T23:10:21 | {} | MEMBER |
ClickHouse | ClickHouse | 800,777,727 | 21,529 | azat | >Stress test (thread) — Fatal message in clickhouse-server.log
2021.03.16 23:51:22.493829 [ 295 ] {} <Fatal> Application: Child process was terminated by signal 9 (KILL). If it is not done by 'forcestop' command or manually, the possible cause is OOM Killer (see 'dmesg' and look at the '/var/log/kern.log' for the de... | 2021-03-17T04:06:23 | 2021-03-17T04:06:40 | {} | MEMBER |
ClickHouse | ClickHouse | 799,774,535 | 21,693 | filimonov | Few perf. regressions in 21.2 reported, seems important
https://github.com/ClickHouse/ClickHouse/labels/v21.2-affected | 2021-03-15T21:47:41 | 2021-03-15T21:47:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 799,883,879 | 21,525 | ucasfl | Related #21376. | 2021-03-16T01:57:30 | 2021-03-16T01:57:30 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 800,292,972 | 21,782 | den-crane | my quote is a copy/paste from the documentation.
I can add the same sentence to the documentation for toDate function if you like. | 2021-03-16T14:15:11 | 2021-03-16T14:15:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 792,984,678 | 21,528 | den-crane | You should try.
Clichouse is not a full substitute for elasticsearch. In many use-cases it is impossible to replace elasticsearch with Clickhouse. | 2021-03-08T18:44:27 | 2021-03-08T18:44:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 793,065,704 | 16,438 | azat | >Next steps:
>
>add time of the last error;
>add exception message of the last error;
>add stack trace of the last error;
#21529 (+ `remote` column, introduced in #19290) | 2021-03-08T20:48:03 | 2021-03-08T20:48:03 | {} | MEMBER |
ClickHouse | ClickHouse | 793,641,190 | 21,542 | baibaichen | it looks this issue depends on https://github.com/ClickHouse/ClickHouse/pull/9404 | 2021-03-09T09:52:06 | 2021-03-09T10:00:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 793,742,957 | 21,502 | phamtai97 | Thank you for your answers. | 2021-03-09T11:22:38 | 2021-03-09T11:22:38 | {} | NONE |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.