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 | 581,640,463 | 8,976 | den-crane | >clickhouse server v18.14.13
And it's not supported. Only 20.1.2, 19.17, 19.16 are supported.
>We added it successfully in all the 5 nodes except 1 node.
??? You need to add new columns only at 3 servers (alter is replicated by ReplicatedMergeTree).
>it said the zookeeper and this node were not in sync
... | 2020-02-03T22:02:58 | 2020-02-03T22:03:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 581,171,674 | 8,894 | kreuzerkrieg | Ok, I have to admit two things.
First I've tested it on my branch stemming from the `stable` instead of testing on master.
Second, I have no idea what to do with it. I've found the root cause, it is the `add_cxx_compile_options` in cmake, more specific - TARGET_PROPERTY:LINKER_LANGUAGE expression generator. when it w... | 2020-02-02T20:05:46 | 2020-02-02T20:05:46 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 581,181,248 | 7,478 | alexey-milovidov | No chance this PR will be finished. | 2020-02-02T21:49:44 | 2020-02-02T21:49:44 | {} | MEMBER |
ClickHouse | ClickHouse | 581,383,448 | 2,755 | simPod | @nvartolomei could you point me to the suggestion? Currently I'm experiencing these exceptions
```Code: 234, e.displayText() = DB::Exception: No active replica has part 202002_449608_449608_0 or covering part (version 20.1.2.4 (official build))```
I always go and recreate the table from scratch. | 2020-02-03T12:07:29 | 2020-02-03T12:07:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 581,540,493 | 8,957 | alexey-milovidov | Yes, I asked @alesapin for this reason. | 2020-02-03T18:03:14 | 2020-02-03T18:03:14 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 581,600,211 | 8,985 | alexey-milovidov | Ok but `try_demangle` -> `tryDemangle`. | 2020-02-03T20:21:27 | 2020-02-03T20:21:27 | {} | MEMBER |
ClickHouse | ClickHouse | 580,851,828 | 8,699 | alexey-milovidov | Ok.
@orloffv The reason is that I tried to solve the issue with a simple patch without complete understanding of the underlying code. | 2020-01-31T18:25:10 | 2020-01-31T18:25:10 | {} | MEMBER |
ClickHouse | ClickHouse | 581,092,182 | 8,740 | alexey-milovidov | @godfreyd I've edited the text. | 2020-02-02T02:58:04 | 2020-02-02T02:58:04 | {} | MEMBER |
ClickHouse | ClickHouse | 581,185,152 | 8,350 | alexey-milovidov | I close this PR because it make documentation worse. | 2020-02-02T22:35:11 | 2020-02-02T22:35:11 | {} | MEMBER |
ClickHouse | ClickHouse | 581,307,148 | 5,183 | ghuname | Where we can find more info about this and all other errors, what is causing them and how to get more info about them? | 2020-02-03T09:01:49 | 2020-02-03T09:01:49 | {} | NONE |
ClickHouse | ClickHouse | 581,336,432 | 8,931 | abyss7 | > Ok... but this is too risky and also not very significant change for backport.
> Better don't risk.
>
> PS. And why 19.11? Probably it's by mistake.
It's just a test - to not risk with backporting something by mistake. 19.11 is already closed, so I decided that it's safe for experiments. Anyway I will clean up... | 2020-02-03T10:13:06 | 2020-02-03T10:13:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 580,905,243 | 8,933 | den-crane | I observe a small difference
sumMapWithOverflow
```
set max_threads=1;
create temporary table sumMapWithOverflowTest as
SELECT
arrayMap(x -> (x % 23), range(50)) AS key,
arrayMap(x -> intDiv(number, x + 1), range(50)) AS val
FROM numbers(1000000) // 10000000
SELECT sumMapWithOverflow(key, val) F... | 2020-01-31T20:47:53 | 2020-01-31T20:53:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 581,030,020 | 8,936 | vzakaznikov | Hi @byx313, the Live View table supposed to work as expected after restart. I will recreate the tables in the article and will check what is going on. | 2020-02-01T13:24:09 | 2020-02-01T13:24:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 581,085,666 | 8,951 | alexey-milovidov | Simple example.
```
milovidov-Latitude-E7470 :) SELECT DISTINCT number % 5 FROM system.numbers
SELECT DISTINCT number % 5
FROM system.numbers
┌─modulo(number, 5)─┐
│ 0 │
│ 1 │
│ 2 │
│ 3 │
│ 4 │
└───────────────────┘
↘ Prog... | 2020-02-02T00:57:31 | 2020-02-02T00:57:31 | {} | MEMBER |
ClickHouse | ClickHouse | 581,094,558 | 8,936 | byx313 | > Hi @byx313, I can't reproduce the issue that you are seeing. Here is my test case.
>
> 1. I enable live views and create a live view table.
>
> ```
> ClickHouse client version 20.2.1.1.
> Connecting to localhost:9000 as user default.
> Connected to ClickHouse server version 20.2.1 revision 54432.
>
> :) S... | 2020-02-02T03:30:29 | 2020-02-02T03:30:29 | {} | NONE |
ClickHouse | ClickHouse | 581,177,024 | 8,836 | alexey-milovidov | The test didn't pass. | 2020-02-02T21:03:19 | 2020-02-02T21:03:19 | {} | MEMBER |
ClickHouse | ClickHouse | 581,293,838 | 8,952 | azat | >The usage of MemoryWriteBuffer is (1) overkill, (2) incorrect.
Agree, done | 2020-02-03T08:27:36 | 2020-02-03T08:27:36 | {} | MEMBER |
ClickHouse | ClickHouse | 581,295,802 | 8,022 | ghost | @alexey-milovidov @alesapin @azat
Hello~
I submitted this PR long time ago.
If any one review this PR, I will very appreciate.
Thank you. | 2020-02-03T08:32:50 | 2020-02-03T08:32:50 | {} | NONE |
ClickHouse | ClickHouse | 581,362,262 | 8,964 | doom369 | Error from the log:
```
2020.02.03 11:12:35.185941 [ 50 ] {e56d27a0-9040-45e8-b1c2-8f3246dddf26} <Error> executeQuery: Code: 49, e.displayText() = DB::Exception: RangeReader read 7524 rows, but 7551 expected. (version 20.1.3.7 (official build)) (from 127.0.0.1:33204) (in query: SELECT * FROM reporting_events WHERE ... | 2020-02-03T11:14:19 | 2020-02-03T11:14:19 | {} | NONE |
ClickHouse | ClickHouse | 581,659,434 | 8,962 | filimonov | Motivation is very simple - it IS a SQL keyword :/
https://docs.microsoft.com/en-us/sql/t-sql/data-types/datetime-transact-sql?view=sql-server-ver15
https://dev.mysql.com/doc/refman/8.0/en/datetime.html
Etc.
In general - check `system.data_type_families` table, it has the answer to that question. | 2020-02-03T22:56:35 | 2020-02-03T22:56:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 580,805,661 | 8,727 | alexey-milovidov | That's ridiculous! Will fix. | 2020-01-31T16:29:04 | 2020-01-31T16:29:04 | {} | MEMBER |
ClickHouse | ClickHouse | 580,820,615 | 8,699 | 4ertus2 | The logic is: we append prewhere actions and check thir sample block (and it has no function column), then we append array join actions (and there's no check after it) and then we append where actions (ant throws on its sample block check). So the checks theirself are OK. I don't think we should implement createColumn(... | 2020-01-31T17:04:48 | 2020-01-31T17:04:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 580,936,955 | 8,692 | den-crane | optimize_move_to_prewhere = 0
```
SELECT
status,
count() AS all
FROM temp.test
ARRAY JOIN nested AS nestedJoined
WHERE (status IN
(
SELECT status
FROM temp.test
ARRAY JOIN nested AS nestedJoined
GROUP BY status
ORDER BY count() DESC
LIMIT 10
)) AND (id IN ('1', '2'))
GROU... | 2020-01-31T22:22:07 | 2020-01-31T22:26:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 581,173,225 | 8,755 | alexey-milovidov | It's ok after revert. | 2020-02-02T20:23:03 | 2020-02-02T20:23:03 | {} | MEMBER |
ClickHouse | ClickHouse | 581,407,833 | 4,050 | simPod | @den-crane Hi, can you please elaborate on
> CH designed to use monthly partitions. The only one explanation to your issue is super tiny partitions. Please show the table DDL (at least partition by section).
What does it imply? Is is wrong to have partitioning based on days? I'm asking because I have daily parts... | 2020-02-03T13:14:36 | 2020-02-03T13:14:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 581,598,314 | 8,022 | alexey-milovidov | I have created the PR with review changes: #8986
When it will be merged, your PR will also be merged automatically. | 2020-02-03T20:16:52 | 2020-02-03T20:16:52 | {} | MEMBER |
ClickHouse | ClickHouse | 581,616,825 | 8,968 | den-crane | Clickhouse version?
Logs? | 2020-02-03T21:02:55 | 2020-02-03T21:02:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 581,659,162 | 6,833 | proxo | I've the same issue i.e. : `Code: 439, e.displayText() = DB::Exception: Cannot schedule a task`,
when creating more than 10 000 buffer tables. It looks like that a buffer table allocates and holds a thread from the global thread pool.
*How to reproduce*
* Run python script with clickhouse driver client:
```
... | 2020-02-03T22:55:53 | 2020-02-03T22:57:46 | {
"+1": 4
} | NONE |
ClickHouse | ClickHouse | 580,717,725 | 8,226 | simPod | Works, thanks! | 2020-01-31T12:39:43 | 2020-01-31T12:39:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 581,163,003 | 8,925 | morozovsk | ```
CREATE TABLE nginx.report_by_query
(`timestamp` DateTime, `country` String, `os` String, `query` String, `n` UInt64)
ENGINE = SummingMergeTree
PARTITION BY toYYYYMM(timestamp)
ORDER BY (timestamp, country, query)
SETTINGS index_granularity = 8192
``` | 2020-02-02T18:28:34 | 2020-02-02T18:41:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 581,496,692 | 8,960 | alexey-milovidov | Created by mistake. | 2020-02-03T16:24:48 | 2020-02-03T16:24:48 | {} | MEMBER |
ClickHouse | ClickHouse | 580,768,324 | 8,933 | alexey-milovidov | > completely on top of the Field class to manage and compare the keys
Some issues with performance. Need to look further.
I will check, do we have any performance tests for `sumMap`...
There were significant complaints on performance of SummingMergeTree in presence of maps, and these complaints were resolved lon... | 2020-01-31T15:00:58 | 2020-01-31T15:00:58 | {} | MEMBER |
ClickHouse | ClickHouse | 581,091,480 | 8,598 | alexey-milovidov | Changelog entry does not describe the changes.
Does it mean that everything was working incorrectly?
| 2020-02-02T02:48:27 | 2020-02-02T02:48:27 | {} | MEMBER |
ClickHouse | ClickHouse | 581,109,344 | 8,925 | filimonov | А структура таблицы? `Show create table nginx.report_by_query` | 2020-02-02T08:04:58 | 2020-02-02T08:04:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 581,172,927 | 8,894 | alexey-milovidov | Ok. | 2020-02-02T20:19:42 | 2020-02-02T20:19:42 | {} | MEMBER |
ClickHouse | ClickHouse | 581,628,046 | 8,837 | AlexJReid | I really didn't expect this to get noticed, but can confirm the workaround works. I've updated the post.
https://mybranch.dev/posts/clickhouse-on-cloud-run/#update | 2020-02-03T21:31:00 | 2020-02-03T21:31:00 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 580,774,086 | 8,933 | alexey-milovidov | No performance tests exist:
```
~/work/ClickHouse/dbms/tests/performance$ grep sumMap *.xml
``` | 2020-01-31T15:15:28 | 2020-01-31T15:41:21 | {} | MEMBER |
ClickHouse | ClickHouse | 580,853,024 | 8,933 | den-crane | I don't really see how to measure speed except run 'optimize final'
```
optimize table X final:
19.13.6.51: Elapsed: 1.482 sec
20.1.3.7: Elapsed: 1.510 sec
20.2.1.2270: Elapsed: 1.914 sec (first PR summap)
20.2.1.2278: Elapsed: 1.888 sec (second (this) PR SummingMergeTree)
I am not sure is it real degradat... | 2020-01-31T18:27:49 | 2020-01-31T18:28:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 580,874,800 | 8,937 | den-crane | duplicate for https://github.com/ClickHouse/ClickHouse/issues/8412
lamda arguments should not be named as columns, use i,j,x,y something
arrayFilter((timestamp, index) -> (index = '1_1') -- wrong
arrayFilter((t, i) -> (i = '1_1') -- right
```
SELECT
countIf(delta_1_1_1_2 > 604800) AS count_delta_1_1_1... | 2020-01-31T19:24:11 | 2020-01-31T19:25:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 581,014,117 | 8,733 | laplab | #### Stateful tests
- `00066_sorting_distributed_many_replicas` fails
Looks like quick fix with fallback to creating separate stream for each shard for `ORDER BY` queries does not work well
#### Stateless tests for all build types
- `00613_shard_distributed_max_execution_time` fails
This test expects that ... | 2020-02-01T10:01:31 | 2020-02-01T10:01:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 581,085,202 | 8,846 | alexey-milovidov | Still don't like the code when we throw/catch in non-exceptional case... | 2020-02-02T00:48:30 | 2020-02-02T00:48:30 | {} | MEMBER |
ClickHouse | ClickHouse | 581,094,327 | 8,840 | alexey-milovidov | @excitoon Please allow edits for your branch. | 2020-02-02T03:25:37 | 2020-02-02T03:25:47 | {} | MEMBER |
ClickHouse | ClickHouse | 581,173,359 | 8,755 | alexey-milovidov | Normal release build should never touch `/usr/include/` files. | 2020-02-02T20:24:14 | 2020-02-02T20:24:14 | {} | MEMBER |
ClickHouse | ClickHouse | 581,181,485 | 7,654 | alexey-milovidov | Conflicted. | 2020-02-02T21:52:38 | 2020-02-02T21:52:38 | {} | MEMBER |
ClickHouse | ClickHouse | 581,237,692 | 8,889 | zombiemonkey | Have isolated this to comparisons in the HAVING clause against a subselect.
We had a few dashboards with the following type of query. This crashes with the above segfault every time.
SELECT uniq(svcs) FROM table GROUP BY toStartOfDay(t_dt) HAVING (count()/48) > (subquery)
The individual queries work fine in is... | 2020-02-03T04:57:50 | 2020-02-03T04:57:50 | {} | NONE |
ClickHouse | ClickHouse | 581,597,625 | 8,986 | alexey-milovidov | It is mergeable but with current state has limited use - will keep under experimental flags for a while. | 2020-02-03T20:15:15 | 2020-02-03T20:15:15 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 581,600,048 | 8,982 | den-crane | CH Version?
What shows:
select configuration_id, count(*) from movement
**where identity(configuration_id)=67**
group by configuration_id; | 2020-02-03T20:21:05 | 2020-02-03T20:21:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 581,614,849 | 8,975 | den-crane | 1. Check results with `set max_block_size=131072`
https://clickhouse.tech/docs/en/query_language/functions/other_functions/#other_functions-runningdifference
runningDifference, neighbor, etc -- reset their value with each block.
---------
2. Check a result of
```
select item, num,neighbor(num, -1) as p... | 2020-02-03T20:57:48 | 2020-02-03T21:00:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 581,054,575 | 8,925 | morozovsk | Запроса в логах нет потому что уровень логгирования был warning, поменял на trace.
И выполнил systemctl enable clickhouse-server.
Получил новую ошибку в то же самоме вермя как предыдущие, только раньше кликхаус просто падал, а сейчас загружает процессор на 100% и при этом недоступен
```
2020.02.01 00:00:05.218606... | 2020-02-01T18:06:50 | 2020-02-01T18:06:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 581,146,358 | 8,598 | filimonov | Allows to put downloaded part to a disks/volumes according to ttl rules (before that change - only by sizes) | 2020-02-02T15:29:25 | 2020-02-02T15:29:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 581,218,311 | 8,957 | alexey-milovidov | @alesapin `No module named confluent.schemaregistry.client` | 2020-02-03T03:08:14 | 2020-02-03T03:08:14 | {} | MEMBER |
ClickHouse | ClickHouse | 581,282,832 | 8,741 | qoega | @alexey-milovidov А ты намеренно положил в ru английскую версию или случайно? | 2020-02-03T07:55:44 | 2020-02-03T07:55:44 | {} | MEMBER |
ClickHouse | ClickHouse | 581,496,912 | 2,755 | nvartolomei | In that case I guess part name is 202002.
```sql
ALTER TABLE table_name DETACH PARTITION '202002';
SELECT sleep(3);
ALTER TABLE table_name ATTACH PARTITION '202002';
```
you can try with and without quotes. | 2020-02-03T16:25:11 | 2020-02-03T16:25:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 580,889,768 | 2,755 | simPod | Me too but no idea how to resume without recreating stucked tables. | 2020-01-31T20:03:34 | 2020-01-31T20:03:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 580,951,154 | 8,860 | den-crane | it fails on long sets [.....] https://github.com/ClickHouse/ClickHouse/issues/8948 | 2020-01-31T23:12:02 | 2020-01-31T23:12:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 581,084,706 | 8,755 | alexey-milovidov | Broken. | 2020-02-02T00:38:58 | 2020-02-02T00:38:58 | {} | MEMBER |
ClickHouse | ClickHouse | 581,085,326 | 8,931 | alexey-milovidov | Ok... but this is too risky and also not very significant change for backport.
Better don't risk.
PS. And why 19.11? Probably it's by mistake. | 2020-02-02T00:50:55 | 2020-02-02T00:50:55 | {} | MEMBER |
ClickHouse | ClickHouse | 581,093,548 | 8,846 | azat | >Still don't like the code when we throw/catch in non-exceptional case...
Me neither, but until it had DRAFT it shouldn't considered seriously.
I'm looking into this and will come back with normal solution or will simply close it | 2020-02-02T03:12:29 | 2020-02-02T03:13:21 | {} | MEMBER |
ClickHouse | ClickHouse | 581,176,850 | 8,840 | alexey-milovidov | Still flapping. | 2020-02-02T21:01:27 | 2020-02-02T21:01:47 | {} | MEMBER |
ClickHouse | ClickHouse | 581,501,663 | 2,755 | simPod | Thanks, I'll try when that happens again. Had to drop tables this time :) | 2020-02-03T16:34:42 | 2020-02-03T16:34:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 580,786,649 | 8,013 | alexey-milovidov | @alesapin said that this PR to be closed. | 2020-01-31T15:46:20 | 2020-01-31T15:46:20 | {} | MEMBER |
ClickHouse | ClickHouse | 580,895,923 | 8,933 | alexey-milovidov | @den-crane We can measure `sumMapWithOverflow` function first. | 2020-01-31T20:20:35 | 2020-01-31T20:20:35 | {} | MEMBER |
ClickHouse | ClickHouse | 581,086,007 | 8,951 | alexey-milovidov | Also doesn't work as expected for HTTP interface:
```
$ curl http://localhost:8123/ -d 'SELECT DISTINCT number % 5 FROM system.numbers'
^C
``` | 2020-02-02T01:03:29 | 2020-02-02T01:03:35 | {} | MEMBER |
ClickHouse | ClickHouse | 581,092,430 | 8,741 | alexey-milovidov | No chance that this will be fixed. | 2020-02-02T03:00:52 | 2020-02-02T03:00:52 | {} | MEMBER |
ClickHouse | ClickHouse | 581,173,114 | 8,755 | alexey-milovidov | It has broken build badly:
```
[3/2169] Building CXX object contrib/boost-cmake/CMakeFiles/boost_iostreams_internal.dir/__/boost/libs/iostreams/src/zlib.cpp.o
FAILED: contrib/boost-cmake/CMakeFiles/boost_iostreams_internal.dir/__/boost/libs/iostreams/src/zlib.cpp.o
/usr/bin/ccache /usr/local/bin/clang++ -DBOOST... | 2020-02-02T20:21:42 | 2020-02-02T20:21:42 | {} | MEMBER |
ClickHouse | ClickHouse | 581,177,260 | 8,780 | alexey-milovidov | Waiting for normal fix. | 2020-02-02T21:05:58 | 2020-02-02T21:05:58 | {} | MEMBER |
ClickHouse | ClickHouse | 581,181,705 | 7,656 | alexey-milovidov | We already have docs for these functions in master. This PR has added duplicated entry. | 2020-02-02T21:54:59 | 2020-02-02T21:54:59 | {} | MEMBER |
ClickHouse | ClickHouse | 581,255,107 | 8,956 | filimonov | Can we downgrade to 1.1.1 stable instead and wait for official 3.0.0 release (it's not even a beta yet)
https://www.openssl.org/blog/blog/2019/11/07/3.0-update/
| 2020-02-03T06:12:59 | 2020-02-03T06:13:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 581,097,750 | 8,952 | azat | >This function is very specific. It works on blocks rather than individual values that is very confusing.
Indeed, nice catch, thanks!
> Maybe name it blockSerializedSize that will resemble blockSize? And we can allow multiple agruments.
Done | 2020-02-02T04:42:02 | 2020-02-02T04:42:02 | {} | MEMBER |
ClickHouse | ClickHouse | 581,257,593 | 8,938 | nmandery | Thank you for the merge and the port of the check form #8945 | 2020-02-03T06:22:35 | 2020-02-03T06:22:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 581,391,892 | 8,971 | 4ertus2 | Related #6767 #6972 | 2020-02-03T12:29:50 | 2020-02-03T12:29:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 581,655,070 | 8,980 | simPod | Ah kk cool. Thanks for info. We did not know what values to expect so we have initially set it rather low. Thanks for confirmation it's fine. It has stopped growing on about 2G. | 2020-02-03T22:43:59 | 2020-02-03T22:43:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 580,724,716 | 8,916 | akuzm | It was too early for this PR. I also need the queue to support shutdown, which requires different pop() signature, and I don't want to change every caller (about ten of them) so I guess I'll just write a separate one. | 2020-01-31T13:02:46 | 2020-01-31T13:02:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 580,881,107 | 5,907 | alexey-milovidov | @nicelulu It is similar but for different case: when SELECT has small limit and will read small amount of data. | 2020-01-31T19:41:40 | 2020-01-31T19:41:40 | {} | MEMBER |
ClickHouse | ClickHouse | 580,946,743 | 8,933 | alexey-milovidov | Small difference is Ok because this function is rarely used.
Let me add your queries as performance test to avoid performance regressions in future. | 2020-01-31T22:55:06 | 2020-01-31T22:55:06 | {} | MEMBER |
ClickHouse | ClickHouse | 581,034,140 | 8,936 | vzakaznikov | So the expected behaviour is that if the Live View table did not execute the stored query it will execute it on the first access. Either when you run the first `SELECT` or the first `WATCH` query. After that the Live View stays synced via INSERT "trigger" to the source table that is specified in the query. If for some ... | 2020-02-01T14:15:46 | 2020-02-01T14:15:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 581,078,428 | 2,755 | nvartolomei | @simPod detach/attach affected partition, suggested by @alesapin some time ago. | 2020-02-01T23:01:17 | 2020-02-01T23:01:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 581,085,035 | 8,846 | alexey-milovidov | > Interesting, why the xxHash64 does not exist?
It is unique to `unbundled` build that lacks some libraries.
We manually exclude some tests for this kind of build. | 2020-02-02T00:45:12 | 2020-02-02T00:45:12 | {} | MEMBER |
ClickHouse | ClickHouse | 581,091,656 | 8,739 | alexey-milovidov | @godfreyd @4ertus2 Всё ещё плохо написано.
Соединение идёт не по "списку ключей", а по ключу, который может представлять собой кортеж. | 2020-02-02T02:52:34 | 2020-02-02T02:52:34 | {} | MEMBER |
ClickHouse | ClickHouse | 581,132,042 | 8,952 | alexey-milovidov | Ok. *Almost* ready.
The usage of `MemoryWriteBuffer` is (1) overkill, (2) incorrect.
1. We should invent `NullWriteBuffer` that will do nothing on `nextImpl` method (`next` will just move cursor to the front and overwrite data). The size of buffer can be smaller, example: 16K for better cache locality.
2. You use th... | 2020-02-02T12:45:29 | 2020-02-02T12:46:30 | {} | MEMBER |
ClickHouse | ClickHouse | 581,181,308 | 7,525 | alexey-milovidov | Ok. | 2020-02-02T21:50:30 | 2020-02-02T21:50:30 | {} | MEMBER |
ClickHouse | ClickHouse | 581,186,509 | 6,663 | alexey-milovidov | Superseded. | 2020-02-02T22:50:07 | 2020-02-02T22:50:07 | {} | MEMBER |
ClickHouse | ClickHouse | 581,381,829 | 8,689 | 4ertus2 | It's most probably related to #7497
There's some strange expectation about previous chain steps here:
https://github.com/ClickHouse/ClickHouse/pull/7497/files#diff-22f780c9dde88bec19a9dac8ce578f1cR719
| 2020-02-03T12:02:45 | 2020-02-03T12:03:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 581,659,426 | 4,050 | den-crane | >Is is wrong to have partitioning based on days?
>I'm asking because I have daily parts, each is over 1 TB.
@simPod it not your case. You parts are not tiny (1TB rather huge). Daily partitions is OK, probably in your case they are strongly recommended.
>So ZK and CH share the instance but not disks.
>Accordi... | 2020-02-03T22:56:33 | 2020-02-03T22:56:48 | {
"heart": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 581,033,230 | 8,936 | vzakaznikov | Hi @byx313, I can't reproduce the issue that you are seeing. Here is my test case.
1. I enable live views and create a live view table.
```
ClickHouse client version 20.2.1.1.
Connecting to localhost:9000 as user default.
Connected to ClickHouse server version 20.2.1 revision 54432.
:) SET allow_experimental_... | 2020-02-01T14:04:06 | 2020-02-01T14:04:06 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 581,063,845 | 8,846 | azat | >2020-01-31 05:45:31 Code: 46. DB::Exception: Received from localhost:9000. DB::Exception: Unknown function xxHash64. Maybe you meant: ['sipHash64','intHash64'].
[Interesting](https://clickhouse-test-reports.s3.yandex.net/8846/de5304cf8ad40c7315b91596a26b936f76622d9f/functional_stateless_tests_(unbundled)/test_run.... | 2020-02-01T19:55:58 | 2020-02-01T19:56:33 | {} | MEMBER |
ClickHouse | ClickHouse | 581,084,405 | 8,949 | alexey-milovidov | Ok.
PS. No idea why performance test failed badly. Probably some server from our CI fleet was overheated :) Will look at the results on next commits. | 2020-02-02T00:33:14 | 2020-02-02T00:33:14 | {} | MEMBER |
ClickHouse | ClickHouse | 581,396,220 | 8,931 | alexey-milovidov | Ok. No need for cleanup. | 2020-02-03T12:42:39 | 2020-02-03T12:42:39 | {} | MEMBER |
ClickHouse | ClickHouse | 581,415,753 | 8,754 | vapongit | const related issue #7666 | 2020-02-03T13:34:11 | 2020-02-03T13:34:11 | {} | NONE |
ClickHouse | ClickHouse | 581,633,652 | 8,987 | den-crane | >We expect our query patterns to span no more than 3 dates
Then daily partitioning is reasonable and will improve your performance.
>but from the documentation it sounds like performance
performance MAY suffer if your query have to process 1000 partitions instead of 10 partitions.
But you need to write SQL... | 2020-02-03T21:45:03 | 2020-02-03T21:45:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 581,186,406 | 7,120 | alexey-milovidov | January is over. | 2020-02-02T22:48:52 | 2020-02-02T22:48:52 | {} | MEMBER |
ClickHouse | ClickHouse | 581,566,111 | 8,022 | alexey-milovidov | Sorry for the delay.
I will try to allocate the time today. | 2020-02-03T19:04:23 | 2020-02-03T19:04:23 | {} | MEMBER |
ClickHouse | ClickHouse | 581,728,977 | 8,889 | SaltTan | Does it crash if `enable_scalar_subquery_optimization ` is disabled? | 2020-02-04T03:41:36 | 2020-02-04T03:41:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 580,729,517 | 8,917 | filimonov | Performance details:
```
# baseline before that PR
┌─version()──┬─block_size─┬─consume_time─┬─consume_speed─┐
│ 20.2.1.2269 │ 65536 │ 122 │ 137518.163934 │
│ 20.2.1.2269 │ 131072 │ 120 │ 139810.133333 │
│ 20.2.1.2269 │ 524288 │ 116 │ 144631.172413 │
│ 20.2.1.2269 │ 1048... | 2020-01-31T13:18:01 | 2020-01-31T13:23:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 580,930,462 | 8,924 | den-crane |
```
create table test1 Engine=MergeTree order by tuple()
as select toUInt64(number%11111111) key, toUInt64(1) value from numbers(10000000);
SELECT key, min(value) FROM test1 group by key format Null;
Elapsed: 0.531 sec. Processed 10.00 million rows
SELECT key, value FROM test1 order by key, value limit 1 by... | 2020-01-31T22:03:31 | 2020-01-31T22:03:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 581,084,343 | 8,952 | alexey-milovidov | This function is very specific. It works on blocks rather than individual values that is very confusing.
Maybe name it `blockSerializedSize` that will resemble `blockSize`? And we can allow multiple agruments. | 2020-02-02T00:32:05 | 2020-02-02T00:32:05 | {} | MEMBER |
ClickHouse | ClickHouse | 581,182,400 | 7,719 | alexey-milovidov | Was documented in newer PR. | 2020-02-02T22:03:21 | 2020-02-02T22:03:21 | {} | MEMBER |
ClickHouse | ClickHouse | 581,518,908 | 8,957 | oandrew | I'm guessing CI doesn't build new `yandex/clickhouse-integration-tests-runner` image from Dockerfile (`dbms/tests/integration/image/Dockerfile`) i.e. someone needs to build and publish it manually | 2020-02-03T17:12:39 | 2020-02-03T17:15:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 581,612,951 | 8,980 | den-crane | Zookeeper is in-memory DB
Check your DB size
```
echo mntr | nc zoo1 2181|grep data_size
zk_approximate_data_size 122079537
```
1GB is tiny for a real production. I use `-Xms8g` | 2020-02-03T20:53:04 | 2020-02-03T20:53:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 610,092,028 | 9,836 | filimonov | > You are correct. Bigger parts are better
Glad you agree.
> Just imaging some worst case.
Try to create 1 mln sub folders in single folder and do a simple `ls` command there
> Not sure the pipelining execution model helps
I'm definitely against that.
That partially loaded table = "inconsistancy". Fo... | 2020-04-06T23:40:38 | 2020-04-06T23:40:38 | {} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.