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ClickHouse | ClickHouse | 691,935,785 | 12,281 | alexey-milovidov | Still outdated. | 2020-09-14T09:26:41 | 2020-09-14T09:26:41 | {} | MEMBER |
ClickHouse | ClickHouse | 692,004,860 | 14,697 | 4ertus2 | Ya check it OK after 'merge master', other tests (except perf) were OK before. | 2020-09-14T11:56:36 | 2020-09-14T11:56:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,178,031 | 13,281 | MaskRay | I see `llvm::orc::LegacyRTDyldObjectLinkingLayer` in the stack trace and thus wonder whether ClickHouse is still using legacy ORCv1 JIT: https://llvm.org/docs/ORCv2.html#transitioning-from-orcv1-to-orcv2
> The majority of the ORCv1 layers and utilities were renamed with a ‘Legacy’ prefix in LLVM 8.0, and have deprec... | 2020-09-14T16:45:34 | 2020-09-14T16:53:46 | {} | NONE |
ClickHouse | ClickHouse | 691,786,900 | 9,375 | jason-godden | @Avogar - thanks so much!
Thanks @alexey-milovidov! | 2020-09-14T03:15:15 | 2020-09-14T03:15:15 | {} | NONE |
ClickHouse | ClickHouse | 691,838,837 | 14,771 | amosbird | > I would suggest something like this:
That's more complicated and not similar to any standard.
> x3 as (select 3) -- error?
It's consistent with other databases. | 2020-09-14T06:18:47 | 2020-09-14T06:19:17 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 691,936,572 | 11,103 | alexey-milovidov | Is it fixed? | 2020-09-14T09:28:13 | 2020-09-14T09:28:13 | {} | MEMBER |
ClickHouse | ClickHouse | 691,936,640 | 13,512 | CurtizJ | Yes, it should be fixed in #13538. | 2020-09-14T09:28:20 | 2020-09-14T09:28:20 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 692,164,812 | 13,156 | vitlibar | Some [performance tests](https://clickhouse-test-reports.s3.yandex.net/13156/0e0ae550e7984f02afce7921c8aa17cb4f85da0a/performance_comparison/report.html#fail1) show slow down. Could you help investigate that? Just add an empty commit or rebase/squash your branch to see if there will be the same performance tests which ... | 2020-09-14T16:20:40 | 2020-09-14T16:50:39 | {} | MEMBER |
ClickHouse | ClickHouse | 692,224,128 | 14,782 | den-crane | Please show ` env|grep -i term ` | 2020-09-14T18:11:09 | 2020-09-14T18:11:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 691,991,932 | 14,697 | 4ertus2 | Stress test was OK before 'merge master'. Most probably It's not related to PR changes. | 2020-09-14T11:26:49 | 2020-09-14T11:26:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,025,838 | 6,847 | alexey-milovidov | Support period for version 19.14 LTS has been ended. | 2020-09-14T12:40:57 | 2020-09-14T12:40:57 | {} | MEMBER |
ClickHouse | ClickHouse | 692,137,718 | 14,514 | bharatnc | @nikitamikhaylov looks like I see the report generation is failing on master and `one of the errors` looks related to this PR:
```
Code: 117, e.displayText() = DB::Exception: Unknown field found in TSV header: 'event_time_microseconds' at position 2
```
https://clickhouse-test-reports.s3.yandex.net/14453/016f707ea... | 2020-09-14T15:34:28 | 2020-09-14T15:34:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,204,632 | 14,813 | KochetovNicolai | > Both inlining + anonymous namespace - to lower the size of symbol table and debug info?
Yes. With inlining some function names are removed from debug info.
In total it saves just ~100M for my local build, however. | 2020-09-14T17:34:48 | 2020-09-14T17:34:48 | {} | MEMBER |
ClickHouse | ClickHouse | 692,216,499 | 14,784 | den-crane | 1. Clickhouse is not a timeseries DB.
2. ALL your skip **indexes are WRONG. Remove them.**
INDEX idxTime (timestamp) TYPE minmax
INDEX idxbTime (b,timestamp) TYPE minmax
skip indexes <> BTree.
b,timestamp === minmax B, minmax timestamp
idxTime is excessive because it's covered by idxbTime.
3. Yo... | 2020-09-14T17:56:53 | 2020-09-14T18:04:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,249,818 | 14,792 | alesapin | Build failure is OOM. Most of the tests passed. | 2020-09-14T19:00:16 | 2020-09-14T19:00:16 | {} | MEMBER |
ClickHouse | ClickHouse | 692,255,874 | 14,646 | alexey-milovidov | @filimonov Sure. We have backported it. | 2020-09-14T19:11:16 | 2020-09-14T19:11:16 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 692,257,466 | 14,747 | Slind14 | This did not happen before with the same version of clickhouse.
@alexey-milovidov I see you added `obsolete-version`, which one would you recommend upgrading to without the smallest possible risk? (we don't need the new features yet)
(the versioning is a bit unclear, it doesn't seem to follow Semantic Versioning)
... | 2020-09-14T19:14:19 | 2020-09-14T19:16:54 | {} | NONE |
ClickHouse | ClickHouse | 691,629,403 | 11,987 | amosbird | Custom settings are now available https://github.com/ClickHouse/ClickHouse/blob/bb832f204c6ae49e7d8dc93ed5da9fe2c0ac2751/tests/queries/0_stateless/01418_custom_settings.sql | 2020-09-13T07:47:00 | 2020-09-13T07:47:00 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 691,719,617 | 14,777 | den-crane | Parts (and files) are immutable in CH. CH never changes anything once written. CH creates a new part (file) and copies & and mutates (or merges) rows from a source part(s).
Mutations (and merges) book twice disk space as original parts. It is done this way because mutations re-write original parts (row by row) to th... | 2020-09-13T20:16:39 | 2020-09-13T20:20:05 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 691,870,545 | 14,732 | glader | > Далее подключить эту таблицу как [словарь](https://clickhouse.tech/docs/en/sql-reference/statements/create/dictionary/), чтобы доставать обновления налету.
Ты имел в виду джойнить большую таблицу со словарем? Насколько я помню, клик не очень хорошо джойнит. | 2020-09-14T07:21:41 | 2020-09-14T07:21:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,019,913 | 8,331 | Vxider | > > > what's the progress of this feature?CAN be Released in this Year?
> >
> >
> > The basic implementation of WindowView is done. We are waiting for a code review now.
> > We currently have another optimized version, and we will consider committing to the community after the basic version is merged.
>
> The ... | 2020-09-14T12:28:26 | 2020-09-14T12:28:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,058,660 | 14,654 | alexey-milovidov | test_dictionaries_all_layouts_and_sources
test_storage_rabbitmq
- bad tests, ignore. | 2020-09-14T13:38:27 | 2020-09-14T13:38:27 | {} | MEMBER |
ClickHouse | ClickHouse | 692,237,545 | 11,948 | nikitamikhaylov | Synchronization check is already fixed in master. And integration test failure seems unrelated. | 2020-09-14T18:37:09 | 2020-09-14T18:37:09 | {} | MEMBER |
ClickHouse | ClickHouse | 692,238,333 | 14,764 | alexey-milovidov | > 1- is this memory consumption normal?
Yes, it's normal.
> 2 - Is it production ready, our server version 20.3.17 revision 54433?
We are still preparing compact parts for production. It is almost production ready but we still find some issues in corner cases. There are recent bugfixes in version 20.8.
> 3 ... | 2020-09-14T18:38:16 | 2020-09-14T18:38:16 | {} | MEMBER |
ClickHouse | ClickHouse | 692,250,290 | 14,805 | alesapin | Tests almost OK, let's check in branch | 2020-09-14T19:01:12 | 2020-09-14T19:01:12 | {} | MEMBER |
ClickHouse | ClickHouse | 691,630,810 | 3,815 | amosbird | > Without passing the variable from parent process I still have to substitute script before running it.
Aren't env vars expanded by shells automatically? You can build a `sql.sh` which exports all your needed variables to ClickHouse.
`SELECT getEnv('ENVVAR1')` will be far away from your expectation as the SQL is ... | 2020-09-13T07:57:58 | 2020-09-13T07:57:58 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 692,028,554 | 14,732 | alexey-milovidov | ClickHouse очень хорошо JOIN-ит, лучше чем кто-либо ещё - если использовать словарь. | 2020-09-14T12:46:26 | 2020-09-14T12:46:26 | {} | MEMBER |
ClickHouse | ClickHouse | 692,057,297 | 14,770 | alexey-milovidov | > Push to dockerhub
Don't needed for backport - Ok. | 2020-09-14T13:36:18 | 2020-09-14T13:36:18 | {} | MEMBER |
ClickHouse | ClickHouse | 692,168,522 | 14,794 | 4ertus2 | > How big the TPC-H data-set? I mean build the memory hash table need cost many memory and time. Maybe the data-set not big enough?
I do not remebmer data-set size exactly. But hash table size was about 10-16 Gb.
> So it should be what you said 'data in both tables is already sorted by joining key' ?
Let me de... | 2020-09-14T16:27:34 | 2020-09-14T16:32:01 | {
"+1": 6
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,198,571 | 14,813 | alexey-milovidov | Both inlining + anonymous namespace - to lower the size of symbol table and debug info? | 2020-09-14T17:23:07 | 2020-09-14T17:23:07 | {} | MEMBER |
ClickHouse | ClickHouse | 692,257,789 | 13,166 | ildus | Thanks! | 2020-09-14T19:14:55 | 2020-09-14T19:14:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 691,636,207 | 3,815 | jangorecki | Agree, although my client is running on same machine as server. Could you please point to a documentation that describes how to export all needed env vars so I can use them inside sql script? | 2020-09-13T08:35:55 | 2020-09-13T08:35:55 | {} | NONE |
ClickHouse | ClickHouse | 691,669,183 | 14,513 | KochetovNicolai | Code looks ok, but we need to investigate crash in tests. | 2020-09-13T13:05:32 | 2020-09-13T13:05:32 | {} | MEMBER |
ClickHouse | ClickHouse | 691,685,157 | 14,774 | den-crane | WA: `alter TABLE ".inner.city_view" ON cluster ch_cluster_all delete WHERE date=20200910` | 2020-09-13T15:23:05 | 2020-09-13T15:23:05 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 691,763,037 | 13,875 | mjodeh | `iowait` on the server jumps suddenly from 4% to 88% and then the `Memory Limit Exceeded` exception is fired after that the server is down.
Also the `context switch` starts rising from 3.2 ksps to 13 ksps and suddenly drops to 0!
We are using `HTTP` interface
1 - Does the partitioning `daily` or `monthly` affect iow... | 2020-09-14T01:32:52 | 2020-09-14T01:33:32 | {} | NONE |
ClickHouse | ClickHouse | 691,801,900 | 14,782 | l1t1 | putty 0.7 is ok | 2020-09-14T04:23:41 | 2020-09-14T04:23:41 | {} | NONE |
ClickHouse | ClickHouse | 692,008,236 | 13,574 | nvartolomei | For the past week I've been experimenting with various approaches for adding "fingerprint" metadata to the parts and moving data between shards.
Orchestrating the whole process is a bit complicated and we reached the conclusion that the best way for now is to add a minimal set of primitives that would allow an opera... | 2020-09-14T12:03:57 | 2020-09-14T12:03:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,209,769 | 14,809 | 4ertus2 | There's a questionable warning in gcc build described here https://gcc.gnu.org/bugzilla/show_bug.cgi?id=55776 | 2020-09-14T17:44:34 | 2020-09-14T17:44:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 691,857,215 | 14,705 | alesapin | AST fuzzer reported https://github.com/ClickHouse/ClickHouse/issues/14700, performance not related. | 2020-09-14T06:56:44 | 2020-09-14T06:56:44 | {} | MEMBER |
ClickHouse | ClickHouse | 691,869,619 | 14,732 | glader | Меня устроят варианты среднего качества. Какой из них наименее плохой? | 2020-09-14T07:20:04 | 2020-09-14T07:20:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 691,673,351 | 13,015 | KochetovNicolai | `AST-based query fuzzer` link is 404 now. Hope it was not important. | 2020-09-13T13:44:50 | 2020-09-13T13:44:50 | {} | MEMBER |
ClickHouse | ClickHouse | 691,791,488 | 14,684 | zhang2014 | > Does not lock for ALTER/... should handle this?
Does the guarantee the execution order of the same table DDLS(it seems that we have no strict guarantee)? | 2020-09-14T03:34:33 | 2020-09-14T03:34:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 691,892,543 | 14,782 | xiashaohua | I have the same problem when i update to clickhouse 20.7.1.1。 | 2020-09-14T08:05:39 | 2020-09-14T08:05:39 | {} | NONE |
ClickHouse | ClickHouse | 692,058,142 | 14,668 | alexey-milovidov | All failures were fixed. | 2020-09-14T13:37:39 | 2020-09-14T13:37:39 | {} | MEMBER |
ClickHouse | ClickHouse | 692,058,909 | 14,670 | alexey-milovidov | All failures were fixed. | 2020-09-14T13:38:45 | 2020-09-14T13:38:45 | {} | MEMBER |
ClickHouse | ClickHouse | 686,577,771 | 12,736 | traceon | @vitlibar The PR is ready (once all tests are OK). The only thing to recheck: are we still OK with `User` entity loitering (an instance is created for each user request, without consulting LDAP server, and stored there forever, even for non-existent users). | 2020-09-03T15:39:26 | 2020-09-14T14:49:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,157,875 | 13,166 | Akazz | Fuzzer failure is due to what is described in #14811 | 2020-09-14T16:08:46 | 2020-09-14T16:08:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 691,639,686 | 3,815 | jangorecki | and when I use sql script rather than query?
```sh
cat "clickhouse/task-clickhouse.sql" | clickhouse-client ...
```
I have to still substitute vars in sql having `.sql.in` template script, isn't it? | 2020-09-13T09:00:40 | 2020-09-13T09:00:55 | {} | NONE |
ClickHouse | ClickHouse | 691,827,921 | 14,535 | zhang2014 | > Integration tests (release) — fail: 1, passed: 814, error: 0
> Performance — Errors while building the report.
test failure unrelated this pr. | 2020-09-14T05:52:55 | 2020-09-14T05:52:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 691,856,054 | 14,769 | alesapin | Fixed here https://github.com/ClickHouse/ClickHouse/pull/14646 | 2020-09-14T06:54:32 | 2020-09-14T06:54:32 | {} | MEMBER |
ClickHouse | ClickHouse | 691,990,643 | 14,796 | 4ertus2 | Related #11414 #12035 | 2020-09-14T11:24:08 | 2020-09-14T11:41:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,057,127 | 14,770 | alexey-milovidov | OOM in build Ok. | 2020-09-14T13:36:01 | 2020-09-14T13:36:01 | {} | MEMBER |
ClickHouse | ClickHouse | 692,226,735 | 14,774 | den-crane | ` .inner` table is a real ReplicatedAggregatingMergeTree table where MV stores it's data.
MATERIALIZED views implicitly create tables to store data and name these tables as .inner.{mvname} | 2020-09-14T18:16:08 | 2020-09-14T18:16:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,250,005 | 14,793 | alesapin | Most tests passed. | 2020-09-14T19:00:40 | 2020-09-14T19:00:40 | {} | MEMBER |
ClickHouse | ClickHouse | 691,622,075 | 14,711 | filimonov | Продублирую тут:
из моих записок - вот тут есть релеватный фрагмент в документации (добавлен недавно)
https://github.com/ClickHouse/ClickHouse/blob/master/docs/en/development/build.md#faster-builds-for-development
Вот какие-то из наборов которыми можно воспользоваться для ускорения билдов:
```
cmake .. -DCMAKE... | 2020-09-13T06:38:45 | 2020-09-13T06:40:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 691,642,986 | 3,815 | amosbird | Then you'd better use https://clickhouse.tech/docs/en/interfaces/cli/#cli-queries-with-parameters
And you can wrap the `clickhouse-client` into a script to set up the parameters you need. | 2020-09-13T09:21:59 | 2020-09-13T09:22:29 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 691,789,416 | 14,781 | zhang2014 | what is the table structure? | 2020-09-14T03:25:43 | 2020-09-14T03:25:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 691,805,382 | 14,781 | chang0391 | no error found, table structure as below
CREATE TABLE `mobile_belong_test` (
`id` bigint(20) NOT NULL AUTO_INCREMENT,
`section` int(11) DEFAULT NULL,
`provinceId` varchar(10) DEFAULT NULL,
`province` varchar(30) DEFAULT NULL,
`cityId` varchar(10) DEFAULT NULL,
`city` varchar(30) DEFAULT NULL,
`sig... | 2020-09-14T04:37:24 | 2020-09-14T04:37:24 | {} | NONE |
ClickHouse | ClickHouse | 691,899,326 | 14,684 | azat | >Does the guarantee the execution order of the same table DDLS(it seems that we have no strict guarantee)?
Let's consider the following examples:
- no `ON CLUSTER` - the order is not strictly defined (since there is tiny probability that the last query can be scheduled earlier then the first, due to some schedule... | 2020-09-14T08:18:52 | 2020-09-14T08:18:52 | {} | MEMBER |
ClickHouse | ClickHouse | 691,953,990 | 11,103 | KochetovNicolai | Not yet.
```sql
EXPLAIN actions = 1
SELECT throwIf(number = 5)
FROM
(
SELECT *
FROM numbers(10)
)
ORDER BY number ASC
LIMIT 1
FORMAT TSV
Expression (Projection)
Actions: PROJECT throwIf(equals(number, 5))
PROJECT throwIf(equals(number, 5))
Limit (preliminary LIMIT)
Limit 1
O... | 2020-09-14T10:03:31 | 2020-09-14T10:03:31 | {} | MEMBER |
ClickHouse | ClickHouse | 692,050,958 | 14,737 | alesapin | First of all your schema is quite suboptimal for your data. You have partition key `toDate(ipdrCreationTime)` and if I understand correctly in each insert you have a lot of different days? In this case, each insert will be split into many parts one for each partition and a lot of files will be created on disk.
> If ... | 2020-09-14T13:26:04 | 2020-09-14T13:26:04 | {} | MEMBER |
ClickHouse | ClickHouse | 692,128,942 | 11,844 | Enmk | @qoega Fixed all notes regarding install paths and default configurations. | 2020-09-14T15:21:11 | 2020-09-14T15:21:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,137,657 | 14,794 | compasses | @4ertus2 very appreciate for your quick answer, please check my comments
> I checked 'partial_merge' vs 'hash' algo when it was implemented on data from TPC-H test. It was 3-4 times slower when all data is in memory. So in general it should be slower.
How big the TPC-H data-set? I mean build the memory hash tabl... | 2020-09-14T15:34:22 | 2020-09-14T15:34:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,209,830 | 14,807 | den-crane | Check errors in CH logs /var/log/clickhouse-server/
Try to remove/ comment empty tags `<root></root> <identity></identity>` | 2020-09-14T17:44:40 | 2020-09-14T17:44:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,219,749 | 13,156 | azat | >Some performance tests show slow down. Could you help investigate that? Just add an empty commit or rebase/squash your branch to see if there will be the same performance tests which slow down.
Ok, sure.
But if you will look into previous performance report (before merging upstream and std::erase replace) you will... | 2020-09-14T18:02:44 | 2020-09-14T18:02:44 | {} | MEMBER |
ClickHouse | ClickHouse | 691,629,263 | 3,815 | jangorecki | @amosbird thanks, that looks promising and ~~flexible enough to address cases mentioned by me above~~. Do you have any idea how can I declare variables (`SET custom_a = 5;`) from command line just before running script? Or if there is `getSetting`'s style way to access environment variables? so I could `SET var1 = SELE... | 2020-09-13T07:46:00 | 2020-09-13T08:36:16 | {} | NONE |
ClickHouse | ClickHouse | 691,763,210 | 14,362 | bobrik | This increased our internal build times significantly, to the point where they started breaching 3h timeout. Previous build of v20.7 took just 1h14m.
On a system with 2 x Xeon E5-2630 v3 (32 logical CPUs in total) I see:
* 92m with job-per-logical-cpu (before)
* 100m with job-per-physical-core (after)
Memory ... | 2020-09-14T01:33:42 | 2020-09-14T01:33:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,057,112 | 14,697 | 4ertus2 | There's a crash in CacheDictionary in Stress test server.log https://clickhouse-test-reports.s3.yandex.net/14697/4d5615cb121712ea2ab497800bda295539b49cd8/stress_test_(memory)/clickhouse-server.log
```
2020.09.14 13:41:08.638417 [ 9598 ] {} <Fatal> BaseDaemon: ########################################
2020.09.14 13:41... | 2020-09-14T13:35:59 | 2020-09-14T13:35:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 691,622,946 | 14,771 | amosbird | For the reference cond.
```
Amos Bird, [13.09.20 14:24]
> For the reference (that both x AS subquery and subquery AS x are Ok):
Hmm, do you want both variants to be treated exactly the same? Currently I left the old one's behavior intact..
Alexey Milovidov, [13.09.20 14:25]
It is treat as scalar subquery?
... | 2020-09-13T06:50:09 | 2020-09-13T06:50:09 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 682,354,574 | 13,159 | amosbird | There is no CTE in clickhouse yet. I'd like to denote the current `WITH` statement as CST (common scalar expression) and table functions are not scalars.
The reason that current `with` doesn't work in `jdbc` or other table functions can be clearly shown in the following query rewrite:
```
with 'where somecolumn... | 2020-08-28T06:34:25 | 2020-09-13T07:37:17 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 691,638,168 | 3,815 | amosbird | It's just bare shell interpolation.
```bash
sql_prelude="
set custom_foo=$foo;
set custom_bar=$bar;
"
query=...
clickhouse client --query "$sql_prelude $query"
``` | 2020-09-13T08:49:49 | 2020-09-13T08:49:49 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 691,800,053 | 14,771 | hczhcz | It seems slightly confusing to me.
```sql
with
x1 as (select 1) -- syntax #1
select
...
from
x1,
(select 2) as x2, -- syntax #2
x3 as (select 3) -- error?
```
I would suggest something like this:
```sql
with query
(select 1) as x
with
(select 1) as y
select
y
fro... | 2020-09-14T04:15:06 | 2020-09-14T04:15:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 691,862,955 | 14,786 | liangli1213 | @BohuTANG @zhang2014 , I don't think this is caused by your code, because if I don't restart ClickHouse service after the replication has been setup, the issue won't occur and the replication can keep going on over night, but whenever I restart the service, the issue happens rapidly. | 2020-09-14T07:07:26 | 2020-09-14T07:07:26 | {} | NONE |
ClickHouse | ClickHouse | 691,928,180 | 13,512 | alexey-milovidov | @CurtizJ is it fixed? | 2020-09-14T09:12:16 | 2020-09-14T09:12:16 | {} | MEMBER |
ClickHouse | ClickHouse | 692,061,060 | 14,414 | 4ertus2 | Unknown crash in stress test in CacheDictionary https://github.com/ClickHouse/ClickHouse/pull/14697#issuecomment-692057112 | 2020-09-14T13:42:16 | 2020-09-14T13:42:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,200,796 | 14,646 | filimonov | to 20.3? | 2020-09-14T17:27:24 | 2020-09-14T17:27:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,248,080 | 14,813 | azat | It will be interesting to see will it helps to reduce build time for gcc10, or the problem wasn't with the amount of debug symbols. | 2020-09-14T18:57:00 | 2020-09-14T18:57:00 | {} | MEMBER |
ClickHouse | ClickHouse | 692,260,301 | 14,804 | alexey-milovidov | Not a bugfix. | 2020-09-14T19:19:19 | 2020-09-14T19:19:19 | {} | MEMBER |
ClickHouse | ClickHouse | 691,669,953 | 14,513 | KochetovNicolai | ```
#0 0x0000000015f524e1 in DB::KeyCondition::isKeyPossiblyWrappedByMonotonicFunctionsImpl(std::__1::shared_ptr<DB::IAST> const&, unsigned long&, std::__1::shared_ptr<DB::IDataType const>&, std::__1::vector<DB::ASTFunction const*, std::__1::allocator<DB::ASTFunction const*> >&) ()
#1 0x0000000015f525b9 in DB::KeyC... | 2020-09-13T13:13:25 | 2020-09-13T13:13:25 | {} | MEMBER |
ClickHouse | ClickHouse | 691,976,119 | 13,538 | vinodkmrm | Thank you @CurtizJ | 2020-09-14T10:51:46 | 2020-09-14T10:51:46 | {} | NONE |
ClickHouse | ClickHouse | 691,976,553 | 14,774 | Awezome | I didn't even know there was .inner table before. I searched .inner table in docs and got nothing . Where can I get more information about it ? Thanks! | 2020-09-14T10:52:49 | 2020-09-14T10:52:49 | {} | NONE |
ClickHouse | ClickHouse | 692,041,501 | 14,794 | 4ertus2 | I checked 'partial_merge' vs 'hash' algo when it was implemented on data from TPC-H test. It was 3-4 times slower when all data is in memory. So in general it should be slower.
Any case, It's expected that in some cases (when data in both tables is already sorted by joining key) MergeJoin would be faster than HashJo... | 2020-09-14T13:10:25 | 2020-09-14T13:10:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,046,504 | 14,802 | den-crane | >_partition_id = '20200914'
1. It's huge mistake overall. CH does not guarantee that partition pruning will work against _partition_id column in the future.
use `toYYYYMMDD(loaded) = 20200914`
or better `loaded >= '2020-09-14 00:00:00' and loaded < '2020-09-15 00:00:00'`
2. _partition_id is a virtual... | 2020-09-14T13:18:34 | 2020-09-14T13:19:07 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,206,362 | 13,281 | MaskRay | The code owner of ORC JIT has announced http://lists.llvm.org/pipermail/llvm-dev/2020-September/144911.html
It seems that ORCv1 will disappear in llvm-12. If you find challenging issues, please reach out to llvm-dev. (I know very little about ORC JIT and was just asked by amosbird about how to react to this issue...) | 2020-09-14T17:38:13 | 2020-09-14T17:38:13 | {} | NONE |
ClickHouse | ClickHouse | 691,852,853 | 14,229 | alexey-milovidov | @4ertus2 is going to remove wide_integer from types.h. | 2020-09-14T06:48:06 | 2020-09-14T06:48:06 | {} | MEMBER |
ClickHouse | ClickHouse | 691,854,597 | 14,646 | alesapin | Performance OK | 2020-09-14T06:51:36 | 2020-09-14T06:51:36 | {} | MEMBER |
ClickHouse | ClickHouse | 691,999,935 | 2,416 | Tasselmi | Thx.~
| |
梁凡
邮箱：tasselmi@yeah.net
|
签名由 网易邮箱大师 定制
On 09/13/2020 01:08, Amos Bird wrote:
to reuse subquery result , please save a kid has suffered so much ！！！
Let's help this poor kid #14771
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub, or unsubscribe. | 2020-09-14T11:45:17 | 2020-09-14T11:45:17 | {} | NONE |
ClickHouse | ClickHouse | 692,057,661 | 14,770 | alexey-milovidov | Stress test:
> AttributeError: 'generator' object has no attribute 'rfind'
Something is wrong with Python script - ignore. | 2020-09-14T13:36:54 | 2020-09-14T13:36:54 | {} | MEMBER |
ClickHouse | ClickHouse | 692,076,673 | 14,755 | vitlibar | > It shouldn't be related because `shutil.copytree` creates `dst` directory. What is more, it requires `dst` directory not exist: https://docs.python.org/2/library/shutil.html#shutil.copytree:
>
> ```
> Recursively copy an entire directory tree rooted at src. The destination directory, named by dst, must not alread... | 2020-09-14T14:08:22 | 2020-09-14T14:08:22 | {} | MEMBER |
ClickHouse | ClickHouse | 692,124,542 | 5,555 | filimonov | Actually the use case I had initially was more about `ATTACH TABLE x EPHEMERALLY` (so don't create metadata for the table at all), but it seems `DETACH TABLE xyz PERMANENTLY` is also needed, and it would be nice to keep metadata for those tables in place. | 2020-09-14T15:14:14 | 2020-09-14T15:14:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,181,247 | 13,281 | alexey-milovidov | We have migrated from llvm-7 to llvm-10 and it was quite hard to do - I just bisected and followed up every commit that broke ABI.
Initially I tried to migrate in single step by switching to the non-Legacy interfeces... but did not succeed, it was very hard to me to grok all the interfaces and how to stick them toge... | 2020-09-14T16:51:21 | 2020-09-14T16:51:21 | {} | MEMBER |
ClickHouse | ClickHouse | 692,250,823 | 14,282 | alexey-milovidov | @l1t1 Here is the changelog for 20.8: https://github.com/ClickHouse/ClickHouse/blob/master/CHANGELOG.md | 2020-09-14T19:02:00 | 2020-09-14T19:02:00 | {} | MEMBER |
ClickHouse | ClickHouse | 465,589,633 | 3,815 | jangorecki | Could we remove `queston` label and add `feature request` instead? I would find that useful to create scripts that can be easily customised for table name to query from.
In my use case I have following query:
```sql
SELECT id1, sum(v1) AS v1 FROM some_tbl GROUP BY id1;
```
And another one
```sql
SELECT query_dur... | 2019-02-20T14:08:55 | 2020-09-13T07:40:23 | {} | NONE |
ClickHouse | ClickHouse | 691,743,254 | 14,229 | bobrik | This commit balloons compilation times: 7c20aa2c621e31bcad50d4bc5eb2384d8a6286bd. Let's look at `DatabaseFactory.cpp` specifically.
Before:
```
10.25user 0.91system 0:11.17elapsed 99%CPU (0avgtext+0avgdata 911068maxresident)k
0inputs+152outputs (0major+235595minor)pagefaults 0swaps
```
After:
```
29.71u... | 2020-09-13T23:46:28 | 2020-09-13T23:46:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,080,771 | 14,737 | ghuname | > First of all your schema is quite suboptimal for your data. You have partition key `toDate(ipdrCreationTime)` and if I understand correctly in each insert you have a lot of different days? In this case, each insert will be split into many parts one for each partition and a lot of files will be created on disk.
Act... | 2020-09-14T14:14:21 | 2020-09-14T14:14:21 | {} | NONE |
ClickHouse | ClickHouse | 691,727,090 | 14,534 | filimonov | Related #14739 #9715 | 2020-09-13T21:20:09 | 2020-09-14T16:03:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,198,049 | 14,813 | alexey-milovidov | Anonymous namespace - :+1:
About inlining - didn't get the reasoning why it will help. | 2020-09-14T17:22:10 | 2020-09-14T17:22:10 | {} | MEMBER |
ClickHouse | ClickHouse | 692,228,122 | 14,750 | den-crane | Try to upgrade to 20.3.18.10 | 2020-09-14T18:18:49 | 2020-09-14T18:18:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 694,106,153 | 14,367 | doom369 | I have a similar issue (probably). after updating 20.3 to 20.6 - our usual queries are no longer executed and failed due to memory limits. After updating to 20.8.2.3 situation is even worth.
"Memory limit (total) exceeded: would use 961.69 MiB (attempt to allocate a chunk of 4676694 bytes), maximum: 886.67 MiB".
I up... | 2020-09-17T09:14:07 | 2020-09-17T09:18:30 | {} | NONE |
ClickHouse | ClickHouse | 694,162,568 | 9,926 | alexey-milovidov | #14886 | 2020-09-17T11:11:21 | 2020-09-17T11:11:21 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.