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 | 676,789,336 | 12,438 | akuzm | > @akuzm We still have a chance to improve performance of clang build.
> And to avoid overallocation...
I think I've merged all parts related to overallocation. I don't know about the clang performance, so I made a comparison with clang 10 in CI: https://clickhouse-test-reports.s3.yandex.net/13617/d293dfcee1dfb6b26... | 2020-08-19T22:51:43 | 2020-08-19T22:51:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 676,904,301 | 13,878 | hexiaoting | @hczhcz @filimonov
```
clickhouse-client -q "create table json_as_string(field String) Engine = Memory"
cat xxx| clickhouse-client -q "insert into json_as_string format JSONAsString"
```
now all my json data is stored in json_as_string table as a string column.
But How can I extract the values of "sentences.c... | 2020-08-20T03:03:27 | 2020-08-20T03:06:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 677,615,688 | 13,917 | alexey-milovidov | Should be fixed in #12550 ? | 2020-08-20T12:11:33 | 2020-08-20T12:11:33 | {} | MEMBER |
ClickHouse | ClickHouse | 677,726,081 | 13,190 | Slach | re-create as separate pull request after failure revert submodule changes
https://github.com/ClickHouse/ClickHouse/pull/13923 | 2020-08-20T15:12:36 | 2020-08-20T15:12:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 677,799,093 | 12,508 | alexey-milovidov | system.clusters data race: https://clickhouse-test-reports.s3.yandex.net/13880/ad30dcf434d24ca447c7533d7c258a7f0d8c67ca/stress_test_(thread)/stderr.log | 2020-08-20T17:29:40 | 2020-08-20T17:29:40 | {} | MEMBER |
ClickHouse | ClickHouse | 678,144,022 | 13,402 | excitoon | > LGTM
> Yet, I suggest that you rename `allow_merges` parameter to what has been proposed in the original issue - `disable_merges`, as it better reflects what the setting does.
> That can be done in a separate PR, of course.
Okay | 2020-08-21T09:36:02 | 2020-08-21T09:36:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 676,349,622 | 13,863 | azat | @vitlibar maybe worth adding a test for this? (to avoid further breakage) | 2020-08-19T13:22:48 | 2020-08-19T13:22:48 | {} | MEMBER |
ClickHouse | ClickHouse | 676,594,744 | 13,887 | KochetovNicolai | `01444_create_table_drop_database_race` - fixed in 30e0cd438ea0f383d9b5431d1acea7eb2fac5966
build check - build broken in master
AST fuzzer - reported #13893 | 2020-08-19T18:40:56 | 2020-08-19T18:40:56 | {} | MEMBER |
ClickHouse | ClickHouse | 676,963,589 | 13,878 | hczhcz | ```sql
with
'{
"id": 1,
"source": "china",
"sentences": [
{ "content": "I loved apples",
"words": [ {"content": "I", "stem": "i", "weight": 5},
{"content": "loved", "stem": "love", "weight": 10},
{"co... | 2020-08-20T03:44:04 | 2020-08-20T03:44:04 | {
"+1": 9,
"heart": 3,
"hooray": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 677,795,457 | 13,896 | alexey-milovidov | Functional test has been broken and fixed in master by @tavplubix | 2020-08-20T17:21:57 | 2020-08-20T17:21:57 | {} | MEMBER |
ClickHouse | ClickHouse | 677,913,075 | 13,930 | den-crane | Expected per documentation:
https://clickhouse.tech/docs/en/engines/table-engines/mergetree-family/collapsingmergetree/#table_engine-collapsingmergetree
_Algorithm_
....
_For each resulting data part ClickHouse **saves**:_
**The first “cancel” and the last “state” rows, if the number of “state” and “ca... | 2020-08-20T21:23:22 | 2020-08-20T21:24:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 677,922,513 | 13,931 | den-crane | >So, would it also be true then that materialized views won't see rows deleted because of TTL settings on the underlying table? That's also handled by the merge process, correct?
Correct.
>It sure seems that there needs to be a simple way to at least force a refresh of a materialized view without needing to comp... | 2020-08-20T21:45:19 | 2020-08-20T21:45:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 678,250,914 | 13,937 | amosbird | I'd suggest using `SYSTEM` queries for this kind of features. | 2020-08-21T11:49:49 | 2020-08-21T11:49:49 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 677,692,905 | 13,761 | abyss7 | Will take a look when it comes out of the draft then. | 2020-08-20T14:15:36 | 2020-08-20T14:15:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 677,893,336 | 13,880 | filimonov | Just a reminder / related - i think that can be merged https://github.com/ClickHouse/ClickHouse/pull/13648 | 2020-08-20T20:42:21 | 2020-08-20T20:42:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 677,990,659 | 3,685 | davidcsisk | I would like to see a simple ALTER command to force refresh of a Materialized View (versus having to drop and re-create the Materialized View). That ALTER...REFRESH command could then simply be cron'd. A Materialized View's content is wrong after an ALTER TABLE...DELETE/UPDATE is performed on the MV's base table (thi... | 2020-08-21T01:40:28 | 2020-08-21T01:44:03 | {} | NONE |
ClickHouse | ClickHouse | 678,280,531 | 12,508 | alexey-milovidov | 01396_inactive_replica_cleanup_nodes - https://github.com/ClickHouse/ClickHouse/pull/13906 | 2020-08-21T13:06:39 | 2020-08-21T13:06:39 | {} | MEMBER |
ClickHouse | ClickHouse | 676,352,139 | 13,726 | alexey-milovidov | It is needed for clickhouse-server, that's why we assign capabilities for clickhouse binary. | 2020-08-19T13:24:31 | 2020-08-19T13:24:31 | {} | MEMBER |
ClickHouse | ClickHouse | 676,759,038 | 12,508 | azat | > 01444_create_table_drop_database_race https://clickhouse-test-reports.s3.yandex.net/13887/25dd0fa4d67daaf0db8d0f16965a6bf1204b3666/functional_stateless_tests_(release).html#fail1 Fix: 30e0cd4
@tavplubix looks like it does not fix the test (https://clickhouse-test-reports.s3.yandex.net/13896/42dd661b6ad84f30869823a... | 2020-08-19T22:17:23 | 2020-08-19T22:17:23 | {} | MEMBER |
ClickHouse | ClickHouse | 677,997,366 | 3,685 | sundy-li | @davidcsisk
Seems it's like a ticked `materialized view`. We used to find it's hard to calculate realtime DAU in ClickHouse with materialized view.
`DAU` (daily active users) represents a distin... | 2020-08-21T02:07:18 | 2020-08-21T02:07:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 678,230,756 | 13,943 | heyciao | Sorry, I just noticed that the whole thing is already mentioned in the docs:
> Data that is inserted to a Buffer table may end up in the subordinate table in a different order and in different blocks. Because of this, a Buffer table is difficult to use for writing to a CollapsingMergeTree correctly. To avoid problem... | 2020-08-21T10:52:40 | 2020-08-21T10:52:40 | {} | NONE |
ClickHouse | ClickHouse | 677,394,297 | 8,706 | erikdubbelboer | This issue is related/similar to https://github.com/ClickHouse/ClickHouse/issues/6481 | 2020-08-20T07:11:48 | 2020-08-20T07:11:48 | {} | NONE |
ClickHouse | ClickHouse | 677,716,611 | 13,611 | KochetovNicolai | Not waiting for perftests. | 2020-08-20T14:56:34 | 2020-08-20T14:56:34 | {} | MEMBER |
ClickHouse | ClickHouse | 677,793,887 | 13,786 | alexey-milovidov | Cannot backport due to changes in Exception constructor. | 2020-08-20T17:18:40 | 2020-08-20T17:18:40 | {} | MEMBER |
ClickHouse | ClickHouse | 677,806,508 | 13,922 | alexey-milovidov | https://github.com/ClickHouse/ClickHouse/issues/13926 | 2020-08-20T17:45:15 | 2020-08-20T17:45:15 | {} | MEMBER |
ClickHouse | ClickHouse | 678,235,897 | 13,939 | filimonov | We were discussing that some time ago with @azat . Actually it is multithreaded already (since `copyData` function is multithreaded internally, i.e. reads several streams in parallel).
Pros:
- parallel commits (that is needed but can also be archived another way)
- several flushes in parallel (especially useful if... | 2020-08-21T11:07:23 | 2020-08-21T11:24:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 678,243,478 | 12,390 | alexey-milovidov | @etah000 Sorry, why close?
I want to merge it. | 2020-08-21T11:28:21 | 2020-08-21T11:28:21 | {} | MEMBER |
ClickHouse | ClickHouse | 678,244,806 | 13,942 | alexey-milovidov | > We can support both settings for a while, in order to achieve better compatibility!
No need for that. | 2020-08-21T11:32:09 | 2020-08-21T11:32:09 | {
"hooray": 1
} | MEMBER |
ClickHouse | ClickHouse | 677,831,881 | 10,341 | azat | > Maybe we can solve the task by introducing another QueryProcessingStage: WithMergeableStateAfterAggregation (it means that aggregate functions were calculated and finalized) and process it inside InterpreterSelectQuery?
@alexey-milovidov #10373 implements this | 2020-08-20T18:39:02 | 2020-08-20T18:39:02 | {} | MEMBER |
ClickHouse | ClickHouse | 677,902,167 | 13,924 | filimonov | https://github.com/ClickHouse/ClickHouse/issues/7663 ? | 2020-08-20T21:00:57 | 2020-08-20T21:00:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 678,197,967 | 13,942 | excitoon | Well, originally idea was to make negative setting, but while working on that I decided to name it `allow_merges`. This PR returns proposed behavior. We can support both settings for a while, in order to achieve better compatibility! | 2020-08-21T10:27:23 | 2020-08-21T10:27:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 678,297,007 | 13,688 | it1804 | I also downloaded the results of the ClickHouse Integration Tests (release) for PR # 13688 (https://clickhouse-test-reports.s3.yandex.net/13688/1006c4f11bce91b4fdf82f575f8f427828347805/integration_tests_(release)/test_dir.tar) And didn't see test_dictionaries_all_layouts_and_sources test was running. The dictionaries a... | 2020-08-21T13:41:54 | 2020-08-21T13:41:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 676,366,878 | 13,726 | filimonov | Are we sure that this code works as expected? And checked everytime?
https://github.com/ClickHouse/ClickHouse/blob/74edcfb87ef79f9c61d5829a9e9db665e5d0bce1/src/Common/hasLinuxCapability.cpp | 2020-08-19T13:34:18 | 2020-08-19T13:34:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 676,388,026 | 13,726 | filimonov | BTW - I don't care actually and it's not something important / the priority is 'very low'.
But it doesn't seem correct/expected behavior especially since we already have `hasLinuxCapability` checks. | 2020-08-19T13:48:22 | 2020-08-19T13:48:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 676,629,141 | 13,190 | blinkov | @Slach try `git submodule sync && git submodule update --init --recursive && git submodule foreach git reset --hard && git submodule foreach git clean -xfd`, but if it's already screwed up this might not help | 2020-08-19T19:54:33 | 2020-08-19T19:54:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 676,635,564 | 13,896 | azat | >ClickHouse build check — 19/20 builds are OK
unbundled build (should be fixed in #13895 ) | 2020-08-19T20:07:09 | 2020-08-19T20:07:09 | {} | MEMBER |
ClickHouse | ClickHouse | 677,487,073 | 13,842 | KochetovNicolai | `01444_create_table_drop_database_race` - flap, was fixed in 30e0cd438ea0f383d9b5431d1acea7eb2fac5966
build was broken in master | 2020-08-20T09:27:02 | 2020-08-20T09:27:02 | {} | MEMBER |
ClickHouse | ClickHouse | 678,189,670 | 13,942 | alexey-milovidov | And motivation is unclear (who ordered this change, why...) | 2020-08-21T10:21:10 | 2020-08-21T10:21:10 | {} | MEMBER |
ClickHouse | ClickHouse | 678,240,486 | 10,419 | woshishen9002 | I also has the same problem. I use
--input_format_allow_errors_num=10 --input_format_allow_errors_ratio=0.1
these two params, and my file only has one bad record(total 10 records). But it still can not pass , Why ??? | 2020-08-21T11:20:05 | 2020-08-21T11:20:05 | {} | NONE |
ClickHouse | ClickHouse | 676,320,462 | 13,726 | filimonov | If i will add needed capabilities - it works:
```
docker run --cap-add=SYS_NICE --cap-add=NET_ADMIN --cap-add=IPC_LOCK --rm -it -v/usr/bin/clickhouse:/clickhouse debian:stretch-slim /clickhouse local -q"select version()"
20.7.1.4186
```
The real question here is : why clickhouse-local require all that stuff? It ... | 2020-08-19T13:03:16 | 2020-08-19T13:03:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 676,553,912 | 13,894 | den-crane | Not a bug. Feature. As far as I remember there is an issue about a setting to enable automatic nullable aggregation results
```
SELECT SUM(cast(-1, 'Nullable(Int8)')) AS aggr
FROM system.one
WHERE 0
┌─aggr─┐
│ ᴺᵁᴸᴸ │
└──────┘
set empty_result_for_aggregation_by_empty_set=1;
SELECT SUM(-1) AS aggr FROM ... | 2020-08-19T17:16:02 | 2020-08-19T17:17:07 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 676,990,823 | 13,878 | hexiaoting | @hczhcz Thanks a lot . That's what I want. ^^ | 2020-08-20T04:02:51 | 2020-08-20T04:02:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 677,392,800 | 6,481 | erikdubbelboer | This is very dangerous. I just did an `ALTER TABLE table DELETE WHERE a = 'b' AND date >= '18-08-2020';` which delete much more data than expected. I know that for parsing speed wrong dates don't throw errors but it would be so much better if for constants in queries it did throw an error. Here speed doesn't matter as ... | 2020-08-20T07:10:54 | 2020-08-20T07:10:54 | {
"+1": 2
} | NONE |
ClickHouse | ClickHouse | 677,640,694 | 13,917 | myrrc | > Should be fixed in #12550 ?
Yup. | 2020-08-20T12:41:26 | 2020-08-20T12:41:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 677,717,076 | 13,027 | millin | @alexey-milovidov Will this be fixed or are named tuples obsolete? | 2020-08-20T14:57:24 | 2020-08-20T14:57:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 677,725,401 | 13,923 | Slach | @blinkov could you look at this PR instead of #13190? | 2020-08-20T15:11:31 | 2020-08-20T15:11:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 678,210,328 | 13,942 | Akazz | This PR is supposed to be a followup to #13402, where this setting (`allow_merges`) is initially introduced. As it is more consistent with the original motivation in #10650 | 2020-08-21T10:36:50 | 2020-08-21T10:40:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 678,255,907 | 13,864 | alexey-milovidov | @filimonov @CurtizJ The difference is that `IDataType::enumerateStreams` called more times. | 2020-08-21T12:04:06 | 2020-08-21T12:04:06 | {} | MEMBER |
ClickHouse | ClickHouse | 676,319,944 | 1,665 | Axelfoley85 | Still got the same issue on clickhouse 19.15.3.6 when importing TabSeparated data. How can I turn on readFloatTextPrecise?
```
cat /tmp/dump.txt
1560038400 3600 12139 12369 foo 15 bar 12345 1236 3.397 \N
cat /tmp/dump.txt | clickhouse-client -q "insert into tmp.testTable format TabSeparated"
clickhouse-clien... | 2020-08-19T13:02:57 | 2020-08-19T13:02:57 | {} | NONE |
ClickHouse | ClickHouse | 676,480,192 | 13,890 | wangxi19 | Hope you get a glimpse of me | 2020-08-19T15:00:34 | 2020-08-19T15:00:34 | {} | NONE |
ClickHouse | ClickHouse | 678,185,681 | 13,928 | alexey-milovidov | TSan report - is it fixed in #13624 ? @nikitamikhaylov | 2020-08-21T10:18:13 | 2020-08-21T10:18:13 | {} | MEMBER |
ClickHouse | ClickHouse | 678,186,809 | 13,928 | nikitamikhaylov | > TSan report - is it fixed in #13624 ? @nikitamikhaylov
Yes. | 2020-08-21T10:19:04 | 2020-08-21T10:19:04 | {} | MEMBER |
ClickHouse | ClickHouse | 678,245,604 | 13,938 | alexey-milovidov | You need to update submodules after
`git checkout tags/v20.6.4.44-stable`:
`git submodule update` | 2020-08-21T11:34:33 | 2020-08-21T11:34:33 | {} | MEMBER |
ClickHouse | ClickHouse | 676,530,890 | 13,190 | Slach | @blinkov I don't know how submodules were changed and pushed to my fork
I always get it only via `git pull --recurse-submodules upstream master`
could you help me revert submodules changes?
when I tried to do it in CLion
),and the problem will occur.
| 2020-08-20T02:42:47 | 2020-08-20T02:42:47 | {} | NONE |
ClickHouse | ClickHouse | 677,744,909 | 11,298 | melin | ColumnExprTupleAccess failure
| 2020-08-20T15:45:05 | 2020-08-20T15:45:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 677,908,003 | 13,931 | den-crane | Materialized view (MV) is a post-insert trigger. Insert to a source table pushes an inserted buffer to MV as well.
MV does select over the inserted buffer (MV never reads the source table except populate stage).
MV does not see `alter update/delete/drop/truncate`.
MV does not see changes from a merge process colla... | 2020-08-20T21:12:53 | 2020-08-20T21:16:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 678,019,265 | 13,906 | tomjiang1987 | > Also need to add a test.
ok, done | 2020-08-21T03:35:17 | 2020-08-21T03:35:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 678,244,585 | 13,942 | alexey-milovidov | Ok.
I mixed it with another setting named "can_be_leader". | 2020-08-21T11:31:34 | 2020-08-21T11:31:34 | {} | MEMBER |
ClickHouse | ClickHouse | 678,282,133 | 13,761 | zhang2014 | @abyss7 Can you help review this PR?
| 2020-08-21T13:10:00 | 2020-08-21T13:10:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 676,556,208 | 13,894 | qoega | Nice. But actually can be very confusing when it encounters somewhere inside part of UNION ALL.
For ANSI compatibility mode we will have to enable it. | 2020-08-19T17:20:56 | 2020-08-19T17:20:56 | {} | MEMBER |
ClickHouse | ClickHouse | 676,614,040 | 12,438 | alexey-milovidov | @akuzm We still have a chance to improve performance of clang build.
And to avoid overallocation... | 2020-08-19T19:22:23 | 2020-08-19T19:22:23 | {} | MEMBER |
ClickHouse | ClickHouse | 676,829,368 | 13,880 | bharatnc | > There is one slight deficiency in this implementation.
>
> So, the list of timezones is generated at release time and committed to sources.
> But when we build ClickHouse without a release, the list of timezones actually used may be different.
> Instead, we can obtain the list of timezones during the build and g... | 2020-08-20T00:34:22 | 2020-08-20T00:34:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 677,801,758 | 13,906 | alexey-milovidov | Also need to add a test. | 2020-08-20T17:35:09 | 2020-08-20T17:35:09 | {} | MEMBER |
ClickHouse | ClickHouse | 678,129,948 | 13,402 | Akazz | This PR definitely could not have affected the failed performance test - [constant_column_search](https://clickhouse-test-reports.s3.yandex.net/13402/16d709964bef99d85dd5eb2c8cdc46669917a0a7/performance_comparison/report.html#fail1) as well as `array_element` | 2020-08-21T09:03:56 | 2020-08-21T09:03:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 676,121,569 | 9,624 | KochetovNicolai | Skip test `01412_cache_dictionary_race` because it causes exception `Directory not empty` while executing `DROP DATABASE IF EXISTS ordinary_db`. This should be fixed in #13756, but was not backported because of conflicts. | 2020-08-19T10:40:18 | 2020-08-19T10:40:18 | {} | MEMBER |
ClickHouse | ClickHouse | 676,361,256 | 13,726 | filimonov | it should be possible to detect/report the lack of capabilities (or even opt-out the functionality).
```
$ docker run --rm -it -v/usr/bin/clickhouse:/clickhouse debian:stretch-slim bash
root@0d4ba34acb09:/# apt-get update > /dev/null && apt-get -y install libcap-ng-utils >/dev/null
debconf: delaying package configu... | 2020-08-19T13:30:34 | 2020-08-19T13:30:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 676,841,097 | 11,059 | compasses | @templarzq Waiting you sharing your story about optimize of merge algorithm of replacingmerge tree 👍 :) | 2020-08-20T01:20:33 | 2020-08-20T01:20:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 677,754,108 | 13,864 | filimonov | It was introduced with polymorphic-parts (PR #8290)
commit bdf41b7c1fe5f4b97833d1efb7b86ba9bff11840 (master before polymorphic-parts)
-с 1 "QPS": 575.5985304060765,
-с 6 "QPS": 470.3612387444902,
-с 12 "QPS": 273.8655580570344,
commit b736029828748b9bfd2e2d3f16ef9f250268b841 (PR #8290)
-с 1 "QPS": 368.2938... | 2020-08-20T16:01:44 | 2020-08-20T16:01:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 677,789,817 | 12,508 | qoega | https://clickhouse-test-reports.s3.yandex.net/13647/e77ab608c8e579caca7131cc2036dbac3d32e582/stress_test_(thread).html#fail1
and
https://clickhouse-test-reports.s3.yandex.net/13928/99b31480fd18b04677ba74bad21e5b3d7d928312/stress_test_(thread).html#fail1
stress test ThreadSanitizer: data race
#1 DB::Cache... | 2020-08-20T17:10:04 | 2020-08-21T09:35:02 | {} | MEMBER |
ClickHouse | ClickHouse | 677,489,751 | 11,059 | templarzq | new version clickhouse final query is fast enough. | 2020-08-20T09:32:45 | 2020-08-20T09:32:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 677,641,188 | 12,550 | myrrc | Resolves #13917 | 2020-08-20T12:42:29 | 2020-08-20T12:42:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 677,717,559 | 13,869 | KochetovNicolai | Looks like we also need `kvm` and `elf` at least | 2020-08-20T14:58:12 | 2020-08-20T14:58:12 | {} | MEMBER |
ClickHouse | ClickHouse | 677,919,487 | 13,931 | davidcsisk | Thank you for the quick reply. So, would it also be true then that materialized views won't see rows deleted because of TTL settings on the underlying table? That's also handled by the merge process, correct? It sure seems that there needs to be a simple way to at least force a refresh of a materialized view without ... | 2020-08-20T21:38:34 | 2020-08-20T21:38:34 | {} | NONE |
ClickHouse | ClickHouse | 678,280,309 | 13,906 | alexey-milovidov | > 01396_inactive_replica_cleanup_nodes
I will merge and report this case of flaky test. | 2020-08-21T13:06:12 | 2020-08-21T13:06:12 | {} | MEMBER |
ClickHouse | ClickHouse | 678,296,552 | 13,688 | it1804 | The test_dictionaries_all_layouts_and_sources test for redis source is actually broken. I figured out why, but I don't know how to fix it. The test creates simultaneously several dictionaries for redis source with one attribute, which differ only in the type of this attribute. In the redis database, only two unique key... | 2020-08-21T13:40:58 | 2020-08-21T13:40:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 676,418,816 | 13,874 | mjodeh | Actually this is so misleading, our use case is to calculate the average time between steps in a funnel query and this unexpected behavior led to wrong calculations as a work around we have to check the length of the output then decide what to do.
On the other hand - correct if I'm wrong - having columns with `Nullabl... | 2020-08-19T14:08:42 | 2020-08-19T14:08:42 | {} | NONE |
ClickHouse | ClickHouse | 676,594,966 | 13,894 | alexey-milovidov | @qoega `empty_result_for_aggregation_by_empty_set` does not make behaviour more standard compliant.
See №5 of https://github.com/ClickHouse/ClickHouse/issues/10540 | 2020-08-19T18:41:24 | 2020-08-19T18:41:24 | {} | MEMBER |
ClickHouse | ClickHouse | 676,755,821 | 13,895 | azat | I've removed the patches for the UInt128.h, since it is required only for gcc10+ (and by required I mean it breaks gcc9), will submit all gcc10 fixes later | 2020-08-19T22:14:59 | 2020-08-19T22:15:16 | {} | MEMBER |
ClickHouse | ClickHouse | 677,491,016 | 13,888 | hczhcz | I would suggest a normal function name like `dateTrunc` + a case insensitive alias like `DATE_TRUNC`. | 2020-08-20T09:35:45 | 2020-08-20T09:35:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 678,287,802 | 13,869 | urgordeadbeef | libprocstat is base system library and it is FreeBSD specific. What are you doing?!!! | 2020-08-21T13:22:05 | 2020-08-21T13:22:05 | {} | NONE |
ClickHouse | ClickHouse | 678,317,971 | 13,862 | UnamedRus | flamegraph from debug build
| 2020-08-21T14:24:49 | 2020-08-21T14:24:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 676,556,560 | 13,894 | qoega | And `empty_result_for_aggregation_by_empty_set` is currently undocumented! | 2020-08-19T17:21:39 | 2020-08-19T17:21:39 | {} | MEMBER |
ClickHouse | ClickHouse | 677,279,710 | 13,894 | hczhcz | It is `max( A U B )` vs `max( { max(A), max(B) } )`. This behavior is useful especially when the operator does not observe the law of association. | 2020-08-20T06:26:52 | 2020-08-20T06:26:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 677,489,067 | 13,611 | KochetovNicolai | 01444_create_table_drop_database_race - was broken in master, fixed
build - broken in master
| 2020-08-20T09:31:17 | 2020-08-20T09:31:17 | {} | MEMBER |
ClickHouse | ClickHouse | 677,604,975 | 13,624 | nikitamikhaylov | Fuzzer error reported here #13917 | 2020-08-20T12:02:36 | 2020-08-20T12:02:36 | {} | MEMBER |
ClickHouse | ClickHouse | 677,795,076 | 12,076 | alexey-milovidov | Please make a new PR if we will continue. | 2020-08-20T17:21:12 | 2020-08-20T17:21:12 | {} | MEMBER |
ClickHouse | ClickHouse | 677,798,215 | 13,880 | alexey-milovidov | Looks like you forgot to commit test reference file. | 2020-08-20T17:27:48 | 2020-08-20T17:27:48 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 676,491,704 | 13,761 | zhang2014 | Need support:
- [x] parseDateTime32BestEffort
- [ ] ~parseDateTime32BestEffortUS~
- [x] parseDateTime32BestEffortOrNull
- [x] parseDateTime32BestEffortOrZero | 2020-08-19T15:20:02 | 2020-08-21T05:10:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 678,203,281 | 13,906 | tomjiang1987 | @tavplubix @alexey-milovidov 01396_inactive_replica_cleanup_nodes seem unrelated to this PR, how to rerun the case? | 2020-08-21T10:31:28 | 2020-08-21T10:31:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 678,208,180 | 13,930 | heyciao | Thanks a lot, you're right (I read that already in the past but I really "understood" it only now).
Then, when that happens, what is the user supposed to do to make the system collapse those rows?
I see that by inserting some additional identical rows (not sure about the order) the system does end up collapsing the... | 2020-08-21T10:35:13 | 2020-08-21T10:35:13 | {} | NONE |
ClickHouse | ClickHouse | 678,247,210 | 13,930 | heyciao | Is my understanding correct if I say that the engine "VersionedCollapsingMergeTree" does not have this limitation (order of inserts can by anything)? | 2020-08-21T11:39:18 | 2020-08-21T11:39:18 | {} | NONE |
ClickHouse | ClickHouse | 678,326,993 | 8,841 | alexey-milovidov | Discussion completed. | 2020-08-21T14:42:17 | 2020-08-21T14:42:17 | {} | MEMBER |
ClickHouse | ClickHouse | 676,369,804 | 13,033 | vladimir-golovchenko | Related PR 'Added date_trunc function' #13888 | 2020-08-19T13:36:16 | 2020-08-19T13:36:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 677,666,992 | 13,820 | bohutang | @alexey-milovidov may i have a review round for this PR? | 2020-08-20T13:32:38 | 2020-08-20T13:32:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 677,797,913 | 13,880 | alexey-milovidov | > Do you recommend a work around here or this sounds good to you?
I think it's Ok. | 2020-08-20T17:27:12 | 2020-08-20T17:27:12 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 678,004,692 | 2,156 | DouMiaoO-Oo | > @raulgregg could you please, open a separate issue?
> Also please add query and table scheme.
@KochetovNicolai Hello, I just opened a issue https://github.com/ClickHouse/ClickHouse/issues/13936 | 2020-08-21T02:35:48 | 2020-08-21T02:35:48 | {} | NONE |
ClickHouse | ClickHouse | 678,097,313 | 13,864 | filimonov | Inside branch - the last good commit is 42c9ea9aa352fc8388ee6c8805e4fb0d7d4f2b9a (moment of forking from master).
![42c9ea9aa352fc8388ee6c8805e4fb0d7d4f2b9a.svg](https://raw.githubusercontent.com/gist/filimonov/48bf4e6b8b34b7596578507bbe750bb2/raw/8fba1ec5ad6a449344da4fa0ec9aad460535c1c6/42c9ea9aa352fc8388ee6c8805e4... | 2020-08-21T07:46:34 | 2020-08-21T07:46:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 678,831,434 | 8,367 | alexey-milovidov | I have checked manually that the queries from "linear_regression" performance test work:
```
ClickHouse client version 20.8.1.1.
Connecting to localhost:9000 as user default.
Connected to ClickHouse server version 20.8.1 revision 54438.
milovidov-desktop :) DROP TABLE IF EXISTS test_model
DROP TABLE IF EXIS... | 2020-08-23T22:14:02 | 2020-08-23T22:14:02 | {} | MEMBER |
ClickHouse | ClickHouse | 679,098,109 | 13,978 | alexey-milovidov | > If I understand correctly, "uncommitted" parts can be merged only with other "uncommitted" parts?
Yes.
> Using current EphemeralLockInZooKeeper or just incrementing number in /partition_id node? If second it's not clear how to make it atomic.
Second. How to make it atomic: read current number, atomically com... | 2020-08-24T12:31:22 | 2020-08-24T12:32:43 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.