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 | 476,324,427 | 4,797 | ghost | > But functions in ClickHouse are invoked for vectors, not for single values. The result will still be constant for a single block of data.
Yes, so returning to the example with Kafka it means that multiple consecutive rows can still have single value of the function.
As a proposal: may be it would be a useful fe... | 2019-03-25T18:32:23 | 2019-03-25T18:32:23 | {} | NONE |
ClickHouse | ClickHouse | 476,360,139 | 4,790 | alexey-milovidov | Missing empty `.reference` file. | 2019-03-25T20:17:04 | 2019-03-25T20:17:04 | {} | MEMBER |
ClickHouse | ClickHouse | 476,447,805 | 4,794 | zhang2014 | @4ertus2 OK, I will try to understand and solve the problem you said : ) | 2019-03-26T02:18:57 | 2019-03-26T02:18:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 476,595,754 | 3,413 | artfabrique | @blinkov also there is one more condition that seems not feasible on using present-day tech.
Here is a real usecase scenario:
Imagine you've got this "soft funnel" and showed it to user
1. A
1.1. X: occurs in 70% of transitions, 700 times per transition A>B
1.2. Y: occurs in 20% of transitions, 200 times per t... | 2019-03-26T12:03:27 | 2019-03-26T20:09:29 | {} | NONE |
ClickHouse | ClickHouse | 477,440,690 | 4,817 | kindred77 | Hi, I got another error:
../contrib/boost/boost/system/error_code.hpp:461: error: undefined reference to 'boost::system::system_category()'
../contrib/boost/boost/system/error_code.hpp:491: error: undefined reference to 'boost::system::system_category()'
../contrib/boost/boost/system/error_code.hpp:733: error: und... | 2019-03-28T04:12:28 | 2019-03-28T04:12:28 | {} | NONE |
ClickHouse | ClickHouse | 477,485,875 | 3,822 | blinkov | @josepowera @lyl2008dsg @likeSakura any idea on how to reproduce this? | 2019-03-28T07:48:20 | 2019-03-28T07:48:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 476,575,381 | 3,096 | blinkov | @fizerkhan is this issue still relevant? | 2019-03-26T10:57:57 | 2019-03-26T10:57:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 476,827,310 | 4,218 | 4ertus2 | #4815 | 2019-03-26T20:08:18 | 2019-03-26T20:08:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 477,549,226 | 3,994 | PhantomPhreak | Yes, i tried. It works, so i solved my problem in following way:
1. Created replicated tables with the same schema as `system.query_log` table, but with additional column `hostname`
2. Created materialized view consuming from local `system.query_log` tables, adding `hostname` field as result of calling `hostname()` ... | 2019-03-28T11:04:22 | 2019-03-28T11:04:41 | {} | NONE |
ClickHouse | ClickHouse | 477,723,228 | 4,820 | nvartolomei | @alexey-milovidov sounds like it ;) | 2019-03-28T18:43:13 | 2019-03-28T18:43:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 476,433,633 | 4,789 | kindred77 | It works, thanks. | 2019-03-26T01:01:46 | 2019-03-26T01:01:46 | {} | NONE |
ClickHouse | ClickHouse | 476,495,742 | 4,070 | Crazylearner30 | Thanks | 2019-03-26T06:43:40 | 2019-03-26T06:43:40 | {} | NONE |
ClickHouse | ClickHouse | 476,546,081 | 4,799 | simPod | @proller my bad, should be k now | 2019-03-26T09:40:42 | 2019-03-26T09:40:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 476,578,762 | 2,555 | blinkov | > On new version i wasn't able to reproduce the crashes, so I'm not sure if it's worth to debug it deeper.
It doesn't make sense to keep this open since we don't support builds from 2018 anymore. | 2019-03-26T11:06:56 | 2019-03-26T11:07:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 476,802,731 | 4,763 | zloidemon | Cmake output:
```
%> cmake .. -DCMAKE_CXX_COMPILER=`which g++-8` -DCMAKE_C_COMPILER=`which gcc-8`
-- The C compiler identification is GNU 8.3.0
-- The CXX compiler identification is GNU 8.3.0
-- Checking whether C compiler has -isysroot
-- Checking whether C compiler has -isysroot - yes
-- Checking whether C c... | 2019-03-26T19:06:10 | 2019-03-26T19:06:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 476,996,854 | 4,761 | filimonov | Each separate insert to MergeTree creates one part stored as a separate directory on filesystem.
So the only acceptable way of inserts to MergeTree is bulk inserts.
Buffer engine definitely works in memory and allows to form a bigger parts, combining multiple smaller inserts into one bigger.
Merges will be ex... | 2019-03-27T06:27:02 | 2019-03-27T06:27:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 476,579,567 | 777 | Slach | wow guys, thanks you for your efforts, i can try actualize and https://github.com/kshvakov/zetcd fork with my https://github.com/Slach/clickhouse-zetcd, i'll report here about results | 2019-03-26T11:09:10 | 2019-03-26T11:09:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 476,623,852 | 4,738 | igor-sh8 | Hi.
Thanks for the info.
Are there some best practices for creating buffer tables ?
If I have server with 20 cpu cores and I need to create two high load buffer tables (insert 300 rows per second). What value of num_layers do I need to set ? For example 16 layers on the first and 16 layers on the second table - is... | 2019-03-26T13:11:36 | 2019-03-26T13:11:36 | {} | NONE |
ClickHouse | ClickHouse | 476,549,497 | 4,800 | 4ertus2 | related #4681
Need retry with #4815 | 2019-03-26T09:49:25 | 2019-03-26T20:10:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 476,506,731 | 4,402 | filimonov | With asan:
```
=================================================================
==4176==ERROR: AddressSanitizer: stack-buffer-overflow on address 0x7fb30f2ca000 at pc 0x000006b87649 bp 0x7fb36b4ac410 sp 0x7fb36b4abbc0
READ of size 1 at 0x7fb30f2ca000 thread T3
#0 0x6b87648 in write (/usr/bin/clickhouse+0x6b... | 2019-03-26T07:30:40 | 2019-03-26T07:30:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 476,579,841 | 3,659 | blinkov | @754154377 do you have any further questions? | 2019-03-26T11:10:02 | 2019-03-26T11:10:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 476,638,471 | 777 | Slach | @kshvakov yeah bro, it's live ;) i commited latest changes for clickhouse-zetcd | 2019-03-26T13:36:02 | 2019-03-26T13:36:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 476,697,602 | 4,674 | smirnov-vs | Плюс я не могу контролировать какими пачками будет приходить файл на сервер во время теста. Ошибка была из-за того, что в буфере хранилась только часть сообщения (такой пакет пришел по сети), а чтение было реализовано так, будто мы можем прочесть всё за один раз.
Теперь я читаю ровно столько сколько надо для каждого с... | 2019-03-26T15:17:56 | 2019-03-26T15:17:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 477,443,185 | 4,800 | ChannelerH | @4ertus2 I have upgraded clickhouse version to 4.1.0. Then everything is ok. Thanks
| 2019-03-28T04:27:59 | 2019-03-28T04:27:59 | {} | NONE |
ClickHouse | ClickHouse | 477,669,119 | 4,774 | Gladdy | @4ertus2 Thanks very much for your time in leaving some very useful feedback! I'll reply to the individual points inline and try to fix them up as well as possible tonight. | 2019-03-28T16:23:46 | 2019-03-28T16:23:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 477,709,941 | 4,774 | Gladdy | @4ertus2 The current approach introduces quite a lot of if constexprs into the Join class - what is your option on rather than having these conditionals, templating in different implementations of the core HashTable (similar-ish to the 2-stage hash table, but then with a normal HT at the base followed by a BST for the ... | 2019-03-28T18:05:59 | 2019-03-28T18:05:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 476,580,008 | 1,432 | RoyBellingan | @blinkov this bug is related to a now quite old version.
I have not checked with the lasted neither if is happening again.
As said Date is a materialized column from the timestamp, that is the only thing I can say.
| 2019-03-26T11:10:36 | 2019-03-26T11:10:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 476,645,920 | 4,674 | alexey-milovidov | Посмотрите тест `00823_capnproto_input.sh`. Он создаёт временный файл со схемой в текущей директории.
Требование на директорию `format_schemas` есть только в случае, если с форматом работает сервер. Клиент может брать схему из любого файла, для которого есть доступ. | 2019-03-26T13:47:02 | 2019-03-26T13:47:02 | {} | MEMBER |
ClickHouse | ClickHouse | 476,655,494 | 4,674 | smirnov-vs | По сути мой тест сделает аналогичное покрытие, как и предложенный. Стоит ли тогда придумывать что-то еще?) | 2019-03-26T14:01:05 | 2019-03-26T14:01:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 477,182,327 | 157 | Abzac | @alexey-milovidov
Wow! Great!
But how can it be used to get user agent info from User-Agent string?
Because search* function and match* filter functions are not very useful in this case. The main problem is you need to have some mapping in memory with UA data to effectively use it. So unfortunately you can not j... | 2019-03-27T14:31:24 | 2019-03-27T14:54:27 | {} | NONE |
ClickHouse | ClickHouse | 477,206,728 | 157 | alexey-milovidov | There is `multiMatchAnyIndex` function that will return an index of matching regexp in array of regexps (starting from 1; or zero if no regexp matched).
Yes, I understand that amount of regexps will be huge for User-Agents. They also are very ambiguous (MSIE, Mozilla, like Gecko...) Maybe this function is not applic... | 2019-03-27T15:22:45 | 2019-03-27T15:22:45 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 477,356,947 | 4,819 | alexey-milovidov | Our floating point parser does not always return the nearest machine representable number (this is documented behaviour). Details are here: https://github.com/yandex/ClickHouse/blob/0eed697623afd251c1908073c4a2850a9bfc9cdc/dbms/src/IO/readFloatText.h#L9:L90
Though, we still have a chance to implement a setting to dy... | 2019-03-27T21:36:03 | 2019-03-27T21:38:07 | {} | MEMBER |
ClickHouse | ClickHouse | 477,681,252 | 4,047 | nvartolomei | @alesapin any clues why would this happen? I'm pretty confident that this is related to connection pooling and getting a closed/close-wait connection from the pool :) | 2019-03-28T16:53:22 | 2019-03-28T16:53:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 476,566,400 | 4,799 | filimonov | That part from [release notes](https://github.com/edenhill/librdkafka/releases/tag/v1.0.0) looks important (as i understand - no actions needed, but looks like that change can bring some improvements).
> ### `max.poll.interval.ms` is enforced
>
> This release adds support for `max.poll.interval.ms` ([KIP-62](htt... | 2019-03-26T10:35:33 | 2019-03-26T10:42:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 476,576,174 | 3,413 | blinkov | @artfabrique is this issue still relevant? | 2019-03-26T10:59:58 | 2019-03-26T10:59:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 476,585,770 | 3,413 | artfabrique | @blinkov Yes! We stuck completely. | 2019-03-26T11:29:42 | 2019-03-26T11:31:54 | {} | NONE |
ClickHouse | ClickHouse | 476,728,838 | 4,674 | alexey-milovidov | Ok, спасибо! | 2019-03-26T16:21:17 | 2019-03-26T16:21:17 | {} | MEMBER |
ClickHouse | ClickHouse | 476,745,621 | 4,642 | alexey-milovidov | Failed. | 2019-03-26T16:56:43 | 2019-03-26T16:56:43 | {} | MEMBER |
ClickHouse | ClickHouse | 477,207,900 | 4,821 | alexey-milovidov | I prefer this field in both query_log and query_thread_log for consistency. | 2019-03-27T15:25:11 | 2019-03-27T15:25:11 | {} | MEMBER |
ClickHouse | ClickHouse | 477,362,782 | 4,629 | alexey-milovidov | The failing test is not related to your changes. | 2019-03-27T21:56:30 | 2019-03-27T21:56:30 | {} | MEMBER |
ClickHouse | ClickHouse | 476,602,552 | 4,674 | smirnov-vs | @alexey-milovidov
После серии тестов я переписал чтение согласно протоколу capnproto, предыдущий код читал, мягко говоря, не очень. Теперь моя сборка заливает данные стабильно.
Я могу написать тест с надуманной схемой и рандомными данными, файл можно засунуть прямо в скрипт, но схему надо засунуть в папку format_sch... | 2019-03-26T12:25:48 | 2019-03-26T12:25:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 476,640,744 | 4,738 | alexey-milovidov | Two layers is enough for at least 1000 inserts per second.
There are very few cases when you need more layers. More layers will lead to higher memory usage and more frequent inserts to the target table. | 2019-03-26T13:39:48 | 2019-03-26T13:39:48 | {} | MEMBER |
ClickHouse | ClickHouse | 476,744,926 | 4,671 | alexey-milovidov | Now the implementation looks perfect. Thank you! | 2019-03-26T16:55:22 | 2019-03-26T16:55:22 | {} | MEMBER |
ClickHouse | ClickHouse | 477,594,199 | 4,546 | adubovikov | thanks. Looks like it will fix it :-) | 2019-03-28T13:27:18 | 2019-03-28T13:27:18 | {} | NONE |
ClickHouse | ClickHouse | 477,595,782 | 4,671 | adubovikov | great job! | 2019-03-28T13:31:33 | 2019-03-28T13:31:33 | {} | NONE |
ClickHouse | ClickHouse | 477,761,143 | 4,774 | Gladdy | @4ertus2 Thanks very much, have merged master and am fixing the issues locally before re-pushing. | 2019-03-28T20:36:13 | 2019-03-28T20:36:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 476,306,673 | 4,797 | alexey-milovidov | > IMHO now() is a constant per query.
Yes. It is documented. now() is a constant expression and it allows to use index for expressions like `x >= now() - INTERVAL 1 MINUTE`
We can provide a separate function that will return precise time of invocation. But functions in ClickHouse are invoked for vectors, not for ... | 2019-03-25T17:44:55 | 2019-03-25T17:44:55 | {} | MEMBER |
ClickHouse | ClickHouse | 476,576,449 | 3,250 | blinkov | @nadarkathiresh have you tried switching to MergeTree engine? | 2019-03-26T11:00:44 | 2019-03-26T11:00:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 476,736,081 | 4,790 | alexey-milovidov | > This seems to have nothing to do with my changes.
True. | 2019-03-26T16:37:24 | 2019-03-26T16:37:24 | {} | MEMBER |
ClickHouse | ClickHouse | 476,739,121 | 4,713 | alexey-milovidov | Still waiting for test fix. | 2019-03-26T16:43:27 | 2019-03-26T16:43:27 | {} | MEMBER |
ClickHouse | ClickHouse | 476,750,051 | 4,439 | alexey-milovidov | Still not ready. | 2019-03-26T17:05:51 | 2019-03-26T17:05:51 | {} | MEMBER |
ClickHouse | ClickHouse | 476,812,872 | 157 | alexey-milovidov | Hyperscan has been added: https://github.com/yandex/ClickHouse/pull/4780 | 2019-03-26T19:32:01 | 2019-03-26T19:32:01 | {
"rocket": 2
} | MEMBER |
ClickHouse | ClickHouse | 477,207,531 | 4,821 | alexey-milovidov | > query_start_time_microseconds
I'm Ok with this variant (separate column with microseconds part). | 2019-03-27T15:24:26 | 2019-03-27T15:25:42 | {} | MEMBER |
ClickHouse | ClickHouse | 477,512,051 | 4,218 | 4ertus2 | fixed in master | 2019-03-28T09:14:56 | 2019-03-28T09:14:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 477,731,965 | 4,774 | 4ertus2 | > @4ertus2 The current approach introduces quite a lot of if constexprs into the Join class - what is your option on rather than having these conditionals, templating in different implementations of the core HashTable (similar-ish to the 2-stage hash table, but then with a normal HT at the base followed by a BST for th... | 2019-03-28T19:07:56 | 2019-03-28T19:09:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 476,575,968 | 1,432 | blinkov | @RoyBellingan do you have any further questions? | 2019-03-26T10:59:29 | 2019-03-26T10:59:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 477,263,224 | 4,047 | nvartolomei | Hitting same exception with v19.1.15 replicating from v18.14.17. Node from which replication is tried does not log anything. ~2% failure rate. Eventually it succeeds.
This might be due to connection pooling which was added since then and trying to use a connection which was closed by the server? | 2019-03-27T17:18:14 | 2019-03-27T20:15:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 477,358,505 | 4,820 | alexey-milovidov | Prefer first replica in order, and if it is unavailable, loop through all other replicas at random? | 2019-03-27T21:41:33 | 2019-03-27T21:41:33 | {} | MEMBER |
ClickHouse | ClickHouse | 477,359,590 | 4,797 | alexey-milovidov | > As a proposal: may be it would be a useful feature to expose Kafka messages metadata as special functions for tables with Kafka engine, so that it is would be possible to explicitly know offset/partition for each row that came as a Kafka message.
Yes, it is technically possible to implement with "virtual columns" ... | 2019-03-27T21:45:16 | 2019-03-27T21:45:37 | {} | MEMBER |
ClickHouse | ClickHouse | 477,488,232 | 3,822 | likeSakura | My approach is simple.Delete the merge table and distributed table that are being merged..Then restart.But this is only a temporary solution.As for the long-term solution and optimization solution, may involve clickhouse's underlying source code and its merger engine.I still don't understand the ground floor@blinkov | 2019-03-28T07:57:03 | 2019-03-28T08:05:44 | {} | NONE |
ClickHouse | ClickHouse | 477,610,482 | 4,796 | KochetovNicolai | I will try to comment your result.
First of all, let me briefly describe `LowCardinality` structure. ClickHouse process data by blocks. Block is a set of columns. But block can't store all data from column. Instead, it stores parts of each column (for example, first 8192 values). This column part can be stored as a ... | 2019-03-28T14:08:41 | 2019-03-28T14:08:41 | {
"+1": 6
} | MEMBER |
ClickHouse | ClickHouse | 477,614,071 | 4,796 | KochetovNicolai | Also, your test for insertion is not completely fair, because for LCDict you read from another table, but for Dict - not. Anyway, insertion to LCDict is slower indeed. What I have tested:
```sql
:) create table Dict2 as Dict
CREATE TABLE Dict2 AS Dict
Ok.
0 rows in set. Elapsed: 0.003 sec.
:) INSERT INT... | 2019-03-28T14:17:33 | 2019-03-28T14:17:33 | {} | MEMBER |
ClickHouse | ClickHouse | 477,712,098 | 4,831 | alex-zaitsev | Could be related to #4215 | 2019-03-28T18:11:51 | 2019-03-28T18:11:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 477,735,108 | 4,724 | alexey-milovidov | You need to fix builds. You can find the relevant info in "Details" link:
https://clickhouse-builds.s3.yandex.net/4724/32c338556c9db19faba6e3a9c7edf7d3b26ce03a/report.html | 2019-03-28T19:17:22 | 2019-03-28T19:17:22 | {} | MEMBER |
ClickHouse | ClickHouse | 477,596,270 | 2,551 | mcarbonneaux | cmake must be ajusted to compile libsasl with librdkafka....or link with os sasl...
I'm working on pull request about that... | 2019-03-28T13:32:54 | 2019-03-28T21:41:46 | {} | NONE |
ClickHouse | ClickHouse | 476,317,698 | 4,797 | ghost | Yes, `now()` is evaluated once per query, but it seems like queries for materialized views are executed for each batch separately and thus have different `now()` values. For example, I have tables like this:
```sql
CREATE TABLE IF NOT EXISTS test_kafka (
message String
) ENGINE = Kafka
SETTINGS
kafka_broker... | 2019-03-25T18:14:22 | 2019-03-25T18:14:22 | {} | NONE |
ClickHouse | ClickHouse | 476,733,692 | 4,402 | alexey-milovidov | BTW, ASan report inside `libunwind` is normal (by design). | 2019-03-26T16:32:17 | 2019-03-26T16:32:17 | {} | MEMBER |
ClickHouse | ClickHouse | 476,833,391 | 4,774 | Gladdy | I've updated the pull request to match the recent refactoring of Join.cpp and my local tests pass again.
There is still some code smell around having to pass in the timestamp, but in order to avoid this without much change to the rest of the codebase it might have to be required to write a specialized HashMap that i... | 2019-03-26T20:21:59 | 2019-03-26T20:21:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 477,487,181 | 3,822 | likeSakura | This operation is due to the large amount of data inserted, the asynchronous merge execution is not completed, which can be understood as oom.Kill the clickhouse-server process by force. | 2019-03-28T07:53:16 | 2019-03-28T07:53:16 | {} | NONE |
ClickHouse | ClickHouse | 477,755,011 | 4,774 | 4ertus2 | > I think your CI might have some issues - it's reporting the error below for commit [e7a10b8](https://github.com/yandex/ClickHouse/commit/e7a10b8a3e82b7abef2b62db70ae219b16007fa7) (https://clickhouse-builds.s3.yandex.net/4774/e7a10b8a3e82b7abef2b62db70ae219b16007fa7/build_log_403388044_1553802981.txt)
>
> ```
> ..... | 2019-03-28T20:18:10 | 2019-03-28T20:30:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 476,754,265 | 4,388 | alexey-milovidov | Unbundled build don't have to be feature complete. And we don't officially support it.
(And it is practically impossible, because we cannot guarantee what version of library we have in particular OS packages.)
We can simply disable brotli in unbundled build. | 2019-03-26T17:14:55 | 2019-03-26T17:15:48 | {} | MEMBER |
ClickHouse | ClickHouse | 477,023,609 | 4,817 | kindred77 | I add below in CMakeLists.txt in ubuntu for occi libs:
target_link_libraries(dbms PUBLIC -locci -lclntsh -lnnz12 -lclntshcore) | 2019-03-27T07:57:57 | 2019-03-27T07:57:57 | {} | NONE |
ClickHouse | ClickHouse | 476,827,836 | 4,681 | 4ertus2 | Reproduced with first patches in #4815 (works with ON, crashes with USING) | 2019-03-26T20:09:36 | 2019-03-27T14:30:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 477,279,053 | 4,827 | blinkov | cc @zhang2014 @sundy-li | 2019-03-27T17:53:06 | 2019-03-27T17:53:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 477,380,639 | 4,747 | alexey-milovidov | > maybe nodeNumber() would be better?
Yes, this name is better. | 2019-03-27T23:06:38 | 2019-03-27T23:06:38 | {} | MEMBER |
ClickHouse | ClickHouse | 476,548,795 | 4,800 | 4ertus2 | Hi,
It looks like you are trying join with nullable columns. There could be some bugs there.
1. Could you get your schema? It's interesting which of joined columns are Nullable and which are not.
2. Could you reproduce crash without GROUP BY and other complex constructions? We need the minimum minimorum sample :... | 2019-03-26T09:47:47 | 2019-03-26T09:48:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 476,578,054 | 777 | blinkov | @kshvakov is it working fine after https://github.com/kshvakov/zetcd/commit/8aebc2946ea5d4166bbc142b1bb26c58981b3705 ? | 2019-03-26T11:05:07 | 2019-03-26T11:05:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 476,641,866 | 4,674 | alexey-milovidov | Мне тоже не нравился старый код - я его с трудом прочитал. Хорошо, что вы его удалили!
Про тесты сейчас напишу... | 2019-03-26T13:41:37 | 2019-03-26T13:41:37 | {} | MEMBER |
ClickHouse | ClickHouse | 476,681,084 | 4,674 | smirnov-vs | Тест должен заливать по HTTP, схема должна быть в той папке, иначе никак. | 2019-03-26T14:47:17 | 2019-03-26T14:47:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 476,744,800 | 4,671 | alexey-milovidov | The test failure `00637_sessions_in_http_interface_and_settings` - unrelated, already fixed in master. | 2019-03-26T16:55:05 | 2019-03-26T16:55:05 | {} | MEMBER |
ClickHouse | ClickHouse | 477,305,380 | 4,824 | alexey-milovidov | Need to update two tests.
PS. Do you want to add `(version a.b.c.d)` instead just `(a.b.c.d)`? | 2019-03-27T19:01:55 | 2019-03-27T19:02:05 | {} | MEMBER |
ClickHouse | ClickHouse | 477,354,637 | 4,388 | alexey-milovidov | Ok. Now the only remaining issue is "unbundled" build.
You can test it with Valgrind if it looks suspicious. Or you can simply disable brotli support for "unbundled" build. | 2019-03-27T21:28:30 | 2019-03-27T21:28:30 | {} | MEMBER |
ClickHouse | ClickHouse | 477,588,502 | 1,287 | ghuname | Can we expect this feature in nearby future? | 2019-03-28T13:11:57 | 2019-03-28T13:11:57 | {
"rocket": 1
} | NONE |
ClickHouse | ClickHouse | 477,804,114 | 4,835 | alexey-milovidov | > CLICKHOUSE-4421 Fix segfault in copier
You did not mention in what cases the issue happens.
Integration test? | 2019-03-28T23:11:35 | 2019-03-28T23:11:35 | {} | MEMBER |
ClickHouse | ClickHouse | 476,503,932 | 4,800 | ChannelerH | clickhouse version is 19.1.6 | 2019-03-26T07:19:14 | 2019-03-26T07:19:14 | {} | NONE |
ClickHouse | ClickHouse | 476,653,173 | 4,674 | smirnov-vs | Хорошо, сделаю аналогично.
Справедливо заметить, что баг появился в случае, когда с форматом работал как раз сервер через HTTP, а тест покроет случай, когда файл читается локально и, скорее всего, за один присест. | 2019-03-26T13:57:53 | 2019-03-26T13:57:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 476,677,640 | 4,674 | alexey-milovidov | Вы исправили проблему, которую не покрывал существующий тест. Значит ваш тест должен чем-то отличаться, чтобы покрыть тот случай, который вы исправили. | 2019-03-26T14:41:21 | 2019-03-26T14:41:21 | {} | MEMBER |
ClickHouse | ClickHouse | 476,735,622 | 4,794 | alexey-milovidov | Functional tests with address sanitizer have failed at predicate expression optimizer:
https://clickhouse-test-reports.s3.yandex.net/4794/c4a7009c390b7f74b9a8e15124cbd82a3d35dafd/functional_stateless_tests_(address).html
(See `stderr.txt`)
It makes clear that the fix is not viable.
| 2019-03-26T16:36:28 | 2019-03-26T16:36:28 | {} | MEMBER |
ClickHouse | ClickHouse | 476,742,481 | 4,763 | proller | Can you show cmake output near `Using line editing libraries` ? | 2019-03-26T16:50:14 | 2019-03-26T16:50:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 477,475,528 | 4,388 | fandyushin | @alexey-milovidov I started debugging yesterday evening and really, this segfault looks very suspicious and its seems for me that this problem is not because of the brotli library. Today I will try to figure out what is wrong. | 2019-03-28T07:08:33 | 2019-03-28T07:08:33 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 477,578,642 | 4,546 | den-crane | https://github.com/yandex/ClickHouse/pull/4671 | 2019-03-28T12:42:33 | 2019-03-28T12:42:33 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 476,579,247 | 3,842 | blinkov | @nightweb @Saugardas @sergio-fry @xdo1 any comments on usecases? | 2019-03-26T11:08:17 | 2019-03-26T11:08:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 476,663,898 | 777 | Felixoid | How performant it by comparing with zookeeper? | 2019-03-26T14:14:21 | 2019-03-26T14:14:21 | {} | MEMBER |
ClickHouse | ClickHouse | 477,415,492 | 3,822 | likeSakura | 这样的操作是由于插入的数据量太大，异步合并执行未完成，可以理解为oom。强行kill掉了clickhouse-server进程。启动就挂也是因为这个。 | 2019-03-28T02:04:52 | 2019-03-28T02:04:52 | {} | NONE |
ClickHouse | ClickHouse | 477,739,648 | 4,837 | nvartolomei | Didn't have time to test, feel free to update PR and merge it as it fits. | 2019-03-28T19:31:12 | 2019-03-28T19:31:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 476,564,291 | 4,790 | TCeason | This seems to have nothing to do with my changes.
> [Functional stateful tests (thread) — fail:20, passed:46](https://clickhouse-test-reports.s3.yandex.net/4790/9d232778fc1e9c829f57a91cd8168ece88fbf4f3/functional_stateful_tests_(thread).html)
| 2019-03-26T10:30:20 | 2019-03-26T10:30:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 477,233,546 | 4,817 | proller | Try use full path to lib.
Or better use find_library https://cmake.org/cmake/help/latest/command/find_library.html | 2019-03-27T16:16:29 | 2019-03-27T16:16:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 457,495,440 | 972 | imvs | I have duplication to.
Here is configuration:
<pre>
<code>
SELECT *
FROM system.clusters
┌─cluster─┬─shard_num─┬─shard_weight─┬─replica_num─┬─host_name────────────────────┬─host_address─┬─port─┬─is_local─┬─user────┬─default_database─┐
│ test │ 1 │ 1 │ 1 │ ch-test02.unix.eklmn.... | 2019-01-25T08:32:36 | 2019-03-28T08:41:39 | {} | NONE |
ClickHouse | ClickHouse | 477,740,124 | 4,837 | nvartolomei | Another alternative solution would be to make underlying class smarter and retry this error for idempotent requests as most http clients do and use GET for data exchange. | 2019-03-28T19:32:34 | 2019-03-28T19:32:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 477,751,384 | 4,774 | Gladdy | I think your CI might have some issues - it's reporting the error below for commit e7a10b8 (https://clickhouse-builds.s3.yandex.net/4774/e7a10b8a3e82b7abef2b62db70ae219b16007fa7/build_log_403388044_1553802981.txt)
```
../dbms/src/Interpreters/Join.cpp:1417:123: error: too few arguments to function call, expected 4, h... | 2019-03-28T20:06:36 | 2019-03-28T20:06:36 | {} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.