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 | 616,985,948 | 10,391 | gglinux | Which version can solve the above two problems? #9860 , #10391 | 2020-04-21T06:43:24 | 2020-04-21T06:43:24 | {} | NONE |
ClickHouse | ClickHouse | 617,323,214 | 10,394 | alexey-milovidov | A link for curious readers: https://zachholman.com/talk/utc-is-enough-for-everyone-right | 2020-04-21T18:03:28 | 2020-04-21T18:03:28 | {} | MEMBER |
ClickHouse | ClickHouse | 617,441,804 | 8,078 | alexey-milovidov | This task is related to materialized views with periodic updates #3685 | 2020-04-21T22:15:14 | 2020-04-21T22:15:14 | {} | MEMBER |
ClickHouse | ClickHouse | 617,460,706 | 10,419 | alexey-milovidov | Look at the `input_format_allow_errors_num`, `input_format_allow_errors_ratio` settings. | 2020-04-21T23:17:57 | 2020-04-21T23:17:57 | {} | MEMBER |
ClickHouse | ClickHouse | 617,465,709 | 10,162 | filimonov | @KochetovNicolai do we know smth about that? | 2020-04-21T23:36:02 | 2020-04-21T23:36:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 617,466,881 | 10,309 | alexey-milovidov | I have checked it manually just in case... | 2020-04-21T23:40:40 | 2020-04-21T23:40:40 | {} | MEMBER |
ClickHouse | ClickHouse | 617,472,894 | 10,007 | alexey-milovidov | We have multiple patch releases since then.
Does the issue reproduce on version 20.1.10 or 20.3.7? | 2020-04-22T00:04:47 | 2020-04-22T00:04:47 | {} | MEMBER |
ClickHouse | ClickHouse | 617,000,011 | 10,385 | qoega | ```
2020-04-21 02:55:39 00306_insert_values_and_expressions: [ FAIL ] 0.64 sec. - return code 134
2020-04-21 02:55:39 Logical error: 'Unexpected literal type inside Array: Array(UInt8). It's a bug'.
2020-04-21 02:55:39 clickhouse-client: ../src/Common/Exception.cpp:40: DB::Exceptio... | 2020-04-21T07:17:28 | 2020-04-21T07:17:28 | {} | MEMBER |
ClickHouse | ClickHouse | 617,322,642 | 9,707 | maqroll | @oandrew
Sorry, I checked against a different build!!
Being able to access the rest of the fields (union support) would be great.
As I see it all the fields inside before should all be defined as Nullable (even if they are not nullable themselves, id for instance). Something like this:
```
before.Value.id Null... | 2020-04-21T18:02:27 | 2020-04-21T18:02:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 617,424,041 | 10,391 | filimonov | It is a duplicate of #7885
Workaround: `create table (...) engine=MySQL(...)` and use that instead. | 2020-04-21T21:28:38 | 2020-04-21T21:28:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 617,461,891 | 10,116 | filimonov | Maybe remote server just stops the query in the middle? Some query complexity limitation configured + overflow mode = break. Check https://github.com/ClickHouse/ClickHouse/blob/master/docs/en/operations/settings/query_complexity.md
It would be nice to get trace logs from both servers.
| 2020-04-21T23:22:06 | 2020-04-21T23:22:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 616,794,489 | 10,336 | den-crane | >proxy_pass http://somedomain.com:8123/;
And somedomain.com = MY_EXTERNAL_IP at this host? not 127.0.0.1 ?
| 2020-04-20T20:35:56 | 2020-04-20T20:35:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 616,818,784 | 10,386 | akuzm | Do you have an example for which the compilation breaks?
For hash table, the existing code does look broken for cases when key type != stored type, that is, any hash map. At first I was surprised because the hash maps are serialized by aggregator, but turns out it uses some custom code for that. This serialization m... | 2020-04-20T21:26:44 | 2020-04-20T21:26:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 616,971,358 | 10,393 | blinkov | @org0000h usually people decide whether they want to use releases marked "stable" or "lts", then the sooner you upgrade inside the chosen option (at least inside a major version), the sooner you receive latest bugfixes. | 2020-04-21T06:03:35 | 2020-04-21T06:03:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 617,198,659 | 10,397 | maison2710 | @qoega yes, I took from master.
I changed to version v20.3.5.21, which is a stable release, but the issue still happened. May you help for this? | 2020-04-21T13:59:59 | 2020-04-21T16:37:51 | {} | NONE |
ClickHouse | ClickHouse | 617,280,959 | 10,411 | alexey-milovidov | This is expected behaviour when you insert values other than -1 and 1 into Sign column of CollapsingMergeTree. | 2020-04-21T16:50:38 | 2020-04-21T16:50:38 | {} | MEMBER |
ClickHouse | ClickHouse | 617,320,994 | 10,394 | alexey-milovidov | @qoega Supporting different kind of calendars: Persian, Thai, Julian... - is not an easy task to do in general.
Maybe we can just do a shortcut by using fixed shift amounts in days. But it still requires research. | 2020-04-21T17:59:27 | 2020-04-21T17:59:27 | {} | MEMBER |
ClickHouse | ClickHouse | 617,350,119 | 9,707 | oandrew | I was thinking about maybe supporting both nullable and non-nullable.
In case of non-nullable it would use default value instead of null e.g. empty string for `String` or 0 for `Int`.
e.g. if you have schema
```
{
"type": "record",
"name": "Root",
"fields": [
{
"name": "before",
"type"... | 2020-04-21T18:54:26 | 2020-04-21T18:54:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 617,473,327 | 6,397 | alexey-milovidov | Does the issue reproduce on recent stable release? | 2020-04-22T00:06:30 | 2020-04-22T00:06:30 | {} | MEMBER |
ClickHouse | ClickHouse | 616,780,568 | 10,384 | alexey-milovidov | Need to add apt key... | 2020-04-20T20:07:44 | 2020-04-20T20:07:44 | {} | MEMBER |
ClickHouse | ClickHouse | 616,835,320 | 10,336 | TimonKK | Yes, and I see my MY_EXTERNAL_IP in clickhouse-server.log | 2020-04-20T22:03:02 | 2020-04-20T22:03:02 | {} | NONE |
ClickHouse | ClickHouse | 616,853,888 | 10,243 | olevole | my report does not apply to 20.3.7.46 (I have not tried this), but you can use my work in the right direction to update the official FreeBSD ports[1]
- log from poudriere: http://www.olevole.ru/clickhouse/12.1-RELEASE/mytest-default/2020-04-21_01h09m25s
- build from poudriere: http://www.olevole.ru/clickhouse/12.1-... | 2020-04-20T22:56:31 | 2020-04-20T22:56:31 | {} | NONE |
ClickHouse | ClickHouse | 616,867,562 | 10,259 | alexey-milovidov | It was "almost" fixed in https://github.com/ClickHouse/ClickHouse/pull/8866 | 2020-04-20T23:43:31 | 2020-04-20T23:43:31 | {} | MEMBER |
ClickHouse | ClickHouse | 617,440,054 | 8,301 | alexey-milovidov | @mayankkt9 You can still work on this :smile: | 2020-04-21T22:10:31 | 2020-04-21T22:10:31 | {} | MEMBER |
ClickHouse | ClickHouse | 617,444,947 | 2,829 | alexey-milovidov | Example: https://github.com/ClickHouse/ClickHouse/pull/10137/files#diff-fb9153de36f137866ae2fe4d54767f3bR51
Available since version 20.4. | 2020-04-21T22:24:05 | 2020-04-21T22:24:05 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 617,470,270 | 6,994 | alexey-milovidov | Fixed in master in #10217 | 2020-04-21T23:54:39 | 2020-04-21T23:54:39 | {} | MEMBER |
ClickHouse | ClickHouse | 617,473,879 | 7,537 | alexey-milovidov | We are using `mariadbclient` library (as opposing to `mysqlclient` library).
It might be a bug in this library that make it incompatible with the `session_track_state_change` feature. | 2020-04-22T00:08:34 | 2020-04-22T00:08:56 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 616,720,648 | 3,987 | blinkov | @yurigazinskiy, do you have any further questions? | 2020-04-20T18:04:56 | 2020-04-20T18:04:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 616,779,983 | 10,384 | alexey-milovidov | We also have to edit Sandbox files after this will be merged. | 2020-04-20T20:06:23 | 2020-04-20T20:06:23 | {} | MEMBER |
ClickHouse | ClickHouse | 616,781,995 | 2,564 | azat | >Those two conditions should be switched.
Indeed and looks pretty damn easy to fix | 2020-04-20T20:10:49 | 2020-04-20T20:10:49 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 616,783,778 | 10,384 | alexey-milovidov | Validating with
`docker build .`
...
Validated successfully. | 2020-04-20T20:14:46 | 2020-04-20T20:20:24 | {} | MEMBER |
ClickHouse | ClickHouse | 616,818,856 | 3,931 | alexey-milovidov | Ok. We can implement addition of intervals of different kinds as "external" addition in "direct product" of kinds, resulting in Tuple of Interval types.
In Postgres it works slightly different: https://www.db-fiddle.com/f/8hpz64orQDo76HhD2a4Asw/0 | 2020-04-20T21:26:54 | 2020-04-20T21:26:54 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 616,821,851 | 3,931 | alexey-milovidov | This task to be reassigned when the time will come. | 2020-04-20T21:33:29 | 2020-04-20T21:33:29 | {} | MEMBER |
ClickHouse | ClickHouse | 616,843,364 | 10,388 | andresokol | Moving from https://github.com/ClickHouse/ClickHouse/pull/9499 due to big structure reorganization in April | 2020-04-20T22:24:04 | 2020-04-20T22:24:04 | {} | NONE |
ClickHouse | ClickHouse | 616,865,613 | 10,385 | alexey-milovidov | Performance test is broken because the query is very long but @akuzm said that the target query speed up 75% with statistical confidence. | 2020-04-20T23:35:43 | 2020-04-20T23:35:43 | {} | MEMBER |
ClickHouse | ClickHouse | 617,444,219 | 9,196 | filimonov | Closes #8667 | 2020-04-21T22:21:55 | 2020-04-21T22:21:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 617,460,103 | 9,270 | alexey-milovidov | @alesapin It's still unclear for me, what will happen if the user did not specify size in cells. | 2020-04-21T23:16:01 | 2020-04-21T23:16:01 | {} | MEMBER |
ClickHouse | ClickHouse | 617,315,396 | 10,403 | alexey-milovidov | When I remove extra pair of braces, I get:
```
SELECT *
FROM generateRandom('`WatchID` UInt64, `JavaEnable` UInt8, `Title` String, `GoodEvent` Int16, `EventTime` DateTime, `EventDate` Date, `CounterID` UInt32, `ClientIP` UInt32, `ClientIP6` FixedString(16), `RegionID` UInt32, `UserID` UInt64, `CounterClass` Int8, ... | 2020-04-21T17:49:03 | 2020-04-21T17:49:03 | {} | MEMBER |
ClickHouse | ClickHouse | 617,347,792 | 10,246 | alexey-milovidov | Thank you! I've committed the fix. | 2020-04-21T18:49:45 | 2020-04-21T18:49:45 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 617,070,034 | 10,397 | qoega | You took 20.4.1 from master? There was no 20.4 stable release https://github.com/ClickHouse/ClickHouse/pulls?q=is%3Apr+is%3Aopen+label%3Arelease | 2020-04-21T09:37:41 | 2020-04-21T09:37:41 | {} | MEMBER |
ClickHouse | ClickHouse | 617,444,758 | 2,829 | alexey-milovidov | The original issue is fixed in #10137 for XML configuration.
| 2020-04-21T22:23:33 | 2020-04-21T22:23:33 | {} | MEMBER |
ClickHouse | ClickHouse | 590,056,672 | 9,270 | alexey-milovidov | @alesapin What will happen if the user did not specify the number of cells? | 2020-02-23T11:04:39 | 2020-04-21T22:53:14 | {} | MEMBER |
ClickHouse | ClickHouse | 617,469,385 | 9,360 | alexey-milovidov | @LiemLin You should upgrade to the latest stable version (20.3). | 2020-04-21T23:51:04 | 2020-04-21T23:51:04 | {} | MEMBER |
ClickHouse | ClickHouse | 617,452,145 | 3,243 | filimonov | Fixed in #9948 | 2020-04-21T22:47:52 | 2020-04-21T22:47:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 617,461,014 | 10,409 | alexey-milovidov | #10423 | 2020-04-21T23:19:04 | 2020-04-21T23:19:04 | {} | MEMBER |
ClickHouse | ClickHouse | 617,030,954 | 10,000 | ildus | Yes, it is basically GROUP BY, and could be implemented better without using `Field`. I will try to look into this regression today. | 2020-04-21T08:21:47 | 2020-04-21T08:21:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 617,311,527 | 10,406 | alexey-milovidov | As far as I remember, we can convert table definition from old to new format on creation time. But we should not do it because it will make impossible to create new replicas for tables of old format.
Add a setting is easy task to do and reasonable.
Also this task is loosely related to the task of support CREATE T... | 2020-04-21T17:41:52 | 2020-04-21T17:41:52 | {} | MEMBER |
ClickHouse | ClickHouse | 617,426,126 | 10,391 | filimonov | > Which version can solve the above two problems? #9860 , #10391
The latest 20.3, or the latest 20.1 (20.1.10 currently) | 2020-04-21T21:33:37 | 2020-04-21T21:33:37 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 617,438,775 | 7,512 | tavplubix | > I didn't found race-conditions test like 00993 or 00992 with DatabaseAtomic. Also, have you tried to run all tests with atomic database instead of ordinary? Myabe we should add this variant to CI.
DatabaseAtomic was enabled by default in this PR during development (until `20e6593`).
I want to add functional test... | 2020-04-21T22:07:07 | 2020-04-21T22:07:07 | {} | MEMBER |
ClickHouse | ClickHouse | 617,442,774 | 10,328 | filimonov | Did try both, normally async. | 2020-04-21T22:17:43 | 2020-04-21T22:17:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 617,463,878 | 10,392 | alexey-milovidov | You can compile in Docker simply by running `make build` inside `docker/builder` directory. | 2020-04-21T23:28:55 | 2020-04-21T23:29:33 | {} | MEMBER |
ClickHouse | ClickHouse | 617,469,649 | 9,040 | alexey-milovidov | Similar to #9270 | 2020-04-21T23:52:06 | 2020-04-21T23:52:06 | {} | MEMBER |
ClickHouse | ClickHouse | 617,472,668 | 10,418 | tavplubix | `00700_decimal_bounds` - `Success when error expected. It expects server error 0, client error 69.`
Looks like a bug in `CAST`:
```
:) select cast(1000*1000*1000, 'Decimal(9, 0)') as should_throw_overflow_exception
┌─should_throw_overflow_exception─┐
│ 1000000000 │
└────────────────────────... | 2020-04-22T00:03:56 | 2020-04-22T00:10:31 | {} | MEMBER |
ClickHouse | ClickHouse | 616,665,421 | 10,326 | den-crane | Yes, it does support but:
MV is an insert trigger. It watches inserts into LEFT FROM TABLE only.
So you can create MV `as select ... from SOMELEFT join SOMERIGHT`, but MV will be triggered by inserts into SOMELEFT.
Plus another problem that JOINs are slow and MV with JOIN with large SOMERIGHT will slow down inserts... | 2020-04-20T16:28:30 | 2020-04-20T16:28:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 616,722,310 | 10,000 | akuzm | Other things that could be improved in the existing implementation (not related to this PR):
* the memory is allocated not on Arena: https://clickhouse.tech/codebrowser/html_report/ClickHouse/src/AggregateFunctions/AggregateFunctionSumMap.h.html#134
* function data uses `std::map`, but we could use ClickHouse HashM... | 2020-04-20T18:08:18 | 2020-04-20T18:08:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 616,776,684 | 10,362 | azat | >max_server_memory_usage should be server setting in config.xml instead of user-level setting. Will do.
Maybe reintroducing `max_memory_usage_for_all_queries` for this particular case will be better? (not sure though) | 2020-04-20T19:59:27 | 2020-04-20T19:59:27 | {} | MEMBER |
ClickHouse | ClickHouse | 616,821,101 | 3,931 | alexey-milovidov | At least Postgres works correctly with respect to time zone adjustments:
https://www.db-fiddle.com/f/x99vPcZa9QFJhhzjAbb5Jh/1 | 2020-04-20T21:31:55 | 2020-04-20T21:31:55 | {} | MEMBER |
ClickHouse | ClickHouse | 616,864,352 | 10,259 | alexey-milovidov | @qoega You mixed up with #10219 | 2020-04-20T23:32:01 | 2020-04-20T23:32:01 | {} | MEMBER |
ClickHouse | ClickHouse | 617,005,267 | 10,246 | ficapy | @alexey-milovidov
I see [this branch](https://github.com/ClickHouse/ClickHouse/pull/10384) being merged and then pulling code to recompile. Now compiles normally, but executes an error.
`DB::ErrnoException: Cannot open file /proc/self/statm, errno: 2, strerror: No such file or directory`
```
build/programs/cli... | 2020-04-21T07:29:27 | 2020-04-21T07:32:51 | {} | NONE |
ClickHouse | ClickHouse | 617,428,383 | 10,398 | filimonov | #10200 ? | 2020-04-21T21:39:13 | 2020-04-21T21:39:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 617,435,258 | 10,415 | azat | >ClickHouse build check — 15/16 builds are OK
AFAICS the clang `[readability-*]` was not introduced by this PR | 2020-04-21T21:57:36 | 2020-04-21T21:57:49 | {} | MEMBER |
ClickHouse | ClickHouse | 617,443,728 | 9,068 | alexey-milovidov | @nikitamikhaylov has fixed this issue without even noticing it :) | 2020-04-21T22:20:27 | 2020-04-21T22:20:27 | {} | MEMBER |
ClickHouse | ClickHouse | 617,445,186 | 2,829 | alexey-milovidov | The implementation of SETTINGS clause for DDL queries is pending. | 2020-04-21T22:24:50 | 2020-04-21T22:24:50 | {} | MEMBER |
ClickHouse | ClickHouse | 616,889,693 | 10,319 | gubinjie | @filimonov Sorry, we checked for a few days and found that it was a problem with our application, which caused data loss, so I closed this bug. | 2020-04-21T01:07:50 | 2020-04-21T01:07:50 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 616,926,612 | 10,200 | AlexanderJLiu | @KochetovNicolai Problem still exists. And **Fatal errors** are same.
**Error message and/or stacktrace:**
There are some **Warnings** before Fatal:
```
2020.04.21 10:48:22.125274 [ 17030 ] {07054086-28ae-47cc-a713-031f3467f70a} <Warning> ConnectionPoolWithFailover: Connection failed at try №1, reason: Cod... | 2020-04-21T03:15:13 | 2020-04-21T05:48:20 | {} | NONE |
ClickHouse | ClickHouse | 617,098,600 | 10,200 | KochetovNicolai | @AlexanderJLiu interesting thing that stacktrace have changed. Right now I can't understand why it may happen, because server just can't parse type name string from remote server. I cannot easily reproduce this segfault too.
Warnings with ` Connection failed` probably say that remote server have also crashed. Can yo... | 2020-04-21T10:38:10 | 2020-04-21T10:38:10 | {} | MEMBER |
ClickHouse | ClickHouse | 617,313,310 | 10,411 | qoega | Error is shown during merge. There is no error in INSERT.
So you can break your production table by inserting one row in Collapsing table. And merge for this table would stop. | 2020-04-21T17:45:02 | 2020-04-21T17:48:43 | {} | MEMBER |
ClickHouse | ClickHouse | 617,318,558 | 10,402 | alexey-milovidov | Your query is simply incorrect - in all senses. | 2020-04-21T17:54:58 | 2020-04-21T17:54:58 | {} | MEMBER |
ClickHouse | ClickHouse | 617,433,842 | 10,328 | alexey-milovidov | Do you use synchronous or asynchronous INSERTs? | 2020-04-21T21:53:47 | 2020-04-21T21:53:47 | {} | MEMBER |
ClickHouse | ClickHouse | 617,438,289 | 10,406 | filimonov | > Can we rewrite metadata from old format to new one after attach?
https://github.com/ClickHouse/ClickHouse/blob/master/utils/convert-month-partitioned-parts/main.cpp | 2020-04-21T22:05:45 | 2020-04-21T22:05:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 617,455,219 | 6,336 | filimonov | Still true.
Regarding performance - you can create several Kafka tables to consume in parallel, together with increasing kafka_max_block_size to values like 512-1024k it usually solves the problem. | 2020-04-21T22:58:18 | 2020-04-21T22:58:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 617,465,779 | 10,332 | alexey-milovidov | It is also possible to change it with manually editing metadata files in filesystem and in ZooKeeper. But it's not recommended because it will require knowledge of ClickHouse internals and extra care. | 2020-04-21T23:36:17 | 2020-04-21T23:36:17 | {} | MEMBER |
ClickHouse | ClickHouse | 616,753,330 | 10,131 | azat | Looks like there are some issues with `distributed_group_by_no_merge` after this patch:
```sql
SELECT sum(number) FROM (SELECT * FROM remote('127.{1,2}', system.numbers) LIMIT 5 SETTINGS distributed_group_by_no_merge = 1)
```
This will `SIGSEGV` the server
I'm not sure is it related to this PR, but looks lik... | 2020-04-20T19:12:23 | 2020-04-20T19:14:41 | {} | MEMBER |
ClickHouse | ClickHouse | 616,787,177 | 10,369 | alexey-milovidov | No difference in perf test - we are lucky. | 2020-04-20T20:21:49 | 2020-04-20T20:21:49 | {} | MEMBER |
ClickHouse | ClickHouse | 617,244,038 | 10,406 | den-crane | > Can we rewrite metadata from old format to new one after attach?
Formats / parts are incompatible. Parts names / ZK metadata are incompatible as well.
There is no proper converter for parts from the old format to the new. It would be nice to have something like `alter table TTT convert to custom_partitioning` but... | 2020-04-21T15:14:44 | 2020-04-21T15:20:35 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 617,453,831 | 9,270 | alexey-milovidov | @alesapin the question is still relevant. | 2020-04-21T22:53:25 | 2020-04-21T22:53:25 | {} | MEMBER |
ClickHouse | ClickHouse | 617,469,879 | 8,352 | alexey-milovidov | The same as #9270 | 2020-04-21T23:52:53 | 2020-04-21T23:52:53 | {} | MEMBER |
ClickHouse | ClickHouse | 617,472,103 | 4,051 | alexey-milovidov | `Illegal Instruction` may have one of the following reasons:
- the server does not support SSE 4.2 instruction set that is required to run official ClickHouse build;
- the server has malfunctioning filesystem, RAM or CPU and the machine code of clickhouse get corrupted. | 2020-04-22T00:01:44 | 2020-04-22T00:01:44 | {} | MEMBER |
ClickHouse | ClickHouse | 616,807,575 | 9,707 | oandrew | @maqroll
`source.name` should work (tested locally with your schema).
`before.name` (or `after.name`) won't work because `before` field is a `union`.
I am planning to add support for such stuff in another PR.
The column name would then specify the path to a field based on avro json encoding (http://avro.apac... | 2020-04-20T21:03:11 | 2020-04-20T21:03:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 617,003,358 | 10,232 | qoega | We need perf tests for this feature and | 2020-04-21T07:25:20 | 2020-04-21T07:29:11 | {} | MEMBER |
ClickHouse | ClickHouse | 617,086,679 | 10,131 | KochetovNicolai | #10399 should fix it. | 2020-04-21T10:11:53 | 2020-04-21T10:11:53 | {} | MEMBER |
ClickHouse | ClickHouse | 617,192,211 | 2,364 | alesapin | I have added a test exactly for this case https://github.com/ClickHouse/ClickHouse/pull/10404. Seems like this bug was fixed, but I don't know when :( I think we can close this issue. | 2020-04-21T13:49:19 | 2020-04-21T13:49:19 | {} | MEMBER |
ClickHouse | ClickHouse | 617,466,764 | 10,309 | alexey-milovidov | Fixed here #10308 | 2020-04-21T23:40:13 | 2020-04-21T23:40:13 | {} | MEMBER |
ClickHouse | ClickHouse | 617,473,120 | 7,728 | alexey-milovidov | Not enough info. | 2020-04-22T00:05:37 | 2020-04-22T00:05:37 | {} | MEMBER |
ClickHouse | ClickHouse | 616,789,925 | 10,378 | alexey-milovidov | Fantastic! That's exactly what I need! | 2020-04-20T20:27:33 | 2020-04-20T20:27:33 | {} | MEMBER |
ClickHouse | ClickHouse | 616,915,549 | 10,326 | BernieJiangDL | > Yes, it does support but:
> MV is an insert trigger. It watches inserts into LEFT FROM TABLE only.
> So you can create MV `as select ... from SOMELEFT join SOMERIGHT`, but MV will be triggered by inserts into SOMELEFT.
> Plus another problem that JOINs are slow and MV with JOIN with large SOMERIGHT will slow down ... | 2020-04-21T02:36:50 | 2020-04-21T02:36:50 | {} | NONE |
ClickHouse | ClickHouse | 617,148,466 | 10,400 | Felixoid | This one duplicates #7019 #7091 #7736 #7939 #8628 | 2020-04-21T12:28:13 | 2020-04-21T12:28:13 | {} | MEMBER |
ClickHouse | ClickHouse | 617,215,432 | 10,406 | qoega | Can we rewrite metadata from old format to new one after attach? | 2020-04-21T14:28:15 | 2020-04-21T14:28:15 | {} | MEMBER |
ClickHouse | ClickHouse | 617,281,287 | 10,411 | alexey-milovidov | ClickHouse tells you about incorrect data clearly and correctly:
```
Incorrect data: Sign = -117 (must be 1 or -1).
``` | 2020-04-21T16:51:09 | 2020-04-21T16:51:09 | {} | MEMBER |
ClickHouse | ClickHouse | 617,317,779 | 10,402 | alexey-milovidov | There is a feature request to allow to INSERT data into ALIAS column under the following conditions:
- alias expression is trivial (level 2: if it's reversible);
- you are not inserting into dependend columns simultaneously (level 2: if you are inserting consistent data). | 2020-04-21T17:53:30 | 2020-04-21T17:53:30 | {} | MEMBER |
ClickHouse | ClickHouse | 617,442,874 | 9,068 | alexey-milovidov | It works exactly as expected. I will add a test.
```
CREATE DICTIONARY testip
(
`network` String,
`test_field` String
)
PRIMARY KEY network
SOURCE(FILE(PATH '/tmp/test.csv' FORMAT CSVWithNames))
LIFETIME(MIN 0 MAX 300)
LAYOUT(IPTRIE())
Received exception from server (version 20.4.1):
Code: 137.... | 2020-04-21T22:18:01 | 2020-04-21T22:18:01 | {} | MEMBER |
ClickHouse | ClickHouse | 617,446,560 | 4,201 | filimonov | That is implemented in #7360 and further PRs, avaliable only with DDL dictionary syntax in clickhouse versions >= 20.1 | 2020-04-21T22:28:53 | 2020-04-21T22:28:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 617,457,672 | 9,270 | alexey-milovidov | The issue is fixed, now you will get the error message:
`Code: 349. DB::Exception: Received from localhost:9000. DB::Exception: Cannot convert NULL value to non-Nullable type: while converting source column col8 to destination column col8.` | 2020-04-21T23:07:10 | 2020-04-21T23:07:10 | {} | MEMBER |
ClickHouse | ClickHouse | 617,269,369 | 10,401 | 4ertus2 | There's a workaround: add alias explicitly `SELECT a < (SELECT 1) s FROM t_dist`.
Query `SELECT a < ((SELECT 1) as s) FROM t_dist` also doesn't work without alias:
```
Not found column less(a, s) in block. There are only columns: less(a, CAST(1, 'UInt8'))
```
Alias fixes it `SELECT a < ((SELECT 1) as s) k FROM t... | 2020-04-21T16:29:26 | 2020-04-21T16:43:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 617,314,667 | 10,403 | alexey-milovidov | The reason is probably that you have extra pair of parentheses around list of types inside generateRandom function. The diagnostics could be better (so I change "bug" to "usability").
Upd. Yes, that is the reason. | 2020-04-21T17:47:40 | 2020-04-21T17:49:30 | {} | MEMBER |
ClickHouse | ClickHouse | 617,317,049 | 10,402 | alexey-milovidov | You are trying to INSERT data into alias column like
```
INSERT INTO mass_table_586 (x, y, s) VALUES ...
```
It is not possible.
And INSERT SELECT cannot silently drop these values. | 2020-04-21T17:52:09 | 2020-04-21T17:52:09 | {} | MEMBER |
ClickHouse | ClickHouse | 617,057,042 | 10,396 | qoega | looks ok | 2020-04-21T09:12:47 | 2020-04-21T09:12:47 | {} | MEMBER |
ClickHouse | ClickHouse | 617,086,973 | 10,394 | qoega | ClickHouse is an OpenSource product, so you can try to implement it yourself.
https://github.com/ClickHouse/ClickHouse/blob/master/CONTRIBUTING.md | 2020-04-21T10:12:29 | 2020-04-21T10:12:29 | {} | MEMBER |
ClickHouse | ClickHouse | 627,878,369 | 7,714 | filimonov | it looks it relates to prewhere and columns which has not value written in column yet (i.e. freshly added, absent in parts). Trying to reproduce...
@d-ganchar i think both disabling `optimize_move_to_prewhere` or running `OPTIMIZE TABLE ... FINAL` should help. | 2020-05-13T09:51:45 | 2020-05-13T09:54:24 | {
"heart": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 627,915,016 | 10,852 | dgrr | What are the naming conventions for the test scripts? I didn't find out on the wiki anything related with test name convention | 2020-05-13T11:14:17 | 2020-05-13T11:14:17 | {} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.