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 | 601,163,344 | 9,716 | alexey-milovidov | Q: How performance test identifies queries for comparison?
A: The queries are identified by their text.
That's good. | 2020-03-19T12:58:46 | 2020-03-19T12:58:46 | {} | MEMBER |
ClickHouse | ClickHouse | 601,264,187 | 9,715 | alexey-milovidov | @4ertus2 Ok. | 2020-03-19T16:03:07 | 2020-03-19T16:03:07 | {} | MEMBER |
ClickHouse | ClickHouse | 601,352,887 | 6,833 | azat | Looks like original issue should be fixed with #9154
Since according to the comment
> Following this article, I setup circular replication with 3 shards across a 3 node cluster. Node 1 had shards 1,2, Node 2 had shards 2,3, and Node 3 had shards 3,1. These were all merge tables, and one distributed table on each... | 2020-03-19T18:45:13 | 2020-03-19T18:45:13 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 600,480,859 | 9,722 | wangzhen11aaa | Use zk-shell `/.../mutations/`
```
> get 0000077325
```
```
format version: 1
create time: 2020-03-18 13:45:02
source replica: cluster01-01-6
block numbers count: 478
19700823 77327
19840414 77331
20181101 72955
20181102 68951
20181201 79785
20181202 79450
20181203 78277
20181204 77593
20181205 791... | 2020-03-18T08:09:02 | 2020-03-18T09:01:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 600,505,460 | 9,720 | filimonov | Thoretically it's possible that your server will be a part of several clusters with different topology (sharded only / replicated only etc/ some subset of hosts etc).
So for replication - the primary source of truth is a zookeeper.
Each replica adds its address to the zookeeper, where other replicas can find it.... | 2020-03-18T09:09:07 | 2020-03-18T09:09:07 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 600,628,040 | 9,722 | den-crane | >But the clickhouse actually will update the partitions with adding the mutation_versions +1
CH always mutates and adds the mutation_versions +1 **to ALL PARTS.** (consistency requirement)
New versions / new parts are created for alll parts using **HARDLINKS**
Only new parts those match `where dt='2020-02-03'` wi... | 2020-03-18T13:39:48 | 2020-03-18T13:41:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 601,071,059 | 9,673 | KochetovNicolai | 01085_max_distributed_connections_http fails :( | 2020-03-19T09:15:09 | 2020-03-19T09:15:09 | {} | MEMBER |
ClickHouse | ClickHouse | 601,358,062 | 9,754 | alexey-milovidov | Discussion in TG chat. | 2020-03-19T18:56:12 | 2020-03-19T18:56:12 | {} | MEMBER |
ClickHouse | ClickHouse | 601,429,206 | 9,742 | filimonov | Implements #4433 | 2020-03-19T21:38:18 | 2020-03-19T21:38:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 601,557,616 | 9,340 | filimonov | Please consider the case when `i=0`. | 2020-03-20T06:44:13 | 2020-03-20T06:44:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 600,179,434 | 9,706 | VonShtirlitz | mkdir /var/lib/clickhouse
chown -R clickhouse:clickhouse /var/lib/clickhouse
rm -rf /nonexistent
sudo apt-get remove clickhouse-client clickhouse-server
sudo apt-get install clickhouse-client clickhouse-server
and now it's working | 2020-03-17T16:49:20 | 2020-03-18T07:10:21 | {} | NONE |
ClickHouse | ClickHouse | 600,547,485 | 9,670 | KochetovNicolai | Well, according to log, threads number 7 does actual mutation:
```
2020.03.15 18:23:27.160160 [ 7 ] {} <Trace> ck.detail: Executing log entry to mutate part 37_0_12020_13_12037 to 37_0_12020_13_12038
2020.03.15 18:23:27.160214 [ 7 ] {} <Debug> DiskLocal: Reserving 3.99 GiB on disk `default`, having unreserved 337.14... | 2020-03-18T10:37:48 | 2020-03-18T10:37:48 | {} | MEMBER |
ClickHouse | ClickHouse | 600,568,784 | 9,715 | alex-zaitsev | Vote for 1, since it is close to SQL standard. | 2020-03-18T11:25:54 | 2020-03-18T11:25:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 600,672,107 | 7,306 | silence44 | @filimonov I have same issue and here is steps to reproduce.
CREATE TABLE mydb.test3 (\`id\` UInt64, \`i\` Int64) ENGINE = MergeTree() PRIMARY KEY id ORDER BY id SETTINGS index_granularity = 8192;
INSERT INTO mydb.test3 VALUES (1, 5756242266);
ALTER TABLE mydb.test3 UPDATE i = 5756242267 WHERE id =1;
Then I... | 2020-03-18T14:58:45 | 2020-03-18T14:58:45 | {} | NONE |
ClickHouse | ClickHouse | 600,675,450 | 9,584 | ShiKaiWi | > Implement this function for string literals is more important.
We can add support for numeric types as a second step.
Actually I am a little confused about this. Does **this function** means `toDateTimeOrNull`? If it does I think it is already able to handle string literals but can't handle numeric parameter:
``... | 2020-03-18T15:04:43 | 2020-03-18T15:04:43 | {} | NONE |
ClickHouse | ClickHouse | 600,876,985 | 9,715 | alexey-milovidov | @filimonov There is a huge amount of use cases that supposed to work. Recent example is one of TPC-DS queries.
You can write `SELECT x + 1 AS x, x FROM (SELECT 1 AS x) t`
Postgres: http://sqlfiddle.com/#!17/0a28f/287
MySQL: http://sqlfiddle.com/#!9/0a28f18/431
Oracle: http://sqlfiddle.com/#!4/ce3d98/2
sqlite: ... | 2020-03-18T21:51:31 | 2020-03-18T21:58:13 | {} | MEMBER |
ClickHouse | ClickHouse | 601,089,480 | 9,458 | alexey-milovidov | Now it looks perfect. | 2020-03-19T09:56:49 | 2020-03-19T09:56:49 | {} | MEMBER |
ClickHouse | ClickHouse | 601,102,645 | 9,721 | amosbird | If you type `format_` the prefix will be highlighted correctly. This is because replxx doesn't do case insensitive comparison when doing LCP https://github.com/AmokHuginnsson/replxx/blob/master/src/replxx_impl.cxx#L830
| 2020-03-19T10:25:08 | 2020-03-19T10:27:28 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 601,163,762 | 8,821 | alesapin | Fixed here https://github.com/ClickHouse/ClickHouse/pull/9223 | 2020-03-19T12:59:43 | 2020-03-19T12:59:43 | {} | MEMBER |
ClickHouse | ClickHouse | 601,338,301 | 9,715 | alex-zaitsev | If prefer_on_column_clashes=column, would the query like below work:
```
SELECT sum(revenue) as revenue,
max(revenue) as max_revenue -- use revenue column
FROM ....
ORDER by revenue -- use alias
``` | 2020-03-19T18:18:27 | 2020-03-19T18:18:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 601,505,546 | 6,892 | akuzm | Split build failed because of CI error -- works locally. | 2020-03-20T03:03:21 | 2020-03-20T03:03:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 601,558,462 | 8,326 | filimonov | Docs need to be updated.
And probably the same should be done for RunningDifference and neighbor | 2020-03-20T06:48:48 | 2020-03-20T06:50:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 600,711,017 | 9,584 | ShiKaiWi | I guess I figured out what's wrong. The issue title says:
> toDateTimeOrNull does not accept second argument with timezone.
Actually `toDateTimeOrNull` does not accept first (not second) argument of numeric type and it does accept second arguement with timezone. | 2020-03-18T15:52:26 | 2020-03-18T15:52:39 | {} | NONE |
ClickHouse | ClickHouse | 601,084,630 | 9,673 | KochetovNicolai | > So revert the last change?
Well, let's not revert anything.
Let's just fix you test and merge this pr. Because it solves the issue and all tests are fine.
Then I check how `max_streams_to_max_threads_ratio` works and fix it if something wrong separetly. | 2020-03-19T09:45:34 | 2020-03-19T09:45:34 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 601,089,231 | 9,458 | alexey-milovidov | Fixed on the new website by https://github.com/ClickHouse/ClickHouse/issues/9458#issuecomment-594123762 | 2020-03-19T09:56:16 | 2020-03-19T09:56:16 | {} | MEMBER |
ClickHouse | ClickHouse | 601,496,405 | 9,721 | amosbird | @alexey-milovidov that's expected because the `merge` table function is also a correct candidate (although I cannot reproduce locally since there is also a `Merge` engine). You can type an additional `t` to complete `MergeTree`.
Or perhaps we can implement smart case completion which, if user inputs any uppercase c... | 2020-03-20T02:08:56 | 2020-03-20T02:10:50 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 600,406,997 | 9,716 | alexey-milovidov | FYI http://sqlfiddle.com/#!9/0a28f18/429 | 2020-03-18T03:32:13 | 2020-03-18T03:32:13 | {} | MEMBER |
ClickHouse | ClickHouse | 600,407,095 | 9,716 | alexey-milovidov | And there is no way to prevent CSE for other functions. | 2020-03-18T03:32:39 | 2020-03-18T03:32:39 | {} | MEMBER |
ClickHouse | ClickHouse | 600,489,666 | 9,607 | qoega | OK, nice.
FunctionCache is added in common, so I thick it is better to add some comments about how to use it. For example to add TODO to support eviction if you want to use it without resetting on each part. | 2020-03-18T08:30:30 | 2020-03-18T08:30:30 | {} | MEMBER |
ClickHouse | ClickHouse | 600,595,065 | 9,682 | abyss7 | Closed in favor of #9727 | 2020-03-18T12:31:41 | 2020-03-18T12:31:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 601,140,618 | 9,716 | akuzm | What do you mean by edited?
The raw data is a TSV that contains query text, run and server index, and elapsed time as reported by the server (in this case, `arithmetic-raw.tsv` and `arithmetic-queries.tsv`). Then the statistics are saved to `arithmetic-report.tsv`. Combined statistics for all tests are in `all-quer... | 2020-03-19T12:02:20 | 2020-03-19T12:02:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 601,360,484 | 9,761 | azat | Nice catch! I was thinking about removing it (in #9154) but decided to make the patch small and clean (in particular for easier backporting) | 2020-03-19T19:01:19 | 2020-03-19T19:01:19 | {} | MEMBER |
ClickHouse | ClickHouse | 601,365,010 | 9,154 | alexey-milovidov | @azat But how it's related to `ParallelInputsProcessor`, that (despite it's name) has no relationship to processors? | 2020-03-19T19:11:01 | 2020-03-19T19:11:01 | {} | MEMBER |
ClickHouse | ClickHouse | 600,508,475 | 9,713 | filimonov | Can you explian a bit how you get that result?
Can you enable trace log (enabled by default, check `/var/log/clickhouse-server/clickhouse-server.log`) and after reproducing the problem - check for all occurrences of problematic part name in that log. | 2020-03-18T09:15:43 | 2020-03-18T09:15:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 601,557,759 | 9,701 | zhouxiujue | Thank you！ | 2020-03-20T06:45:03 | 2020-03-20T06:45:03 | {} | NONE |
ClickHouse | ClickHouse | 600,458,997 | 9,646 | Jokser | > > Renaming works as well, will add a test for that.
>
> For folders? I remember there was some issues on using MergeTree on fuse because directory rename is considered as instant/atomic op. (it is not on s3
Directory renaming in DiskS3 just renames directory on local FS where metadata files are located. | 2020-03-18T07:06:50 | 2020-03-18T07:06:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 600,558,846 | 9,724 | zhao-xiaogang | configuration file: /etc/clickhouse-server/users.d/xyz.xml
remove allow_databases, DESCRIBE TABLE functions successful
```xml
<?xml version="1.0"?>
<yandex>
<users>
<xyz>
<password></password>
<!-- remove allow_databases, DESCRIBE TABLE functions successful
<allow_databa... | 2020-03-18T11:02:48 | 2020-03-18T11:02:48 | {} | NONE |
ClickHouse | ClickHouse | 600,570,765 | 9,573 | raghuteja | @alesapin
We have updated clickhouse version to `20.1.6.30`
Thanks for the suggestion.
Now that issue is not coming but it is showing timeout
```
2020.03.18 11:27:58.488195 [ 53 ] {} <Error> void DB::StorageKafka::threadFunc(): std::exception. Code: 1001, type: cppkafka::HandleException, e.what() = Local:... | 2020-03-18T11:31:05 | 2020-03-18T11:31:41 | {} | NONE |
ClickHouse | ClickHouse | 600,887,898 | 9,578 | filimonov | Thanks for confirming! | 2020-03-18T22:20:50 | 2020-03-18T22:20:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 600,953,566 | 9,741 | den-crane | CH has specific design -- all nodes are independent without the singe point of failure so nodes has zero knowledge about other nodes.
ReplicatedMT knows nothing about a sharding and knows how to publish information about its parts using ZK.
All shards ReplicatedMT are different tables.
| 2020-03-19T02:26:02 | 2020-03-19T02:26:02 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 601,157,026 | 9,715 | filimonov | > @filimonov There is a huge amount of use cases that supposed to work. Recent example is one of TPC-DS queries.
>
> You can write SELECT x + 1 AS x, x FROM (SELECT 1 AS x) t
That is how standard SQL work :)
Clickhouse work another way:
```sql
SELECT sin(pk) as pk
FROM table
WHERE pk = 10;
```
in click... | 2020-03-19T12:43:28 | 2020-03-19T12:45:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 601,558,199 | 9,768 | YiuRULE | I could be wrong and maybe it's an unstable feature or a deprecated one in the master branch, but it seems to work perfectly on `20.3.3` | 2020-03-20T06:47:29 | 2020-03-20T06:47:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 600,619,765 | 9,715 | alexey-milovidov | @4ertus2 What do you think is the most simple and natural in terms of the implementation and the user experience? | 2020-03-18T13:23:03 | 2020-03-18T13:23:15 | {} | MEMBER |
ClickHouse | ClickHouse | 600,817,769 | 9,716 | akuzm | [metric-deviation.txt](https://github.com/ClickHouse/ClickHouse/files/4350975/metric-deviation.txt)
As usual, page faults and allocator things show the most variation, but now also the query profiler overruns. Need to investigate the reason for its recent slowdown. | 2020-03-18T19:28:25 | 2020-03-18T19:28:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 601,087,221 | 9,716 | alexey-milovidov | 3 "slower":
https://clickhouse-test-reports.s3.yandex.net/9716/f04c2da1805a96b86b675ca2a8572f3e6c8f9b02/performance_comparison/report.html
But there are the edited queries.
How performance test identifies queries for comparison? Does it just use query text or something more sophisticated? | 2020-03-19T09:51:39 | 2020-03-19T09:51:39 | {} | MEMBER |
ClickHouse | ClickHouse | 601,121,860 | 9,721 | amosbird | https://github.com/ClickHouse-Extras/replxx/pull/3 | 2020-03-19T11:11:42 | 2020-03-19T11:11:42 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 601,373,800 | 9,715 | alexey-milovidov | @alex-zaitsev No, this query won't work with the proposed solution. | 2020-03-19T19:24:49 | 2020-03-19T19:24:49 | {} | MEMBER |
ClickHouse | ClickHouse | 601,511,780 | 9,742 | alexey-milovidov | @hczhcz could you please fix it? | 2020-03-20T03:41:26 | 2020-03-20T03:41:26 | {} | MEMBER |
ClickHouse | ClickHouse | 601,556,974 | 5,485 | filimonov | [neighbor](https://clickhouse.tech/docs/en/query_language/functions/other_functions/#neighbor
) function introduced in #5925 is avaliable since 19.14 address that issue.
But it works currently inside each block separately:
```
SELECT
number,
neighbor(toNullable(number), 1)
FROM numbers(12)
ORDER B... | 2020-03-20T06:41:02 | 2020-03-20T06:41:02 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 600,544,938 | 9,670 | KochetovNicolai | A. and B. takes almost no memory.
| 2020-03-18T10:32:34 | 2020-03-18T10:32:34 | {} | MEMBER |
ClickHouse | ClickHouse | 601,232,582 | 9,574 | Enmk | > How is it going?
Was busy with other things, will come back to this soon. | 2020-03-19T15:08:19 | 2020-03-19T15:08:19 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 601,426,929 | 9,673 | azat | >Functional stateless tests (thread) — fail: 20, passed: 1522, skipped: 8
00111_shard_external_sort_distributed breaks everything after, but I cannot find a lot from the logs (maybe clickhouse-server had been killed by OOM, especially TSAN version)
P.S. recent changes passes all tests except for 01085_max_distrib... | 2020-03-19T21:32:14 | 2020-03-19T21:32:30 | {} | MEMBER |
ClickHouse | ClickHouse | 601,429,036 | 4,433 | filimonov | #9742 implements the feature | 2020-03-19T21:37:49 | 2020-03-19T21:37:49 | {
"rocket": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 600,561,263 | 9,670 | KochetovNicolai | This still may happen if you have "fat" rows.
Maybe temporarily decreasing of `max_block_size` for default user may help. | 2020-03-18T11:08:16 | 2020-03-18T11:08:16 | {} | MEMBER |
ClickHouse | ClickHouse | 597,715,191 | 9,606 | alesapin | Need to remove:
- [x] AlterDataPartTransaction
- [x] alterDataPart
- [x] analyzeAlterConversions
- [x] columns_lock from MergeTreeDataPart (finally)
- [x] removeEmptyColumnsFromPart
- [x] Simplify `mutatePartToTemporaryPart` (hell of different columns sets)
Make code less clumsy. | 2020-03-11T15:54:42 | 2020-03-18T13:17:46 | {} | MEMBER |
ClickHouse | ClickHouse | 600,832,035 | 9,715 | den-crane | >SELECT avg(number) AS number, max(number) FROM numbers(10)
I would prefer a DB error: `ambiguous column name/expression: [number]. Qualify explicit table aliases. `
>If aliased expression contains aggregate function,
>alias should not be resolved inside another aggregate function.
The next day you'll get a ... | 2020-03-18T20:02:48 | 2020-03-18T20:07:16 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 600,870,011 | 9,715 | filimonov | Aliases which 'hide' the original column value - doesn't follow SQL standard and is counterintuitive sometimes, but since we have it already everywhere, I'm against introducing some exceptions. The option proposed by @den-crane is the safest. | 2020-03-18T21:33:42 | 2020-03-18T21:33:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 600,966,302 | 9,622 | YiuRULE | @qoega a bit difficult to test the feature, it doesn't seem we have right now a test on the prometheus exporter, except one who just see if we have an output and see if every metrics begin by `ClickHouse` :/ | 2020-03-19T03:18:47 | 2020-03-19T03:18:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 601,077,883 | 9,673 | azat | >Actually, it seems that reference file is wrong.
Yep, you are right! | 2020-03-19T09:30:41 | 2020-03-19T09:30:41 | {} | MEMBER |
ClickHouse | ClickHouse | 601,162,389 | 9,715 | alexey-milovidov | > 2. If alias name clashes with the column name, the substitution of this alias should be cancelled.
The proposed solution is to add a setting `prefer_column_name_to_alias` with default 1.
```
SELECT sin(pk) as pk
FROM table
WHERE pk = 10;
```
If there is a column named `pk` in the table, then `pk` will r... | 2020-03-19T12:56:37 | 2020-03-19T12:56:37 | {} | MEMBER |
ClickHouse | ClickHouse | 600,491,261 | 9,622 | qoega | Can we add a test for this feature? | 2020-03-18T08:34:40 | 2020-03-18T08:34:40 | {} | MEMBER |
ClickHouse | ClickHouse | 600,766,323 | 9,715 | akuzm | Being able to refer to aliases from SELECT list is a distinctive ClickHouse thing, which is sometimes convenient. Current behavior of substituting always + erroring out actually seems preferrable to me, because it works the same way always, unconditionally.
The use case doens't look compelling, maybe there is a more... | 2020-03-18T17:33:35 | 2020-03-18T17:33:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 601,076,020 | 9,673 | KochetovNicolai | Actually, it seems that reference file is wrong. | 2020-03-19T09:26:07 | 2020-03-19T09:26:07 | {} | MEMBER |
ClickHouse | ClickHouse | 601,083,720 | 9,721 | alexey-milovidov | 
| 2020-03-19T09:43:34 | 2020-03-19T09:43:34 | {} | MEMBER |
ClickHouse | ClickHouse | 601,090,943 | 9,730 | KochetovNicolai | Reproduced:
```
:) select * from system.one limit 1 format JSON
SELECT *
FROM system.one
LIMIT 1
FORMAT JSON
{
"meta":
[
{
"name": "dummy",
"type": "UInt8"
}
],
"data":
[
{
"dummy": 0
}
],
"rows": 1,
"rows_before_limit_at_least": 1,
"statistics":
{
"elap... | 2020-03-19T10:00:08 | 2020-03-19T10:00:08 | {} | MEMBER |
ClickHouse | ClickHouse | 601,559,469 | 9,340 | leiysky | @filimonov Adding an if to deal with it is fine.
I'm just curious about whether there is any concern about optimization or not, since there are many magic codes in this project to do so. | 2020-03-20T06:53:49 | 2020-03-20T06:53:49 | {} | NONE |
ClickHouse | ClickHouse | 600,694,887 | 9,595 | AlexAkulov | Кажется удалось достать лог с запросом полностью
```
2020.03.18 18:15:49.924090 [ 722 ] {32b3f564-e718-4c72-ae85-495b10d5a645} <Debug> clickhouse.origins_transitions (SelectExecutor): Key condition: u
nknown
2020.03.18 18:15:49.924123 [ 722 ] {32b3f564-e718-4c72-ae85-495b10d5a645} <Debug> clickhouse.origins_transit... | 2020-03-18T15:39:04 | 2020-03-18T15:39:04 | {} | NONE |
ClickHouse | ClickHouse | 600,725,867 | 4,165 | silence44 | I also have same problem but with integers. Described more detailed at #7306 | 2020-03-18T16:18:18 | 2020-03-18T16:18:52 | {} | NONE |
ClickHouse | ClickHouse | 601,049,799 | 7,306 | silence44 | Hello @filimonov . Query that you provided works.
But that is not a solution for me, because I have to store all column types and do casting on every query, handle column type changes etc.
Update behavior have to be similar to insert in this case, but currently it is different.
May you please label this issue as... | 2020-03-19T08:26:35 | 2020-03-19T08:26:35 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 601,427,468 | 7,306 | filimonov | Related issues: #4165 #3746 #7442 #6778 | 2020-03-19T21:33:40 | 2020-03-19T21:33:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 601,511,080 | 9,742 | alexey-milovidov | There is a bug:
```
2020.03.20 02:37:23.729114 [ 15306 ] {2508016b-5e04-4815-b037-68b9238b972d} <Debug> executeQuery: (from [::1]:54184) SELECT splitByString('', '')
2020.03.20 02:37:23.735104 [ 147520 ] {} <Fatal> BaseDaemon: ########################################
2020.03.20 02:37:23.735170 [ 147520 ] {} <Fata... | 2020-03-20T03:37:13 | 2020-03-20T03:37:13 | {} | MEMBER |
ClickHouse | ClickHouse | 600,402,192 | 9,622 | YiuRULE | @azat @alexey-milovidov solved the issues, it doesn't segfault anymore, values are now properly quoted and build has been fixed for clang-8 on spitted build | 2020-03-18T03:10:32 | 2020-03-18T03:10:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 600,406,798 | 9,716 | alexey-milovidov | For `rand` function you can pass arbitrary argument as a "tag". | 2020-03-18T03:31:19 | 2020-03-18T03:31:19 | {} | MEMBER |
ClickHouse | ClickHouse | 600,708,416 | 2,820 | den-crane | Because a client tries to change the parameter `set result_overflow_mode='break` to limit rows returned by a query. And it needs READ_ONLY mode =2 | 2020-03-18T15:50:55 | 2020-03-18T15:50:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 601,083,280 | 9,721 | alexey-milovidov | @amosbird I found strange issue:
when I type `merge` or `replica` and press TAB, there is no highlight of matched prefix by color.
while when I type, `max_rows` or `split`, it works as expected. | 2020-03-19T09:42:39 | 2020-03-19T09:42:39 | {} | MEMBER |
ClickHouse | ClickHouse | 601,298,783 | 9,755 | blinkov | > will you add readme?
probably, for now it's right behind shebang of shell scripts, this is just an experiment | 2020-03-19T16:59:42 | 2020-03-19T17:00:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 601,349,466 | 9,154 | alexey-milovidov | The original issue
> ParallelAggregatingBlockInputStream::Handler::onFinish/onFinishThread
should not affect versions 20.3+, because `experimental_use_processors` is enabled. | 2020-03-19T18:37:58 | 2020-03-19T18:37:58 | {} | MEMBER |
ClickHouse | ClickHouse | 601,359,522 | 9,154 | azat | >should not affect versions 20.3+, because experimental_use_processors is enabled.
It was found on 20.2.1.2337-23 (even though AFAIK there wasn't official release), and AFAICS the processors has been enabled:
```
$ git show v20.2.1.2337-testing:dbms/src/Core/Settings.h | fgrep processors
M(SettingBool, expe... | 2020-03-19T18:59:09 | 2020-03-19T18:59:09 | {} | MEMBER |
ClickHouse | ClickHouse | 601,429,132 | 9,154 | azat | >@azat But how it's related to ParallelInputsProcessor, that (despite it's name) has no relationship to processors?
Hm, query that triggers it had been initiated by the Distributed engine, and I guess [it was possible](https://github.com/ClickHouse/ClickHouse/pull/8929/files#diff-674a9476cad518ac771ee9d89fa4f282R92)... | 2020-03-19T21:38:02 | 2020-03-19T21:40:04 | {} | MEMBER |
ClickHouse | ClickHouse | 600,480,998 | 9,606 | qoega | Please check that changes in docs are not needed after this changes https://clickhouse.tech/docs/en/query_language/alter/#query_language_queries_alter | 2020-03-18T08:09:25 | 2020-03-18T08:09:25 | {} | MEMBER |
ClickHouse | ClickHouse | 600,618,891 | 9,715 | alexey-milovidov | @den-crane Please advise.
(you can link all the existing issues if you want) | 2020-03-18T13:21:16 | 2020-03-18T13:21:55 | {} | MEMBER |
ClickHouse | ClickHouse | 600,859,390 | 9,724 | vitlibar | Already fixed in https://github.com/ClickHouse/ClickHouse/pull/9065 | 2020-03-18T21:09:01 | 2020-03-18T21:09:01 | {} | MEMBER |
ClickHouse | ClickHouse | 601,087,505 | 9,726 | alexey-milovidov | > Integration tests (release) Pending — Started
Did not finish in CI. | 2020-03-19T09:52:24 | 2020-03-19T09:52:24 | {} | MEMBER |
ClickHouse | ClickHouse | 601,121,536 | 9,721 | alexey-milovidov | Ok, now we have to patch it, because it's a little regresstion. | 2020-03-19T11:10:50 | 2020-03-19T11:10:50 | {} | MEMBER |
ClickHouse | ClickHouse | 601,263,276 | 9,715 | 4ertus2 | I think we could have all described cases under the setting: prefer column, prefer alias, throw an error. With 'prefer column' as default cause it's most SQL compatible.
`prefer_on_column_clashes` `column | alias | error`.
| 2020-03-19T16:01:42 | 2020-03-19T16:01:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 601,348,489 | 9,154 | alexey-milovidov | @azat query profiler test was temporary broken in master a few weeks ago, then fixed.
See https://github.com/ClickHouse/ClickHouse/pull/9472#issuecomment-593719032 | 2020-03-19T18:35:42 | 2020-03-19T18:38:10 | {} | MEMBER |
ClickHouse | ClickHouse | 601,374,364 | 9,715 | alexey-milovidov | @amosbird
> If we choose the second route, perhaps we can also disable alias for certain cases like #8329
It still looks too hacky. | 2020-03-19T19:26:07 | 2020-03-19T19:26:07 | {} | MEMBER |
ClickHouse | ClickHouse | 601,412,710 | 9,715 | alex-zaitsev | > @alex-zaitsev No, this query won't work with the proposed solution.
That's why I think you option #1 is better. It would work for this case as well. | 2020-03-19T20:56:18 | 2020-03-19T20:56:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 600,933,527 | 9,722 | wangzhen11aaa | > > But the clickhouse actually will update the partitions with adding the mutation_versions +1
>
> CH always mutates and adds the mutation_versions +1 **to ALL PARTS.** (consistency requirement)
> New versions / new parts are created for alll parts using **HARDLINKS**
> Only new parts those match `where dt='2020-... | 2020-03-19T01:07:14 | 2020-03-19T01:07:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 601,076,687 | 9,673 | KochetovNicolai | ```c++
if (max_streams > 1 && !is_remote)
max_streams *= settings.max_streams_to_max_threads_ratio;
```
I think we need to limit the number of threads here too. | 2020-03-19T09:27:40 | 2020-03-19T09:27:40 | {} | MEMBER |
ClickHouse | ClickHouse | 601,088,439 | 9,574 | alexey-milovidov | How is it going? | 2020-03-19T09:54:25 | 2020-03-19T09:54:25 | {} | MEMBER |
ClickHouse | ClickHouse | 601,096,518 | 9,706 | VonShtirlitz | 
I looked into the postinstall code and also did not find a problem. Сlean installation showed that everything is ok. most likely a problem elsewhere.
close the issue | 2020-03-19T10:12:52 | 2020-03-19T10:12:52 | {} | NONE |
ClickHouse | ClickHouse | 601,471,003 | 9,715 | alexey-milovidov | Ok, reasonable... | 2020-03-20T00:03:45 | 2020-03-20T00:03:45 | {} | MEMBER |
ClickHouse | ClickHouse | 600,567,117 | 9,681 | l1t1 | i compared the pdfs, not webpages
namely
https://clickhouse.tech/docs/zh/single/clickhouse_zh.pdf page 149
and
https://clickhouse.tech/docs/en/single/clickhouse_en.pdf page 151
and i tried many pdf readers, they all shows the same results. | 2020-03-18T11:21:58 | 2020-03-18T11:21:58 | {} | NONE |
ClickHouse | ClickHouse | 600,570,743 | 9,681 | l1t1 | and tables of the webpages are all align.
https://clickhouse.tech/docs/en/operations/table_engines/merge/
all contents are in blcaks
https://clickhouse.tech/docs/zh/operations/table_engines/merge/
some parts of the contents are in colors | 2020-03-18T11:31:02 | 2020-03-18T11:31:02 | {} | NONE |
ClickHouse | ClickHouse | 600,872,920 | 7,306 | filimonov | Can you try:
```
ALTER TABLE mydb.test3 UPDATE i = toInt64(5756242267) WHERE id =1
``` | 2020-03-18T21:41:24 | 2020-03-18T21:41:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 601,078,098 | 9,673 | azat | >I think we need to limit the number of threads here too.
So revert the [last change](https://github.com/ClickHouse/ClickHouse/pull/9673#discussion_r394522947)? | 2020-03-19T09:31:12 | 2020-03-19T09:31:12 | {} | MEMBER |
ClickHouse | ClickHouse | 601,130,151 | 9,715 | amosbird | > You can write `SELECT x + 1 AS x, x FROM (SELECT 1 AS x) t`
That doesn't hide the original column under the SQL standard iiuc. If we choose the second route, perhaps we can also disable alias for certain cases like https://github.com/ClickHouse/ClickHouse/pull/8329 ? | 2020-03-19T11:32:55 | 2020-03-19T11:37:31 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 601,363,159 | 9,721 | alexey-milovidov | @amosbird Sorry, one more minor issue has been found.
If you type `Merg` (with intent to find MergeTree table) and press TAB, then the line is replaced to lowercase `merge` despite the fact that the `MergeTree` variant is also present in the completion list. | 2020-03-19T19:06:55 | 2020-03-19T19:06:55 | {} | MEMBER |
ClickHouse | ClickHouse | 579,637,106 | 8,573 | filimonov | > And version of openssl ?
At that point CH should use bundled openssl (unless you rebuild ClickHouse with custom flags).
20.1 actually uses trunc version [3.0.0-dev](https://github.com/ClickHouse-Extras/openssl/blob/c74e7895eb0d219d4007775eec134dd2bcd9d1ae/VERSION) directly from master branch with few customiza... | 2020-01-29T07:53:47 | 2020-01-29T08:53:10 | {} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.