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 | 658,735,446 | 11,364 | akuzm | `SELECT arrayMap(x -> arrayJoin([1]), [1, 2])` | 2020-07-15T12:22:09 | 2020-07-15T12:22:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 658,078,960 | 12,322 | wanghaifen | it is not work sometimes single node drop，it is not work sometimes use on cluster ck_test_cluster。
accessory is the clickhouse cluster configuration files and the detailed information of the bug and the create table sql file of ReplicatedMergeTree table,Distributed table
I hope I can make that clear,Thanks for the... | 2020-07-14T09:33:16 | 2020-07-14T09:33:16 | {} | NONE |
ClickHouse | ClickHouse | 658,103,973 | 10,223 | blinkov | 1. There's actually a much more serious issue with rendering around page breaks, I've tried some easy approaches, but they didn't work.
2. As of “tables” that are pretty-printed query results, we'd better just get rid of them in favor of normal HTML tables.
3. Why font stops being monospace during print also deserves... | 2020-07-14T10:30:19 | 2020-07-14T10:30:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 658,129,961 | 10,561 | alesapin | `test_moves_to_disk_work` no fix in test in this branch. It's ok. | 2020-07-14T11:34:59 | 2020-07-14T11:34:59 | {} | MEMBER |
ClickHouse | ClickHouse | 658,168,040 | 12,443 | akuzm | > Ok, probably it is because I did not merge with master and performance of master was improved.
> @akuzm Why it is not merged with master before running tests?
It's more complicated -- it choose the master for comparison that is newer than the master it was merged to before build, because the perf test task cloned... | 2020-07-14T13:06:37 | 2020-07-14T13:06:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 658,336,812 | 5,956 | alexey-milovidov | This behaviour is better, because there will be no benefits for overflow.
The settings `decimal_check_overflow` only applied on calculations on decimals, not for parsing.
Please add the relevant info to the documentation. | 2020-07-14T18:21:24 | 2020-07-14T18:21:24 | {} | MEMBER |
ClickHouse | ClickHouse | 658,338,007 | 7,442 | alexey-milovidov | Fixed in https://github.com/ClickHouse/ClickHouse/pull/4243
```
ALTER TABLE test_decimal_array
UPDATE v = CAST([2.], 'Array(Decimal(38, 18))') WHERE d = today()
Ok.
``` | 2020-07-14T18:23:47 | 2020-07-14T18:23:47 | {} | MEMBER |
ClickHouse | ClickHouse | 658,499,561 | 12,488 | alexey-milovidov | CC @filimonov @azat | 2020-07-15T01:54:10 | 2020-07-15T01:54:10 | {} | MEMBER |
ClickHouse | ClickHouse | 658,741,590 | 12,520 | akuzm | What is a `LEFT SEMI JOIN`? `SEMI` selects left side where matching right side exists, and `LEFT` selects left side always. Together it doesn't make sense. I guess our definition of semi join includes right side columns, so this should select all left rows and any matching right row? Saying this is more standard is a s... | 2020-07-15T12:36:21 | 2020-07-15T12:36:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 658,057,764 | 12,195 | alexey-milovidov | We need this feature :) | 2020-07-14T08:49:59 | 2020-07-14T08:49:59 | {} | MEMBER |
ClickHouse | ClickHouse | 658,066,732 | 12,322 | wanghaifen | it is not work sometimes single node drop
| |
18101350634
|
|
邮箱：18101350634@163.com
|
Signature is customized by Netease Mail Master
On 07/10/2020 04:37, Denis Zhuravlev wrote:
BTW I don't understand is it work OK or not
single node drop:
ALTER TABLE market.vb_xue_order_sku_user_lesson_detail_5_minutes_replica... | 2020-07-14T09:08:04 | 2020-07-14T09:08:04 | {} | NONE |
ClickHouse | ClickHouse | 658,199,815 | 10,594 | wutchzone | I am sorry that it took me so long. For safety a hid IP addresses
*When error is not present*:
```
sudo netstat -nap
Active Internet connections (servers and established)
Proto Recv-Q Send-Q Local Address Foreign Address State PID/Program name
tcp 0 0 hidden:9000 h... | 2020-07-14T14:04:49 | 2020-07-14T14:04:49 | {} | NONE |
ClickHouse | ClickHouse | 658,357,507 | 7,701 | alexey-milovidov | @CurtizJ Could you please clarify the status of this issue? | 2020-07-14T19:03:50 | 2020-07-14T19:03:50 | {} | MEMBER |
ClickHouse | ClickHouse | 658,361,953 | 12,499 | ksvladimir | Actually even more generic solution could be `neighbor(id, -1) != id`, but it has the same problem: it works in SELECT but not in WHERE. | 2020-07-14T19:12:54 | 2020-07-14T19:12:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 658,502,184 | 4,586 | alexey-milovidov | @vitlibar has fixed it in https://github.com/ClickHouse/ClickHouse/pull/10234 exactly as it was proposed. | 2020-07-15T02:04:16 | 2020-07-15T02:04:16 | {} | MEMBER |
ClickHouse | ClickHouse | 658,523,275 | 12,432 | alexey-milovidov | Just throw exception if FORMAT is specified for query sent via MySQL or PostgreSQL protocol.
About
```
explain select 1;
```
default format should not override MySQLWire (I'm not sure about the specific mechanics in code, but it should work this way). | 2020-07-15T03:24:19 | 2020-07-15T03:24:19 | {} | MEMBER |
ClickHouse | ClickHouse | 658,637,625 | 12,294 | alesapin | 
| 2020-07-15T08:51:10 | 2020-07-15T08:51:10 | {} | MEMBER |
ClickHouse | ClickHouse | 658,741,090 | 12,516 | 4ertus2 | Most probable you have not deterministic result of the query. It's not correlated to `OUTER` keyword for sure. This word is just ignored at parser stage of query analisys.
In future version it could force `ALL JOIN` to override `set join_default_strictness=ANY`. But for now parser reads and ignores `OUTER` keyword. | 2020-07-15T12:35:07 | 2020-07-15T12:41:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 658,259,768 | 12,081 | den-crane | @calebeaires
https://repo.yandex.ru/clickhouse/deb/prestable/main/
```
cat /etc/apt/sources.list.d/clickhouse.list
deb http://repo.yandex.ru/clickhouse/deb/stable/ main/
deb http://repo.yandex.ru/clickhouse/deb/testing/ main/
deb http://repo.yandex.ru/clickhouse/deb/prestable/ main/
apt-get update
apt-... | 2020-07-14T15:52:14 | 2020-07-14T15:52:31 | {
"hooray": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 658,324,131 | 12,288 | alexey-milovidov | > clickhouse-client ignores queries after insert in multiquery mode
The data for insert query is parsed until end of line. | 2020-07-14T17:55:32 | 2020-07-14T17:55:32 | {} | MEMBER |
ClickHouse | ClickHouse | 658,348,293 | 12,498 | alexey-milovidov | There is nothing special about connection timeout. | 2020-07-14T18:44:44 | 2020-07-14T18:44:44 | {} | MEMBER |
ClickHouse | ClickHouse | 658,352,390 | 6,813 | alexey-milovidov | @vitlibar Please check if this issue still exists. | 2020-07-14T18:53:24 | 2020-07-14T18:53:24 | {} | MEMBER |
ClickHouse | ClickHouse | 658,381,673 | 12,499 | ksvladimir | That indeed works! In fact, even this works:
```sql
SELECT *
FROM
(
SELECT *
FROM tbl AS dst
ARRAY JOIN related
WHERE related IN
(
SELECT id
FROM tbl AS src
WHERE data < 3
)
)
WHERE runningDifferenceStartingWithFirstValue(id) != 0
┌─id─┬─data─┬─relat... | 2020-07-14T19:55:55 | 2020-07-14T19:55:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 658,369,782 | 12,499 | den-crane | hmm,
```
SELECT id
FROM
(
SELECT
id,
(runningDifferenceStartingWithFirstValue(id)) AS d
FROM tbl
ARRAY JOIN related
WHERE related IN
(
SELECT id
FROM tbl AS src
WHERE data < 3
)
)
WHERE d != 0
┌─id─┐
│ 2 │
│ 3 │
│ 4 │
└────┘
... | 2020-07-14T19:29:07 | 2020-07-14T19:51:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 658,405,343 | 11,731 | alexey-milovidov | I don't think it's easy to do, let's ask @alesapin who has implemented RENAME COLUMN. | 2020-07-14T20:48:08 | 2020-07-14T20:48:08 | {} | MEMBER |
ClickHouse | ClickHouse | 658,428,064 | 12,503 | alexey-milovidov | @nikvas0 A bug in DatabaseLazy has been found by stress test:
```
milovidov@milovidov-desktop:~/Downloads$ grep --text -F '[ 26936 ]' clickhouse-server.log
2020.07.14 23:31:18.349876 [ 26936 ] {} <Fatal> BaseDaemon: ########################################
2020.07.14 23:31:18.350695 [ 26936 ] {} <Fatal> BaseDaem... | 2020-07-14T21:41:28 | 2020-07-14T21:41:28 | {} | MEMBER |
ClickHouse | ClickHouse | 658,485,701 | 10,429 | l1t1 | @den-crane thanks | 2020-07-15T01:00:08 | 2020-07-15T01:00:08 | {} | NONE |
ClickHouse | ClickHouse | 658,492,612 | 12,506 | alexey-milovidov | > 1359.79 sec. - Timeout!
To execute this test, another modification is required: https://github.com/ClickHouse/ClickHouse/pull/12505 | 2020-07-15T01:26:49 | 2020-07-15T01:26:49 | {} | MEMBER |
ClickHouse | ClickHouse | 658,535,184 | 12,444 | bharatnc | I presume that this table will contain everything that's available under `/usr/share/zoneinfo` ? | 2020-07-15T04:14:36 | 2020-07-15T04:14:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 658,669,063 | 12,382 | alesapin | I'm not completely sure, about theses restrictions, but I think it's OK as a temporary solution (better than stucked mutation).
| 2020-07-15T09:50:08 | 2020-07-15T09:50:08 | {} | MEMBER |
ClickHouse | ClickHouse | 658,062,501 | 9,624 | alesapin | UBsan failure doesn't lead to a real bug. | 2020-07-14T08:59:50 | 2020-07-14T08:59:58 | {} | MEMBER |
ClickHouse | ClickHouse | 658,343,277 | 11,591 | alexey-milovidov | You mean:
"
I need to support for empty data stream in `RowBinaryWithNamesAndTypes`.
Although it is definitely incorrect (no names and types provided), it will be convenient in some places in the code for implementation of some features that I need.
" | 2020-07-14T18:34:15 | 2020-07-14T18:34:15 | {} | MEMBER |
ClickHouse | ClickHouse | 658,354,121 | 7,754 | alexey-milovidov | Fixed in recent releases. | 2020-07-14T18:57:06 | 2020-07-14T18:57:06 | {} | MEMBER |
ClickHouse | ClickHouse | 658,364,362 | 12,501 | calebeaires | Here is how I could manage this.
1. First: find all mutation passing database by database
```sql
clickhouse-client --user=SOMEUSER --passoword=SOMEPASS
select * from system.mutations where database = 'db_1';
KILL MUTATION where database = 'db_1' SYNC;
select * from system.mutations where database = 'db_2'... | 2020-07-14T19:17:57 | 2020-07-14T19:17:57 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 658,403,130 | 1,053 | alexey-milovidov | We don't consider using *shared libraries* for implementation of plugins due to maintainability issues.
In fact, most of the usages of shared libraries were removed from ClickHouse, examples:
- compiled queries via runtime template instantiation - deprecated after #6435 in #6646 (comiled LLVM expressions still exist)... | 2020-07-14T20:43:15 | 2020-07-14T20:44:26 | {} | MEMBER |
ClickHouse | ClickHouse | 658,664,881 | 11,737 | 4ertus2 | Continuation is here #12515 | 2020-07-15T09:42:10 | 2020-07-15T09:42:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 658,747,048 | 12,520 | 4ertus2 | > What is a `LEFT SEMI JOIN`?
It is present on internet. Ex. https://stackoverflow.com/questions/21738784/difference-between-inner-join-and-left-semi-join
`SEMI` means not `INNER` or `OUTER`. `SEMI JOIN` is `LEFT SEMI JOIN` with omitted `LEFT`. It's just like `JOIN` is `INNER JOIN` with omitted `INNER`. | 2020-07-15T12:48:38 | 2020-07-15T12:50:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 658,224,809 | 12,496 | alesapin | Also, it requires a small fix in CI. | 2020-07-14T14:49:16 | 2020-07-14T14:55:09 | {} | MEMBER |
ClickHouse | ClickHouse | 658,286,779 | 12,499 | den-crane | 1. `LIMIT BY` has no sense without `ORDER BY`
2. CH SQL allows to omit grouped columns in select section but in reality you query does `SELECT count(), id FROM limit_by` , and it needs `group by id` section.
Seems you need
```
SELECT uniqExact(id) FROM limit_by
┌─uniqExact(id)─┐
│ 2 │
└──────... | 2020-07-14T16:40:33 | 2020-07-14T16:45:59 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 658,334,155 | 4,823 | alexey-milovidov | https://github.com/ClickHouse/ClickHouse/pull/10388 | 2020-07-14T18:16:03 | 2020-07-14T18:16:03 | {} | MEMBER |
ClickHouse | ClickHouse | 658,344,285 | 10,429 | alexey-milovidov | The format is available in version 20.6-prestable. | 2020-07-14T18:36:21 | 2020-07-14T18:36:21 | {} | MEMBER |
ClickHouse | ClickHouse | 658,344,230 | 12,499 | den-crane | Probably there is no way to avoid arrayJoin
you can use `uniqExactIf(id,data>2)` (or uniqExactIf(data,data>2))
```sql
SELECT
uniqExactIf(id, data > 2) AS c1,
uniqExactIf(id, data > 3) AS c2
FROM tbl
WHERE arrayJoin(related) IN
(
SELECT id
FROM tbl
WHERE data < 3
)
┌─c1─┬─c2─┐
│ 2 │ ... | 2020-07-14T18:36:14 | 2020-07-14T18:36:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 658,494,211 | 11,815 | alexey-milovidov | Continued in https://github.com/ClickHouse/ClickHouse/issues/12508 | 2020-07-15T01:33:32 | 2020-07-15T01:33:32 | {} | MEMBER |
ClickHouse | ClickHouse | 658,522,631 | 12,443 | alexey-milovidov | Performance test Ok. | 2020-07-15T03:21:58 | 2020-07-15T03:21:58 | {} | MEMBER |
ClickHouse | ClickHouse | 658,752,198 | 12,520 | akuzm | > > What is a `LEFT SEMI JOIN`?
>
> It is present on internet. Ex. https://stackoverflow.com/questions/21738784/difference-between-inner-join-and-left-semi-join
> `SEMI` means not `INNER` or `OUTER`. `SEMI JOIN` is `LEFT SEMI JOIN` with omitted `LEFT`. It's just like `JOIN` is `INNER JOIN` with omitted `INNER`.
... | 2020-07-15T12:59:36 | 2020-07-15T12:59:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 658,097,403 | 12,386 | alexey-milovidov | Ok. | 2020-07-14T10:14:53 | 2020-07-14T10:14:53 | {} | MEMBER |
ClickHouse | ClickHouse | 658,318,014 | 12,499 | den-crane | Please provide the data sample and desired result.
I don't think you need ` limit by / ARRAY JOIN` . From your description you need just `count / group by ` | 2020-07-14T17:43:18 | 2020-07-14T17:43:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 658,349,167 | 12,498 | alexey-milovidov | It uses port 90090 because some replica advertised this port as the port that should be used for connection.
The port is advertised in ZooKeeper in replicas/replica_name/host node.
Most likely it was configured this way. | 2020-07-14T18:46:36 | 2020-07-14T18:46:36 | {} | MEMBER |
ClickHouse | ClickHouse | 658,459,924 | 12,498 | alexey-milovidov | And the issue is fixed long time ago: https://github.com/ClickHouse/ClickHouse/pull/9221 | 2020-07-14T23:21:34 | 2020-07-14T23:21:34 | {} | MEMBER |
ClickHouse | ClickHouse | 658,729,132 | 12,517 | blinkov | superseded by #12520 | 2020-07-15T12:06:38 | 2020-07-15T12:06:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 658,061,916 | 9,624 | alesapin | Asan has known failure
```
==273==ERROR: AddressSanitizer: allocator is out of memory trying to allocate 0x20000 bytes
==273==ERROR: AddressSanitizer failed to allocate 0x2000 (8192) bytes of InternalMmapVector (error code: 12)
ERROR: Failed to mmap
```
| 2020-07-14T08:58:39 | 2020-07-14T08:58:39 | {} | MEMBER |
ClickHouse | ClickHouse | 658,297,365 | 12,499 | ksvladimir | Thanks for the quick answer! Understanding that `id` is being added as a queried column helps, and the suggested query solves my question.
The underlying problem I'm trying to solve is slightly more complex – I'd like to perform `uniqExact` over one column, but aggregate over other columns:
```sql
SELECT countIf... | 2020-07-14T17:01:22 | 2020-07-14T17:01:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 658,331,633 | 12,499 | ksvladimir | **Details** (I tried to simplify it and remove all unrelated stuff):
```sql
CREATE TABLE tbl (id UInt32, data UInt32, related Array(UInt32)) ENGINE MergeTree ORDER BY id;
INSERT INTO tbl VALUES (1, 1, []), (2, 2, [1]), (3, 3, [1]), (4, 4, [1, 2])
```
```
SELECT * FROM tbl
┌─id─┬─data─┬─related─┐
│ 1 │ 1... | 2020-07-14T18:10:57 | 2020-07-14T18:10:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 658,523,604 | 12,503 | alexey-milovidov | All test failures are reported in #12508 | 2020-07-15T03:25:28 | 2020-07-15T03:25:28 | {} | MEMBER |
ClickHouse | ClickHouse | 658,060,826 | 9,624 | alesapin | msan failures are not interesting, no supressions in SSL and infinite recursion in libunwind. | 2020-07-14T08:56:20 | 2020-07-14T08:56:20 | {} | MEMBER |
ClickHouse | ClickHouse | 658,340,843 | 4,903 | alexey-milovidov | Current status: ORC output format is available since version 20.6. | 2020-07-14T18:29:27 | 2020-07-14T18:29:27 | {} | MEMBER |
ClickHouse | ClickHouse | 658,341,749 | 11,932 | alexey-milovidov | @santoshdvn
> hdfs dfs -cat <hdfs_parquetfiles> | ~/ch/chc -h -u --password --query="insert into test.testing_clickhouse_FORMAT Parquet";
Your query has extra underscore.
The correct query is:
> hdfs dfs -cat <hdfs_parquetfiles> | ~/ch/chc -h -u --password --query="insert into test.testing_clickhouse FORMAT ... | 2020-07-14T18:31:13 | 2020-07-14T18:31:13 | {} | MEMBER |
ClickHouse | ClickHouse | 658,358,727 | 7,673 | alexey-milovidov | Workaround: `toDateTimeOrZero`. | 2020-07-14T19:06:21 | 2020-07-14T19:06:21 | {} | MEMBER |
ClickHouse | ClickHouse | 658,422,944 | 12,458 | alexey-milovidov | Yandex synchronization check is broken after merging this PR. | 2020-07-14T21:28:24 | 2020-07-14T21:28:24 | {} | MEMBER |
ClickHouse | ClickHouse | 658,479,263 | 10,429 | l1t1 | I downloaded the latest rpm file (version 20.7.1.4080) ,it shows error, isn't the format merged into it?
```
ClickHouse client version 20.7.1.4080 (official build).
Connecting to localhost:9000 as user default.
Connected to ClickHouse server version 20.7.1 revision 54437.
2aad45d60838 :) select version();
SEL... | 2020-07-15T00:32:40 | 2020-07-15T00:32:40 | {} | NONE |
ClickHouse | ClickHouse | 658,269,313 | 12,498 | den-crane | try to upgrade to 20.1.16.120 | 2020-07-14T16:08:50 | 2020-07-14T16:08:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 658,286,073 | 12,293 | 4ertus2 | We could have troubles with `prefer_alias_to_column_name` logic if we enable it. Probably we need some investigation here. | 2020-07-14T16:39:15 | 2020-07-14T16:39:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 658,324,346 | 12,288 | alexey-milovidov | It always worked this way.
We can add exception for VALUES format. | 2020-07-14T17:55:55 | 2020-07-14T17:55:55 | {} | MEMBER |
ClickHouse | ClickHouse | 658,338,761 | 7,306 | alexey-milovidov | We should apply type conversion automatically. | 2020-07-14T18:25:19 | 2020-07-14T18:25:19 | {} | MEMBER |
ClickHouse | ClickHouse | 658,425,482 | 12,498 | Tri0L | No luck.
Version: 20.1.16.120
```
2020.07.14 21:33:15.519596 [ 8 ] {} <Error> db.table: DB::StorageReplicatedMergeTree::queueTask()::<lambda(DB::StorageReplicatedMergeTree::LogEntryPtr&)>: Poco::Exception. Code: 1000, e.code() = 0, e.displayText() = Timeout: connect timed out: 10.2.1.24:90090. 0x102b981f Poco::E... | 2020-07-14T21:34:30 | 2020-07-14T21:34:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 658,492,926 | 11,815 | alexey-milovidov | 00933_test_fix_extra_seek_on_compressed_cache
https://clickhouse-test-reports.s3.yandex.net/12503/cd7ed94a085d1c810bff7a84abca1a1786e7e2a2/functional_stateless_tests_(release,_databaseatomic)/test_run.txt.out.log
Maybe another read can happen during select query (system log)? | 2020-07-15T01:28:16 | 2020-07-15T01:28:16 | {} | MEMBER |
ClickHouse | ClickHouse | 658,501,111 | 7,680 | alexey-milovidov | @vitlibar Please clarify how (and if) it's possible to configure with the new RBAC feature. | 2020-07-15T02:00:20 | 2020-07-15T02:00:20 | {} | MEMBER |
ClickHouse | ClickHouse | 658,741,394 | 12,294 | alesapin | Ok, I think we can merge it now, but first let's check that it can be built with normal builds. | 2020-07-15T12:35:50 | 2020-07-15T12:40:00 | {} | MEMBER |
ClickHouse | ClickHouse | 658,159,366 | 12,407 | akuzm | The test failures seem to be the same as in 20.3 branch, I'll merge. | 2020-07-14T12:47:49 | 2020-07-14T12:47:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 658,253,815 | 12,081 | calebeaires | @alexey-milovidov is there some way to test it? I have found one bug os alter table deletion that it is a problem on of my server, but not in another.
I have tried to update like this, but the release does not have the **next** version
`sudo apt install clickhouse-server=20.6` | 2020-07-14T15:41:13 | 2020-07-14T15:41:13 | {} | NONE |
ClickHouse | ClickHouse | 658,323,799 | 12,288 | alexey-milovidov | > Same in interactive mode - try to paste those into clickhouse-client -mn interactive session
Does not reproduce.
```
milovidov@milovidov-desktop:~/work/ClickHouse$ clickhouse-client -n
ClickHouse client version 20.7.1.1.
Connecting to localhost:9000 as user default.
Connected to ClickHouse server version 20... | 2020-07-14T17:54:50 | 2020-07-14T17:54:50 | {} | MEMBER |
ClickHouse | ClickHouse | 658,346,473 | 7,814 | alexey-milovidov | It can be implemented with two independent settings:
- `output_format_json_each_row_pretty_print` for indentation and whitespaces;
- `output_format_json_each_row_enclose_in_array` to enclose everything in single array; | 2020-07-14T18:40:47 | 2020-07-14T18:40:47 | {} | MEMBER |
ClickHouse | ClickHouse | 658,357,235 | 12,499 | ksvladimir | Awesome, `uniqExactIf` is exactly what I needed! It made me appreciate the power of combinations, thanks.
In the meantime, I found another solution that almost works and is about 50% faster:
```
SELECT countIf(data>2) AS c1, countIf(data>3) AS c2
FROM tbl ARRAY JOIN related
WHERE related IN (SELECT id FROM tbl W... | 2020-07-14T19:03:17 | 2020-07-14T19:03:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 658,507,094 | 11,666 | alexey-milovidov | https://github.com/ClickHouse/ClickHouse/pull/11666/files#diff-9d5603f2961d81cbf91c69c0d568399fR1
This is derivative work but you forgot to add information about license and author. | 2020-07-15T02:23:27 | 2020-07-15T02:23:27 | {} | MEMBER |
ClickHouse | ClickHouse | 658,539,953 | 12,302 | azat | > Do I understand correctly that it's not worth adding tests?
Yes (it won't be determined) | 2020-07-15T04:34:56 | 2020-07-15T04:34:56 | {} | MEMBER |
ClickHouse | ClickHouse | 658,558,943 | 3,162 | hzruandd | > We are experiencing a similar issue, using the same ClickHouse version (18.12.17).
>
> We are running a multi-tenant setup with separate databases, all having the same tables. The issue occurs on different databases, but thus far only with the below specific query (according to our logs).
>
> We have checked lo... | 2020-07-15T05:46:01 | 2020-07-15T05:46:01 | {} | NONE |
ClickHouse | ClickHouse | 658,564,721 | 12,510 | ritaank | When using `CREATE ROLE` in a similar situation, the 'actual' exitcode seems to be `0`, not `237` as shown in the `CREATE USER` example above. It seems that this exitcode of `0` is unique to `CREATE ROLE`. Details below.
```
ClickHouse client version 20.6.1.4066 (official build).
Connecting to localhost:9000 as us... | 2020-07-15T06:05:10 | 2020-07-15T06:05:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 658,738,423 | 10,690 | XavierQuiniou | Just a few words to conclude.
Our problem was solved by completing <tt>escape_sequences # replaceEscapeSequences (..)</tt>, with a step
additional parsing. We have written a parser that builds a very simple token tree.
The branches are created by the expression detection <tt>(.....)</tt>.
This technique thus allo... | 2020-07-15T12:29:10 | 2020-07-15T12:29:10 | {} | NONE |
ClickHouse | ClickHouse | 658,062,124 | 9,624 | alesapin | Also no integration tests. Run them manually. | 2020-07-14T08:59:04 | 2020-07-14T08:59:04 | {} | MEMBER |
ClickHouse | ClickHouse | 658,164,369 | 12,489 | den-crane | probably the same as https://github.com/ClickHouse/ClickHouse/issues/10471
default database is used on shards
Technically it's not a bug, because this behavior existed from the beginning, though I believe it's not documented.
```
use system
select * from remote(two_shards,system,one) where dummy in (select d... | 2020-07-14T12:58:55 | 2020-07-14T13:26:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 658,178,963 | 10,223 | l1t1 | yes, it is not only the table align question, but also the mono font for script and keywords question. | 2020-07-14T13:27:56 | 2020-07-14T13:27:56 | {} | NONE |
ClickHouse | ClickHouse | 658,358,088 | 7,697 | alexey-milovidov | @filimonov Please either accept or close. | 2020-07-14T19:05:01 | 2020-07-14T19:05:01 | {} | MEMBER |
ClickHouse | ClickHouse | 658,458,213 | 12,498 | alexey-milovidov | The port value is 9009 as expected, but the message in log is concatenated. How it should read:
```
2020.07.14 21:33:15.519596 [ 8 ] {} <Error> db.table: DB::StorageReplicatedMergeTree::queueTask()::<lambda(DB::StorageReplicatedMergeTree::LogEntryPtr&)>: Poco::Exception. Code: 1000, e.code() = 0, e.displayText() = Ti... | 2020-07-14T23:15:31 | 2020-07-14T23:15:31 | {} | MEMBER |
ClickHouse | ClickHouse | 658,497,678 | 12,496 | alexey-milovidov | > Stress test (address) — Some exception occurred :(
Failure in "Sandbox" infrastructure - it does not support operation when single datacenter is not available.
> Exception: Path with logs doesn't exists: /place/sandbox-data/tasks/2/1/724017812/server_log/clickhouse-server.log or /place/sandbox-data/tasks/2/1/7... | 2020-07-15T01:46:38 | 2020-07-15T01:48:37 | {} | MEMBER |
ClickHouse | ClickHouse | 658,421,007 | 12,418 | alexey-milovidov | > Irrelevant failure due to test is flaky.
BTW, performance test has failed not because it is flaky but because performance on master has been improved. | 2020-07-14T21:23:41 | 2020-07-14T21:23:41 | {} | MEMBER |
ClickHouse | ClickHouse | 658,485,026 | 10,429 | den-crane |
@l1t1
https://github.com/s-mx/ClickHouse/blob/fd46ef428a0176cd70845de824bcfd141bb7b4b2/docs/en/operations/settings/settings.md#output_format_pretty_grid_charset-output-format-pretty-grid-charset
```
SET output_format_pretty_grid_charset = 'ASCII';
SELECT version()
+-version()---+
| 20.6.1.4066 |
+------... | 2020-07-15T00:57:29 | 2020-07-15T00:58:56 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 658,493,218 | 11,815 | alexey-milovidov | 00981_topK_topKWeighted_long:
https://clickhouse-test-reports.s3.yandex.net/12278/e1543c18414e566fe7cbaaa3342e061357dfbf51/functional_stateless_tests_(release,_databaseatomic)/test_run.txt.out.log
Reproduced second time since 2020-05-26 02:29:39. | 2020-07-15T01:29:27 | 2020-07-15T01:29:27 | {} | MEMBER |
ClickHouse | ClickHouse | 658,741,313 | 12,294 | alesapin | 
| 2020-07-15T12:35:37 | 2020-07-15T12:35:37 | {} | MEMBER |
ClickHouse | ClickHouse | 658,752,872 | 12,511 | CurtizJ | Is the field `latest_fail_reason` empty in `system.mutations` table? And are there any exceptions in log?
Also it will be helpful, if you will share the create statement of table. | 2020-07-15T13:00:53 | 2020-07-15T13:01:44 | {} | MEMBER |
ClickHouse | ClickHouse | 658,328,847 | 3,326 | alexey-milovidov | clickhouse-client reconnects to the server on client-side errors. This is required due to possible protocol desynchronization errors.
Asking of password is necessary because client does not keep entered password for further reuse, because it's insecure
(but it can still reside in memory).
Although reconnection is ... | 2020-07-14T18:05:11 | 2020-07-14T18:05:11 | {} | MEMBER |
ClickHouse | ClickHouse | 658,428,300 | 12,503 | alexey-milovidov | Arcadia (Yandex Synchronization) is broken in this PR: https://github.com/ClickHouse/ClickHouse/pull/12458 | 2020-07-14T21:42:07 | 2020-07-14T21:42:07 | {} | MEMBER |
ClickHouse | ClickHouse | 658,087,150 | 12,386 | fastio | > Another question - do you feel that it's necessary to have a separate name for this storage / table function?
> Maybe if s3 will just magically support COS is also Ok?
We have integrated COS with Appache/Hive,Spark,Alluxio,HDFS etc.
Our customers are familiar with keyword COS, and using it on other systems.
Now... | 2020-07-14T09:51:50 | 2020-07-14T09:51:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 658,096,522 | 10,223 | alexey-milovidov | @blinkov I have downloaded the latest version and can confirm that tables don't align. | 2020-07-14T10:12:46 | 2020-07-14T10:13:03 | {} | MEMBER |
ClickHouse | ClickHouse | 658,298,636 | 12,482 | 4ertus2 | Related #4567 | 2020-07-14T17:03:55 | 2020-07-14T17:03:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 658,331,694 | 8,247 | alexey-milovidov | @CurtizJ Could you please look why index analysis does not respect type changes due to ARRAY JOIN? | 2020-07-14T18:11:03 | 2020-07-14T18:11:03 | {} | MEMBER |
ClickHouse | ClickHouse | 658,350,563 | 7,317 | alexey-milovidov | > a function with a null argument should ALWAYS produce a null result.
It already works this way and always worked this way.
```
SELECT toYear(NULL)
┌─toYear(NULL)─┐
│ ᴺᵁᴸᴸ │
└──────────────┘
``` | 2020-07-14T18:49:39 | 2020-07-14T18:49:39 | {} | MEMBER |
ClickHouse | ClickHouse | 658,353,730 | 7,764 | alexey-milovidov | Fixed in recent ClickHouse releases. | 2020-07-14T18:56:13 | 2020-07-14T18:56:13 | {} | MEMBER |
ClickHouse | ClickHouse | 658,554,736 | 12,507 | ritaank | This seems to be a general occurrence, occurring not just in ```CREATE USER```, but every time the ```IF NOT EXISTS``` clause is used when the designated object with that name already does exist. | 2020-07-15T05:30:42 | 2020-07-15T05:30:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,405,458 | 12,266 | akuzm | ```
Performance — 2 slower, 139 unstable
```
Nice, this is better than the usual 200 unstable queries. | 2020-07-08T09:30:27 | 2020-07-08T09:30:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,440,147 | 11,781 | filimonov | https://github.com/ClickHouse/ClickHouse/commit/4d01fb3cbce2971f19a66ac97e85092523487125 | 2020-07-08T10:43:57 | 2020-07-08T10:43:57 | {} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.