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 | 733,923,025 | 17,275 | azat | >Conflicting files
>.gitmodules
Rebased | 2020-11-25T20:07:23 | 2020-11-25T20:07:23 | {} | MEMBER |
ClickHouse | ClickHouse | 734,009,141 | 17,376 | alexey-milovidov | Changelog is using PR description - it's Ok. | 2020-11-26T00:39:26 | 2020-11-26T00:39:26 | {} | MEMBER |
ClickHouse | ClickHouse | 734,015,969 | 17,385 | alexey-milovidov | I have validated the instruction once again:
```
$ curl https://clickhouse-datasets.s3.yandex.net/hits/tsv/hits_v1.tsv.xz | unxz --threads=`nproc` > hits_v1.tsv
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left S... | 2020-11-26T01:06:51 | 2020-11-26T01:07:23 | {} | MEMBER |
ClickHouse | ClickHouse | 733,694,631 | 16,804 | KochetovNicolai | Integration test - not connected
Performance - speed up up to 4.8 times:
```
0.566 | 0.116 | -4.858x | -0.795 | 0.786 | ip_trie | 3 | \n SELECT dictGetFloat32(\'default.dict_ip_trie\', \'val\', tuple(IPv6StringToNum(ip)))\n FROM table_ip_from_dict\n WHERE ver == 6\n LIMIT 500000 FORMAT Null\n
-- | -- | -- | -- | --... | 2020-11-25T13:05:10 | 2020-11-26T05:49:41 | {} | MEMBER |
ClickHouse | ClickHouse | 734,184,823 | 17,393 | ahutsunshine | IC. Works for me. Can you help explain my second question?
```
How does the multiple columns as shards implement? For example, if I take host(string) and time(DateTime) as sharding keys, how to calculate the hash and the shard num?
``` | 2020-11-26T09:32:40 | 2020-11-26T09:32:53 | {} | NONE |
ClickHouse | ClickHouse | 734,199,534 | 1,782 | bulletRush | for ReplacingMergeTree, yandex/clickhouse-server:latest(4fee701cac0d, 4 weeks ago), this sql
```sql
SELECT xxx_id
FROM t_xxxx
FINAL
LIMIT 10, 10
FORMAT JSON
```
return without `total`:
```json
"rows": 10,
"rows_before_limit_at_least": 20,
"statistics":
{
"elapsed": 0.003157,
"rows_rea... | 2020-11-26T10:00:17 | 2020-11-26T10:00:17 | {} | NONE |
ClickHouse | ClickHouse | 733,851,540 | 17,380 | den-crane | Another issue that Distributed table tries to insert into underlying table indefinitely.
It needs some stopper. For example after some number of tries/errors block insert into Distributed for example to signal to a client.
https://github.com/ClickHouse/ClickHouse/issues/12530#issuecomment-658902526 | 2020-11-25T17:37:54 | 2020-11-25T17:37:54 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 733,881,453 | 17,236 | den-crane | >It's not the most user-friendly way
??
It was thousands requests to implement user management using SQL in CH.
Now you can consider XML user management as obsolete. | 2020-11-25T18:33:28 | 2020-11-25T18:33:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 733,919,135 | 17,385 | nikitamikhaylov | That's interesting. Maybe you have corrupted file? Will there be a correct string if you do something like `grep -rin -A 5 -B 5 'бассе» в соверх контекст песни'` ? | 2020-11-25T19:58:27 | 2020-11-25T19:58:27 | {} | MEMBER |
ClickHouse | ClickHouse | 733,906,407 | 16,400 | den-crane | @filipbacic it's another agg. function -- sortof-any, not max.
```
select argMax(x,y) from (select 1 x, 2 y union all select 2 x, 2 y)
--
1
```
y = 2 for both of x , so using any we show any(x)=1.
```
select any(x), anyLast(x) from (select null x union all select 1)
┌─any(x)─┬─anyLast(x)─┐
│ 1 │ ... | 2020-11-25T19:26:55 | 2020-11-25T19:30:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 734,001,751 | 17,415 | alexey-milovidov | Partial solution:
- add MemoryTracker::BlockerInThread to destructor of ProcessListEntry. | 2020-11-26T00:11:21 | 2020-11-26T00:11:21 | {} | MEMBER |
ClickHouse | ClickHouse | 734,003,184 | 17,386 | filimonov | That sounds really strange. I guess the reason is some silly tiny error like using the wrong path or smth like that.
Try to remove your builded binary, try to start it (should fail).
Try to make some syntax error in your cpp file and compile (should fail). | 2020-11-26T00:16:23 | 2020-11-26T00:16:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 734,008,207 | 17,410 | alexey-milovidov | What's the purpose of this test?
Looks like it checks that the per-partition node in `block_numbers` is not removed? But why we need to ensure that? | 2020-11-26T00:36:20 | 2020-11-26T00:36:20 | {} | MEMBER |
ClickHouse | ClickHouse | 734,012,232 | 17,055 | l1t1 | AMD VM , here the table a is a small table contains 49 rows
```
VM_0_52_centos :) set max_memory_usage='40G', max_bytes_before_external_group_by='20G';
SET max_memory_usage = '40G', max_bytes_before_external_group_by = '20G'
Query id: c61b725f-7d42-4358-a11b-d80f261e0a12
[VM_0_52_centos] 2020.11.26 08:44:59.... | 2020-11-26T00:52:06 | 2020-11-26T00:52:06 | {} | NONE |
ClickHouse | ClickHouse | 734,027,269 | 14,099 | alexey-milovidov | CC @azat | 2020-11-26T01:52:11 | 2020-11-26T01:52:11 | {} | MEMBER |
ClickHouse | ClickHouse | 734,063,480 | 16,664 | alexey-milovidov | And it's fixed in https://github.com/ClickHouse/ClickHouse/pull/15111 | 2020-11-26T04:15:47 | 2020-11-26T04:15:47 | {} | MEMBER |
ClickHouse | ClickHouse | 734,124,714 | 17,415 | azat | >Higher values of max_untracked_memory can lead to std::terminate for the same reason.
And by "higher" you mean? Are you talking about the case when 1024 is not enough to destroy the object? Yes, this is possible.
>Looks like you want is to disable memory checks when memory usage is incremented by small amounts (... | 2020-11-26T07:26:22 | 2020-11-26T07:26:22 | {} | MEMBER |
ClickHouse | ClickHouse | 733,814,995 | 17,143 | antisaa | I found a solution:
```sql
CREATE temporary table tmp1 AS
SELECT
number AS t,
number AS v,
'x' AS x
FROM system.numbers
WHERE number IN [1,5,7];
CREATE temporary table tmp2 AS
SELECT
number AS t,
'x' AS x
FROM system.numbers
WHERE number<20
LIMIT 20
SELECT tt.t AS t, vv.v AS v FROM tmp2 tt ASOF JO... | 2020-11-25T16:32:43 | 2020-11-25T16:32:43 | {} | NONE |
ClickHouse | ClickHouse | 733,825,911 | 16,400 | den-crane | ```sql
SELECT
max(inserted_on),
argMax(string_column, inserted_on) AS non_nullable_column,
argMax(tuple(nullable_string_column), inserted_on).1 AS nullable_column_current
FROM test.argMax_test AS amt
┌────max(inserted_on)─┬─non_nullable_column─┬─nullable_column_current─┐
│ 2020-11-25 17:47:22 │ 1 ... | 2020-11-25T16:51:59 | 2020-11-25T16:51:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 733,974,164 | 17,396 | nikitamikhaylov | Integration tests failures are completely unrelated to the subject of this PR. | 2020-11-25T22:30:20 | 2020-11-25T22:30:20 | {} | MEMBER |
ClickHouse | ClickHouse | 734,005,130 | 17,414 | filimonov | +1. I saw that problem 'in the wild' several times. It's definitely counterintuitive right now. | 2020-11-26T00:24:09 | 2020-11-26T00:24:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 734,008,649 | 17,382 | alexey-milovidov | It does not work in CI, because ClickHouse cannot be installed after this PR.
I recommend to look at the perf results of the similar PR: https://github.com/ClickHouse/ClickHouse/pull/17111 | 2020-11-26T00:37:45 | 2020-11-26T00:37:45 | {} | MEMBER |
ClickHouse | ClickHouse | 734,048,139 | 17,386 | Joeywzr | I think I've found the problem. I forget to add './' before 'clickhouse'. Silly question :(. Thanks a lot! | 2020-11-26T03:13:27 | 2020-11-26T03:19:47 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 734,170,378 | 17,404 | hexiaoting | this query runs ok: return 7 rows.
with CH version 20.13.1.1. | 2020-11-26T09:06:22 | 2020-11-26T09:06:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 733,767,547 | 17,387 | DmitriZhukov | @den-crane
sql client time is 18:04
`select now() , toString(now())`
`2020-11-25 15:04:31, 2020-11-25 16:04:31`
Server is located in UTC +2 Germany
My Computer is located in UTC+3 Moscow
So _now()_ gives correct time zone of UTC
_toString(now())_ gives incorrect time zone
| 2020-11-25T15:14:36 | 2020-11-25T15:14:36 | {} | NONE |
ClickHouse | ClickHouse | 733,866,661 | 17,346 | ralphM78 | Will PARTITION REPLACE work between a replacingmergetree table and a replicatedreplacingmergetree table? Same host, different databases.
Ralph | 2020-11-25T18:07:21 | 2020-11-25T18:07:21 | {} | NONE |
ClickHouse | ClickHouse | 733,873,340 | 17,376 | azat | @alexey-milovidov can you also remove the "pr-improvement" label? (although note sure that it is used for something, i.e. changelog generation) | 2020-11-25T18:18:41 | 2020-11-25T18:18:41 | {} | MEMBER |
ClickHouse | ClickHouse | 733,998,103 | 17,415 | alexey-milovidov | Higher values of `max_untracked_memory` can lead to std::terminate for the same reason. | 2020-11-26T00:01:37 | 2020-11-26T00:01:37 | {} | MEMBER |
ClickHouse | ClickHouse | 734,059,525 | 17,145 | alexey-milovidov | RabbitMQ test failure is also fixed. | 2020-11-26T03:58:39 | 2020-11-26T03:58:39 | {} | MEMBER |
ClickHouse | ClickHouse | 734,059,697 | 17,145 | alexey-milovidov | Maybe also add a test for getLeastSupertype?
It can be done with function `if` or array construction `toTypeName([x, y])`. | 2020-11-26T03:59:23 | 2020-11-26T03:59:23 | {} | MEMBER |
ClickHouse | ClickHouse | 733,694,229 | 17,386 | filimonov | What do you see when you recompile the project / i.e. when run `ninja`?
Check timestamp of your binary, and a way how you run it. | 2020-11-25T13:04:21 | 2020-11-25T13:04:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 733,696,153 | 17,387 | DmitriZhukov | ```
select timezone()
Europe/Berlin
```
I think it is TCP as I set connection string by server ip address
Some info
in my SQL client I do
`insert into temp (datetime) values('2020-11-24 15:22:00')`
but see what it stores
`select * from temp order by datetime`
results
`2020-11-24 14:22:00`
So it ... | 2020-11-25T13:08:02 | 2020-11-25T13:09:14 | {} | NONE |
ClickHouse | ClickHouse | 733,934,910 | 15,419 | akuzm | Now we're having performance problems again: https://clickhouse-test-reports.s3.yandex.net/15419/1a9c3135291ef8c0b468c7bb58b08d886ae6f10d/performance_comparison/report.html#changes-in-performance
But before that it was OK and even improved, and the code didn't change: https://clickhouse-test-reports.s3.yandex.net/15... | 2020-11-25T20:37:30 | 2020-11-25T20:37:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 733,974,468 | 17,394 | nikitamikhaylov | Integration tests failures are completely unrelated to the subject of this PR. | 2020-11-25T22:31:26 | 2020-11-25T22:31:26 | {} | MEMBER |
ClickHouse | ClickHouse | 733,976,140 | 14,414 | nikitamikhaylov | Leak in openssl: https://gist.github.com/nikitamikhaylov/3ddede6efb865ff25f905f4d11c9d232 | 2020-11-25T22:37:41 | 2020-11-25T22:37:41 | {} | MEMBER |
ClickHouse | ClickHouse | 734,039,420 | 17,055 | l1t1 | i increase the cpu of centos vbox from 1 to 12, nothing changed
| 2020-11-26T02:39:09 | 2020-11-26T02:39:09 | {} | NONE |
ClickHouse | ClickHouse | 734,165,881 | 16,000 | decaseal | Decided to refuse to add these functions | 2020-11-26T08:57:47 | 2020-11-26T08:57:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 733,695,289 | 17,379 | filimonov | Remove CMakeCache, or... just "install clang from brew" how the message says. | 2020-11-25T13:06:20 | 2020-11-25T13:06:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 733,814,279 | 17,386 | Joeywzr | @filimonov Thanks. Here it is.
~~~
[wzr@sugon205 build]$ ninja
[0/2] Re-checking globbed directories...
[14/14] cd /home/wzr/ch113/ClickH..._symlink clickhouse clickhouse-server
~~~
The whole process is that I modify a code file, then I recompile it using `ninja`, and then generate a new clickhouse file in `build... | 2020-11-25T16:31:22 | 2020-11-25T16:31:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 733,883,117 | 17,255 | azat | >By the way, both materialized_view_parallel_insert and parallel_inserts tests are now broken (take too long).
Indeed, the reason is that SETTINGS was not applied for INSERT query there, here is an RFC patch that addresses this issue - #17414 | 2020-11-25T18:36:39 | 2020-11-25T18:36:55 | {} | MEMBER |
ClickHouse | ClickHouse | 733,922,205 | 16,947 | azat | >Conflicting files
>src/Interpreters/InterpreterSelectWithUnionQuery.h
Going to resolve them after the review (to avoid doing it constantly) | 2020-11-25T20:05:16 | 2020-11-25T20:05:16 | {} | MEMBER |
ClickHouse | ClickHouse | 733,735,217 | 17,387 | den-crane | JDBC client (more exactly JAVA itself) converts DateTime to TZ of JVM (to TZ of your Desktop).
check `select now() , toString(now())`
https://github.com/ClickHouse/clickhouse-jdbc/issues/202 | 2020-11-25T14:19:50 | 2020-11-25T14:20:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 733,755,966 | 16,621 | filimonov | > @filimonov could you give your opinion, please?
My $0.02:
2 - promising, but time consuming / hard to predict all the corner cases etc. Do you have a time for that?
3 - easier upgrade path, but i doubt we will be able to make upgrade 100% smooth anyway
1 - is acceptable (that increment is definitely importan... | 2020-11-25T14:55:39 | 2020-11-25T14:55:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 733,756,828 | 17,376 | alexey-milovidov | Ok. @azat Also need to cleanup perf test. | 2020-11-25T14:57:01 | 2020-11-25T14:57:01 | {} | MEMBER |
ClickHouse | ClickHouse | 733,827,932 | 17,387 | den-crane | Вы как / чем запросы на от себя выполняете? | 2020-11-25T16:55:19 | 2020-11-25T16:55:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 733,830,761 | 16,400 | den-crane | the same with other agg functions
```
select groupArray(x) from (select 1 as x union all select null)
┌─groupArray(x)─┐
│ [1] │
└───────────────┘
select groupArray(tuple(x)) from (select 1 as x union all select null)
┌─groupArray(tuple(x))─┐
│ [(1),(NULL)] │
└──────────────────────┘
se... | 2020-11-25T17:00:38 | 2020-11-25T17:02:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 734,003,735 | 17,346 | filimonov | If the schema (set of columns / keys / partitioning etc) is the same - then the answer is yes, definitely. You can move the partitions between Replicated and non-Replicated tables. | 2020-11-26T00:18:45 | 2020-11-26T00:19:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 734,014,506 | 17,385 | alexey-milovidov | It may happen if you mixed up the CREATE TABLE statements for `hits_v1` with `hits_100m` (they are similar but different tables). | 2020-11-26T01:01:21 | 2020-11-26T01:01:21 | {} | MEMBER |
ClickHouse | ClickHouse | 734,026,681 | 14,099 | alexey-milovidov | Please share `perf report`s. | 2020-11-26T01:49:54 | 2020-11-26T01:49:54 | {} | MEMBER |
ClickHouse | ClickHouse | 734,027,382 | 17,055 | l1t1 | because the physical memory on win10 is 16GB, and free of Ubuntu 18.04 is 12GB, I `set max_memory_usage='11G', max_bytes_before_external_group_by='10G'` | 2020-11-26T01:52:47 | 2020-11-26T01:52:47 | {} | NONE |
ClickHouse | ClickHouse | 734,085,865 | 17,393 | ahutsunshine | Anyone can help this? | 2020-11-26T05:37:44 | 2020-11-26T05:37:44 | {} | NONE |
ClickHouse | ClickHouse | 733,747,584 | 9,287 | baibaichen | @alexey-milovidov, the following codes copied from `movePartitionToTable`, why isn't atomic? Because
1. Data is visible at the destination table and at the same time is invisible at the source table.
2. In case of failure, the source data is unchanged.
3. Source data is deleted after transaction commit.
```
... | 2020-11-25T14:41:31 | 2020-11-25T14:41:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 733,758,905 | 17,375 | alexey-milovidov | https://github.com/ClickHouse/ClickHouse/issues/17404 | 2020-11-25T15:00:24 | 2020-11-25T15:00:24 | {} | MEMBER |
ClickHouse | ClickHouse | 733,783,377 | 17,387 | den-crane | ```
select timezone() -> Europe/Berlin
datetime DateTime
insert into temp (datetime) values('2020-11-25 00:10:00')
```
При инсерте КХ воспринимает ('2020-11-25 00:10:00') как время в Europe/Berlin, конвертирует в unixtime в UTC и сохраняет в таблице.
```
select * from temp where
datetime between '2020-... | 2020-11-25T15:40:53 | 2020-11-25T15:44:01 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 733,797,013 | 17,403 | den-crane | Please show clickhouse-server logs for this slow query.
Try to ping **-s 1500** | 2020-11-25T16:01:41 | 2020-11-25T16:02:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 733,886,004 | 17,402 | nikitamikhaylov | Integration tests failures are completely unrelated to the subject of this PR. | 2020-11-25T18:42:24 | 2020-11-25T18:42:24 | {} | MEMBER |
ClickHouse | ClickHouse | 734,030,576 | 17,055 | l1t1 | and the result of vitualbox centos7 on win7 64bits, also better than AMD, but worse than Win10 WSL2
![2020-11-26_100127](https://user-images.githubusercontent.com/23301704/100299519-e1067f80-2fc... | 2020-11-26T02:05:23 | 2020-11-26T02:05:23 | {} | NONE |
ClickHouse | ClickHouse | 734,055,771 | 17,347 | alexey-milovidov | Maybe add a test for empty substring? | 2020-11-26T03:42:47 | 2020-11-26T03:42:47 | {} | MEMBER |
ClickHouse | ClickHouse | 734,059,380 | 17,145 | alexey-milovidov | > 01247_optimize_distributed_group_by_sharding_key
The fix is merged in https://github.com/ClickHouse/ClickHouse/pull/17254 | 2020-11-26T03:57:58 | 2020-11-26T03:57:58 | {} | MEMBER |
ClickHouse | ClickHouse | 734,061,801 | 17,416 | alexey-milovidov | > docker: Error response from daemon: toomanyrequests: You have reached your pull rate limit. You may increase the limit by authenticating and upgrading: https://www.docker.com/increase-rate-limit.
| 2020-11-26T04:08:45 | 2020-11-26T04:08:45 | {} | MEMBER |
ClickHouse | ClickHouse | 734,090,284 | 17,393 | alexey-milovidov | DateTime is processed as UInt32 containing the unix timestamp.
Date is processed as UInt16 containing the number of days since unix epoch.
Float32 is processed as the corresponding 4 bytes in memory. Float64 as 8 bytes. | 2020-11-26T05:52:57 | 2020-11-26T05:52:57 | {} | MEMBER |
ClickHouse | ClickHouse | 734,132,488 | 17,415 | alexey-milovidov | > And by "higher" you mean? Are you talking about the case when 1024 is not enough to destroy the object? Yes, this is possible.
No. If max untracked memory is 1000000 and the current value of untracked memory is 999999. | 2020-11-26T07:45:52 | 2020-11-26T07:45:52 | {} | MEMBER |
ClickHouse | ClickHouse | 733,777,384 | 17,387 | den-crane | I suggest to configure at your desktop
use_time_zone=Europe/Moscow
use_server_time_zone_for_dates=false
use_server_time_zone=false
When JDBC will not convert time. And you will see DateTime values in Berlin TZ
| 2020-11-25T15:30:47 | 2020-11-25T15:31:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 733,878,125 | 16,400 | den-crane | > @den-crane Thank you, definitely more elegant and accurate than my solution.
>
> Will this be fixed in the future so that there is no need for those kinds of workarounds?
It's nothing to fix. It's a natural behaviour for agg. functions against NULLs.
```
select count(x) from (select 1 as x union all select ... | 2020-11-25T18:27:31 | 2020-11-25T18:27:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 733,998,858 | 17,415 | alexey-milovidov | Looks like you want is to disable memory checks when memory usage is incremented by small amounts (this is not the same as large max_untracked_memory, because max_untracked_memory is accumulated).
But it will prevent correct memory tracking e.g. when using AggregateFunction states that do small allocations on heap (mi... | 2020-11-26T00:04:16 | 2020-11-26T00:04:57 | {} | MEMBER |
ClickHouse | ClickHouse | 734,090,681 | 17,393 | alexey-milovidov | Example:
```
milovidov-desktop :) SELECT hex(reinterpretAsString(1.0))
SELECT hex(reinterpretAsString(1.))
Query id: 2cebc77b-aa1f-4a2a-bd48-144fee820f8c
┌─hex(reinterpretAsString(1.))─┐
│ 000000000000F03F │
└──────────────────────────────┘
1 rows in set. Elapsed: 0.002 sec.
milovidov-de... | 2020-11-26T05:54:14 | 2020-11-26T05:54:14 | {} | MEMBER |
ClickHouse | ClickHouse | 734,131,057 | 15,111 | vitlibar | > ```
> $ ./clickhouse-grpc-client.py
> 'list' object has no attribute 'join'
> ```
There was a small mistake in error-handling in this python script, and you have to install grpcio-tools: `pip3 install grpcio-tools`. Fixed in https://github.com/ClickHouse/ClickHouse/pull/17435/commits/dbd136672f074404530fabd917f... | 2020-11-26T07:42:23 | 2020-11-26T07:42:35 | {} | MEMBER |
ClickHouse | ClickHouse | 734,143,248 | 17,436 | ggchangan | Other Information:
It only takes about 2 seconds when I select all data from fwmrm_oltp.lu_dma, why when in join clause, it takes about 10s to `SELECT(country_id, code, description) ON fwmrm_oltp.lu_dma` | 2020-11-26T08:10:53 | 2020-11-26T08:10:53 | {} | NONE |
ClickHouse | ClickHouse | 733,716,330 | 16,400 | filipbacic | Hi,
I have the same problem with argMax function, it does not take NULL values into account:
Reproducible code:
```
CREATE TABLE test.argMax_test (
inserted_on DateTime DEFAULT now(),
string_column String,
nullable_string_column Nullable(String)
) ENGINE = Log;
INSERT INTO test.argMax_test (string_column... | 2020-11-25T13:47:25 | 2020-11-25T13:47:25 | {} | NONE |
ClickHouse | ClickHouse | 733,728,345 | 17,271 | alesapin | Yes, failures not related. | 2020-11-25T14:08:30 | 2020-11-25T14:08:30 | {} | MEMBER |
ClickHouse | ClickHouse | 733,788,427 | 17,403 | den-crane | It could be anything. Network issue. Slow query execution. System swapping.
Why do you think that the issue is in http ?
How we can reproduce this issue? | 2020-11-25T15:48:31 | 2020-11-25T15:48:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 733,867,229 | 16,400 | filipbacic | @den-crane Thank you, definitely more elegant and accurate than my solution.
Will this be fixed in the future so that there is no need for those kinds of workarounds? | 2020-11-25T18:08:21 | 2020-11-25T18:08:21 | {} | NONE |
ClickHouse | ClickHouse | 734,087,353 | 15,111 | alexey-milovidov | ```
$ ./clickhouse-grpc-client.py
'list' object has no attribute 'join'
Bye
```
```
<!-- gRPC protocol (see src/Server/grpc_protos/clickhouse_grpc.proto for the API) -->
<grpc_port>9001</grpc_port>
<grpc>
<enable_ssl>false</enable_ssl>
<!-- The following two files are... | 2020-11-26T05:42:52 | 2020-11-26T05:42:52 | {} | MEMBER |
ClickHouse | ClickHouse | 734,125,441 | 17,410 | alesapin | > What's the purpose of this test?
> Looks like it checks that the per-partition node in `block_numbers` is not removed? But why we need to ensure that?
It checks that if we have a partition without any parts mutation for such partition finishes successfully. | 2020-11-26T07:28:18 | 2020-11-26T07:28:18 | {} | MEMBER |
ClickHouse | ClickHouse | 734,156,652 | 16,877 | alesapin | For some reason, epoll version works slower... investigating | 2020-11-26T08:39:42 | 2020-11-26T08:39:42 | {} | MEMBER |
ClickHouse | ClickHouse | 733,793,319 | 17,403 | audricbizouarn | if I ping the server at the same time, there is no problem.
And if I query clickhouse in tcp mode I don't have a problem either.
I don't see swap with htop.
I sometimes have no process running (SHOW PROCESSLIST) and yet I have the problem | 2020-11-25T15:56:03 | 2020-11-25T15:56:03 | {} | NONE |
ClickHouse | ClickHouse | 733,856,459 | 17,236 | AndrewNik | @xinstein Nope, I had to create user by sql commands. It's not the most user-friendly way, but it is only thing I've found | 2020-11-25T17:47:51 | 2020-11-25T17:47:51 | {} | NONE |
ClickHouse | ClickHouse | 734,109,043 | 17,380 | filimonov | > Workaround: set insert_distributed_sync = 1.
Actually async in desired in many cases. For example it can do proper retries into Replicated. It properly deals with partial failure. It easily survive that some server is temporary unavailiable or we loose the connection during sending the data. | 2020-11-26T06:48:02 | 2020-11-26T06:48:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 733,775,599 | 17,387 | den-crane | hm,
>sql client time is 18:04
>So now() gives correct time zone of UTC
No it's incorrect. You should get 18:04
Check your JDBC settings.
```
use_time_zone
use_server_time_zone_for_dates
use_server_time_zone
```
Another question what do you want to achieve? Because now you observe a normal behavi... | 2020-11-25T15:27:53 | 2020-11-25T15:27:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 733,836,890 | 17,387 | den-crane | Потому что переводится из Берлин в Москва. Видимо JVM старая и не знает про отмену перевода на летнее / зимнее.
Поставьте
на клиенте поставить
```
use_time_zone=Europe/Moscow
use_server_time_zone_for_dates=false
use_server_time_zone=false
```
и будет конвертировать Europe/Moscow -> UTC ->Europe/Moscow и буде... | 2020-11-25T17:11:46 | 2020-11-25T17:11:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 733,858,189 | 9,621 | den-crane | It seems Curtiz will never remove it. I think it pretty important task. | 2020-11-25T17:51:13 | 2020-11-25T17:51:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 733,901,420 | 16,400 | filipbacic | > > @den-crane Thank you, definitely more elegant and accurate than my solution.
> > Will this be fixed in the future so that there is no need for those kinds of workarounds?
>
> It's nothing to fix. It's a natural behaviour for agg. functions against NULLs.
>
> ```
> select count(x) from (select 1 as x union a... | 2020-11-25T19:15:11 | 2020-11-25T19:15:11 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 734,063,371 | 16,664 | alexey-milovidov | > Intergration tests failed for unknown reason.
Because ClickHouse cannot be installed due to
> Compatibility check — glibc check failed | 2020-11-26T04:15:21 | 2020-11-26T04:15:21 | {} | MEMBER |
ClickHouse | ClickHouse | 734,141,928 | 17,379 | hoverzheng | Hi @filimonov
I deleted CMakeCache.txt, and ran **ninjia**，but got the same error：
$ ninja
[0/2] Re-checking globbed directories...
[1/2] Re-running CMake...
-- The C compiler identification is AppleClang 10.0.1.10010046
-- The CXX compiler identification is AppleClang 10.0.1.10010046
-- Detecting C compile... | 2020-11-26T08:07:55 | 2020-11-26T08:08:14 | {} | NONE |
ClickHouse | ClickHouse | 734,013,359 | 17,380 | alexey-milovidov | Workaround: set insert_distributed_sync = 1. | 2020-11-26T00:56:52 | 2020-11-26T00:56:52 | {} | MEMBER |
ClickHouse | ClickHouse | 734,026,924 | 14,099 | alexey-milovidov | Maybe it because localhost optimization will be disabled? | 2020-11-26T01:50:48 | 2020-11-26T01:50:48 | {} | MEMBER |
ClickHouse | ClickHouse | 734,118,642 | 17,376 | azat | >Performance — 6 errors, 4 too long, 8 faster, 7 slower, 40 unstable
Failures looks unrelated (after date_time_short has been updated) | 2020-11-26T07:11:10 | 2020-11-26T07:11:10 | {} | MEMBER |
ClickHouse | ClickHouse | 733,731,684 | 16,899 | alesapin | Ok, now I know more https://stackoverflow.com/a/3527632 | 2020-11-25T14:14:07 | 2020-11-25T14:14:07 | {} | MEMBER |
ClickHouse | ClickHouse | 733,913,468 | 16,400 | filipbacic | @den-crane I would say that any() comes after max, first max(y) is performed (in argMax(x,y)) and then for all records which have y = max(y) any(x) is performed. So if there is just one record with y = max(y) then whatever there is in x it should be returned (in other words any(x) = x when there is just one record). | 2020-11-25T19:43:59 | 2020-11-25T19:43:59 | {} | NONE |
ClickHouse | ClickHouse | 733,924,111 | 16,767 | azat | Rebased against upstream/master to fix some CI checks (previous HEAD was 2f4ae45254126cc09c84b4725413752a13598e7d) | 2020-11-25T20:10:04 | 2020-11-25T20:10:04 | {} | MEMBER |
ClickHouse | ClickHouse | 734,015,447 | 17,055 | filimonov | And the same on Windows 10 wsl? | 2020-11-26T01:04:36 | 2020-11-26T01:05:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 734,026,910 | 17,055 | l1t1 | Windows 10 wsl2 Ubuntu 18.04
```
DESKTOP-RS3EG9A.localdomain :) SET send_logs_level='trace';
SET send_logs_level = 'trace'
Query id: 6fb839c0-3045-40dd-8427-366bb7b82edb
Ok.
0 rows in set. Elapsed: 0.001 sec.
DESKTOP-RS3EG9A.localdomain :) set max_memory_usage='11G', max_bytes_before_external_group_by... | 2020-11-26T01:50:44 | 2020-11-26T01:50:44 | {} | NONE |
ClickHouse | ClickHouse | 733,919,591 | 17,349 | nikitamikhaylov | `git submodule sync && git submodule update --init --recursive && git submodule foreach git reset --hard && git submodule foreach git clean -xfd` | 2020-11-25T19:59:42 | 2020-11-25T19:59:42 | {} | MEMBER |
ClickHouse | ClickHouse | 733,974,279 | 17,395 | nikitamikhaylov | Integration tests failures are completely unrelated to the subject of this PR. | 2020-11-25T22:30:38 | 2020-11-25T22:30:38 | {} | MEMBER |
ClickHouse | ClickHouse | 734,007,494 | 17,243 | l1t1 | I use https://mirrors.tuna.tsinghua.edu.cn/centos/7/isos/x86_64/CentOS-7-x86_64-Minimal-2009.iso and run clickhouse on it
| 2020-11-26T00:33:30 | 2020-11-26T00:33:30 | {
"hooray": 1
} | NONE |
ClickHouse | ClickHouse | 734,133,454 | 17,415 | alexey-milovidov | > But yes it is possible to get inaccurate memory accounting.
We can implement another mode - when memory is accounted but exception is not thrown. | 2020-11-26T07:48:03 | 2020-11-26T07:48:03 | {} | MEMBER |
ClickHouse | ClickHouse | 734,198,960 | 17,296 | KochetovNicolai | Nice fix!
I have added some more tests in #17439 | 2020-11-26T09:59:09 | 2020-11-26T09:59:09 | {} | MEMBER |
ClickHouse | ClickHouse | 735,466,464 | 17,585 | azat | >But seems that rewrite can be fixed to take this into account
But this may conflicts with #16947, so I would wait sometime maybe someone will take a look at it | 2020-11-29T22:15:09 | 2020-11-29T22:15:09 | {} | MEMBER |
ClickHouse | ClickHouse | 735,468,657 | 14,099 | filimonov | Here are some details.
https://altinity.com/blog/2018/5/10/circular-replication-cluster-topology-in-clickhouse
BTW: quite a lot of people are interested in that, but the concept described in that article works badly...
| 2020-11-29T22:34:33 | 2020-11-29T22:34:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 735,599,307 | 17,052 | MrSsunlight | > Check this:
> https://den-crane.github.io/Everything_you_should_know_about_materialized_views_commented.pdf
> https://youtu.be/ckChUkC3Pns?list=PLO3lfQbpDVI-hyw4MyqxEk3rDHw95SzxJ&t=9371
>
> Without a reproduction it's hard to say what is wrong with your setup.
Thanks i study | 2020-11-30T07:13:50 | 2020-11-30T07:13:50 | {} | NONE |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.