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 | 655,240,975 | 12,163 | alexey-milovidov | ```
clickhouse-benchmark <<< "SELECT uniqState(x) FROM (SELECT materialize(1) AS k, toNullable(1) AS x FROM numbers(1)) GROUP BY k WITH TOTALS"
``` | 2020-07-08T02:21:02 | 2020-07-08T02:21:02 | {} | MEMBER |
ClickHouse | ClickHouse | 655,241,631 | 12,163 | alexey-milovidov | ```
SELECT uniqState(x) FROM (SELECT toNullable(1) AS x) WITH TOTALS
``` | 2020-07-08T02:23:05 | 2020-07-08T02:23:05 | {} | MEMBER |
ClickHouse | ClickHouse | 654,668,497 | 12,127 | thyn | As I understand AggregateFunction is working as I need. Am I right?
`
CREATE TABLE temp.test11 (
d Date,
f4 AggregateFunction (medianExact, UInt32)
)
ENGINE = AggregatingMergeTree()
PARTITION BY tuple()
ORDER BY d
```
```
insert into temp.test11 (d,f4) select toDate('2020-01-01'), medianExa... | 2020-07-07T07:53:58 | 2020-07-07T07:55:19 | {} | NONE |
ClickHouse | ClickHouse | 654,453,942 | 12,127 | thyn | > > anyLastIf in select is not working because it produces default value which rewrites current value.
>
> Hm... With nulls - too?
Thank you for idea. Didn't use NULL values because didn't see any reason for that.
```
CREATE TABLE temp.test5 (
d Date,
f1 SimpleAggregateFunction(anyLast, Nullabl... | 2020-07-06T20:39:46 | 2020-07-07T08:34:01 | {} | NONE |
ClickHouse | ClickHouse | 649,541,164 | 9,336 | bohutang | Comments:
Add `select @@version` #11815 | 2020-06-25T13:28:20 | 2020-07-07T09:19:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 654,759,342 | 3,767 | filimonov | Fixed in 19.8, most probably by #5238 | 2020-07-07T10:29:42 | 2020-07-07T10:29:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,004,450 | 12,195 | qoega | I added a can be tested label to see difference in perf tests. In current version it seems invalid. There could be more complex logic when we can set less threads for trivial insert queries. | 2020-07-07T17:13:42 | 2020-07-07T17:13:42 | {} | MEMBER |
ClickHouse | ClickHouse | 655,217,451 | 12,257 | alexey-milovidov | @filimonov If it's already fixed, why we need this task? | 2020-07-08T00:58:49 | 2020-07-08T00:58:49 | {} | MEMBER |
ClickHouse | ClickHouse | 655,257,869 | 7,529 | cailin-lai | Some new updates, after upgrading version to v20.1.6, looks the issue has gone. We've confirmed this in many add column stages.
Thanks to Team! | 2020-07-08T03:18:38 | 2020-07-08T03:18:38 | {} | NONE |
ClickHouse | ClickHouse | 655,261,693 | 12,275 | alexey-milovidov | @alesapin Tests did not run in 5 hours. | 2020-07-08T03:33:09 | 2020-07-08T03:33:09 | {} | MEMBER |
ClickHouse | ClickHouse | 655,349,350 | 12,257 | filimonov | Maybe we don't need that.
Do we know why the stack trace was missing in that case? Is it fixed for the general case? Can the problem pop up again in some other place / with other exceptions? | 2020-07-08T07:44:54 | 2020-07-08T07:44:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 654,508,856 | 12,091 | alexey-milovidov | Does not help. | 2020-07-06T23:13:25 | 2020-07-06T23:13:25 | {} | MEMBER |
ClickHouse | ClickHouse | 655,183,177 | 12,076 | alexey-milovidov | Did the experiment go well? | 2020-07-07T23:04:25 | 2020-07-07T23:04:25 | {} | MEMBER |
ClickHouse | ClickHouse | 655,239,313 | 12,163 | alexey-milovidov | ```
clickhouse-benchmark <<< "SELECT k, finalizeAggregation(uniqState(x)) FROM (SELECT intDiv(number, 3) AS k, toNullable(1) AS x FROM system.numbers LIMIT 100000) GROUP BY k WITH TOTALS"
``` | 2020-07-08T02:15:22 | 2020-07-08T02:15:22 | {} | MEMBER |
ClickHouse | ClickHouse | 655,239,917 | 12,163 | alexey-milovidov | ```
$ clickhouse-benchmark <<< "SELECT uniqState(x) FROM (SELECT intDiv(number, 3) AS k, toNullable(1) AS x FROM system.numbers LIMIT 100000) GROUP BY k WITH TOTALS"
Loaded 1 queries.
An error occurred while processing the query 'SELECT uniqState(x) FROM (SELECT intDiv(number, 3) AS k, toNullable(1) AS x FROM system... | 2020-07-08T02:17:22 | 2020-07-08T02:17:30 | {} | MEMBER |
ClickHouse | ClickHouse | 655,240,528 | 12,111 | alexey-milovidov | Please link all the issues found by this fuzzer to let people know how good this tool is. | 2020-07-08T02:19:24 | 2020-07-08T02:19:24 | {} | MEMBER |
ClickHouse | ClickHouse | 655,354,422 | 12,269 | filimonov | Relates to https://github.com/ClickHouse/ClickHouse/issues/12098 ? | 2020-07-08T07:55:03 | 2020-07-08T07:55:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,396,349 | 12,261 | filimonov | 1) Why ReplacingMergeTree not AggregatingMergeTree?
2) How do you examine the size? Do you do optimize before, how many states (i.e. different ds,ext1 combinations) do your table store?
```
clickhouse-local -q 'select groupBitmapState(number) as state from numbers(50000000)' | clickhouse-local --structure 'st... | 2020-07-08T09:13:37 | 2020-07-08T09:18:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 654,806,164 | 12,195 | ucasfl | > Can you check speed of insert if numbers_mt is used?
There is no difference between numbers and numbers_mt @qoega | 2020-07-07T11:58:30 | 2020-07-07T11:58:30 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 654,932,737 | 12,160 | abyss7 | Looks great! Have you tested it manually? | 2020-07-07T15:13:12 | 2020-07-07T15:13:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,232,086 | 7,380 | alexey-milovidov | #12263 | 2020-07-08T01:50:15 | 2020-07-08T01:50:15 | {} | MEMBER |
ClickHouse | ClickHouse | 655,240,813 | 12,163 | alexey-milovidov | ```
clickhouse-benchmark <<< "SELECT uniqState(x) FROM (SELECT materialize(1) AS k, toNullable(1) AS x FROM numbers(10)) GROUP BY k WITH TOTALS"
``` | 2020-07-08T02:20:27 | 2020-07-08T02:20:27 | {} | MEMBER |
ClickHouse | ClickHouse | 655,298,967 | 12,279 | alexey-milovidov | @alesapin Tests did not run in 4 hours. | 2020-07-08T05:41:39 | 2020-07-08T05:41:39 | {} | MEMBER |
ClickHouse | ClickHouse | 654,666,949 | 10,142 | filimonov | In newer versions the exception message get is worse:
```
SELECT ifNull(CAST(NULL, 'Nullable(Decimal(38, 10))'), 0)
Received exception from server (version 20.4.6):
Code: 43. DB::Exception: Received from localhost:9000. DB::Exception: Invalid types of arguments 2 and 3 of if.
0 rows in set. Elapsed: 0.004 s... | 2020-07-07T07:51:03 | 2020-07-07T07:51:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 654,672,094 | 7,714 | filimonov | Fixed in https://github.com/ClickHouse/ClickHouse/pull/9612 | 2020-07-07T08:00:29 | 2020-07-07T08:00:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 654,951,713 | 12,168 | alesapin | Ok, let's merge | 2020-07-07T15:43:22 | 2020-07-07T15:43:22 | {} | MEMBER |
ClickHouse | ClickHouse | 655,054,891 | 12,196 | Akazz | From what I can see, performance tests are OK! | 2020-07-07T18:50:25 | 2020-07-07T18:50:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,241,770 | 12,163 | alexey-milovidov | ```
SELECT uniqState(toNullable(1)) WITH TOTALS
``` | 2020-07-08T02:23:31 | 2020-07-08T02:23:31 | {} | MEMBER |
ClickHouse | ClickHouse | 654,995,451 | 12,195 | den-crane | Probably https://github.com/ClickHouse/ClickHouse/issues/5907 (set max_threads=1) has sense only if max_insert_threads=1 | 2020-07-07T16:58:20 | 2020-07-07T16:58:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,173,153 | 12,275 | alexey-milovidov | No need to implement for "most subtype" because it is used only in `arrayIntersect` and it works already due to "default implementation for low cardinality". | 2020-07-07T22:39:21 | 2020-07-07T22:39:21 | {} | MEMBER |
ClickHouse | ClickHouse | 655,346,598 | 12,279 | alesapin | > @alesapin Tests did not run in 4 hours.
Problems in CI, fixing... | 2020-07-08T07:39:06 | 2020-07-08T07:39:06 | {} | MEMBER |
ClickHouse | ClickHouse | 654,582,031 | 11,564 | bobrik | I made a patch based on my previous comment. It's geared towards queries for continuous PK ranges, and it's too ugly to show in public, but it works for my use case.
```
Found regular ranges for part 20200707_52684_52918_6 with 1775 steps (2296 total marks in the part)
Regular ranges for part 20200707_52684_52918_... | 2020-07-07T03:48:48 | 2020-07-07T03:48:48 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 654,899,448 | 12,177 | alexey-milovidov | Ok.
But TCPHandler doesn't have session ids. | 2020-07-07T14:23:45 | 2020-07-07T14:52:48 | {} | MEMBER |
ClickHouse | ClickHouse | 655,010,206 | 11,572 | laingawbl | Looks like I get this on 20.5.2 as well:
```sql
:) CREATE TABLE foo
(
timestamp DateTime CODEC(DoubleDelta, LZ4),
val UInt64
)
ENGINE = SummingMergeTree()
PARTITION BY toYYYYMM(timestamp)
ORDER BY (timestamp)
:) INSERT INTO foo VALUES ('2020-01-01 00:00:00', 1)
:) INSERT INTO foo VALUES ('20... | 2020-07-07T17:23:44 | 2020-07-07T17:23:44 | {} | NONE |
ClickHouse | ClickHouse | 655,063,558 | 8,212 | alexey-milovidov | `SELECT 'hello' UNION ALL SELECT toLowCardinality('hello')` | 2020-07-07T19:06:26 | 2020-07-07T19:06:26 | {} | MEMBER |
ClickHouse | ClickHouse | 655,136,445 | 12,179 | azat | >std::optional<Block> does not make sense, because it implies that passing empty block is somewhat acceptable.
But injectiveness of zero-argument function is degenerate case.
@alexey-milovidov thoughts? create proper arguments just for checking injectiveness?
Right now the only place that requires arguments for ... | 2020-07-07T21:10:07 | 2020-07-07T21:10:07 | {} | MEMBER |
ClickHouse | ClickHouse | 655,235,508 | 12,163 | alexey-milovidov | @akuzm Does not reproduce. | 2020-07-08T02:02:09 | 2020-07-08T02:02:09 | {} | MEMBER |
ClickHouse | ClickHouse | 655,237,801 | 12,163 | alexey-milovidov | ```
clickhouse-benchmark <<< "SELECT k, finalizeAggregation(quantilesTimingState(0.5)(x)) FROM (WITH toNullable(number = 3 ? 3 : 1) AS d SELECT intDiv(number, 3) AS k, number % d AS x FROM system.numbers LIMIT 100000) GROUP BY k WITH TOTALS"
``` | 2020-07-08T02:10:05 | 2020-07-08T02:10:05 | {} | MEMBER |
ClickHouse | ClickHouse | 655,250,630 | 12,163 | alexey-milovidov | @akuzm It is broken here: https://github.com/ClickHouse/ClickHouse/pull/10890
Versions 20.5 and 20.4 are affected. | 2020-07-08T02:53:30 | 2020-07-08T02:53:30 | {} | MEMBER |
ClickHouse | ClickHouse | 655,301,859 | 12,264 | alexey-milovidov | Minimal example:
`SELECT a, sum(b) FROM (SELECT 1 AS a, 1 AS b, 0 AS c) GROUP BY a HAVING c` | 2020-07-08T05:50:41 | 2020-07-08T05:50:41 | {} | MEMBER |
ClickHouse | ClickHouse | 655,334,174 | 12,284 | blinkov | @healiseu first of all, you are using almost one-year-old version, while looking at documentation intended for `master`. Unfortunately, the documentation is still lagging behind code, so the best place to look for examples is the `tests` folder.
So the format you are looking for is probably `ArrowStream` https://git... | 2020-07-08T07:13:00 | 2020-07-08T07:13:00 | {
"+1": 5
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 654,718,625 | 9,336 | bohutang | Comments:
Add `KILL QUERY [connection_id]` #12152 | 2020-07-07T09:19:27 | 2020-07-07T09:19:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 654,997,970 | 11,971 | akuzm | omg, the memory profiling now takes 25% of query no.2 in questdb_sum_int32. These results won't make sense, I've reverted the profiling patch, will rerun the tests. | 2020-07-07T17:02:22 | 2020-07-07T17:02:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,131,264 | 11,572 | alexey-milovidov | It should not give "Output of TreeExecutor is not sorted." exception nevertheless. | 2020-07-07T20:59:30 | 2020-07-07T20:59:30 | {} | MEMBER |
ClickHouse | ClickHouse | 654,646,816 | 12,185 | blinkov | @shinoi2 isn't it similar to <https://clickhouse.tech/docs/en/sql-reference/operators/#operator-interval>? | 2020-07-07T07:11:12 | 2020-07-07T07:11:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 654,807,374 | 12,177 | myrrc | Proposed changes:
Currently in `Interpreters/ProcessList.h` we have the hierarchy
```
ProcessList =>
(Username => RAM usage info);
(Username => ProcessListForUser)
```
the latter responsible for tracking user memory (and limiting it by the `max_memory_usage_for_user` setting).
As we introduce another ... | 2020-07-07T12:01:10 | 2020-07-07T12:01:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,132,595 | 12,182 | azat | @alexey-milovidov I decided not to expose per-user tracking outside, and just limit amount of each tracker to the total_memory_tracker (which is periodically synced with process resident size), thoughts? | 2020-07-07T21:02:11 | 2020-07-07T21:02:11 | {} | MEMBER |
ClickHouse | ClickHouse | 655,240,317 | 12,163 | alexey-milovidov | ```
clickhouse-benchmark <<< "SELECT uniqState(x) FROM (SELECT intDiv(number, 3) AS k, toNullable(1) AS x FROM numbers(10)) GROUP BY k WITH TOTALS"
``` | 2020-07-08T02:18:39 | 2020-07-08T02:18:39 | {} | MEMBER |
ClickHouse | ClickHouse | 655,369,868 | 12,269 | inakivb | I think it doesn't as there is no update expression, it's not mysql and there isn't error output | 2020-07-08T08:24:19 | 2020-07-08T08:24:19 | {} | NONE |
ClickHouse | ClickHouse | 655,401,411 | 12,261 | filimonov | Please show how you get that result (exact queries you run), i.e. where that 190 Mb number cames from?
And you didn't answer other questions above. | 2020-07-08T09:22:47 | 2020-07-08T09:24:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 654,506,335 | 12,179 | alexey-milovidov | Sorry, now it looks even more strange.
`std::optional<Block>` does not make sense, because it implies that passing empty block is somewhat acceptable.
But injectiveness of zero-argument function is degenerate case. | 2020-07-06T23:04:22 | 2020-07-06T23:04:22 | {} | MEMBER |
ClickHouse | ClickHouse | 654,751,590 | 12,185 | filimonov | Is it what you are looking for?
```
WITH
toDate('2018-01-01') AS date,
toDateTime('2018-01-01 00:00:00') AS date_time
SELECT
date + INTERVAL 1 YEAR AS add_years_with_date,
date_time + INTERVAL 1 YEAR AS add_years_with_date_time
┌─add_years_with_date─┬─add_years_with_date_time─┐
│ 2... | 2020-07-07T10:15:18 | 2020-07-07T10:15:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 654,993,251 | 11,971 | akuzm | 40% expected variability for query no. 0 from questdb_sum_32 -- rather bad, but why? Let's open 'output.7z' and look at the variability of metrics for this query (`questdb_sum_32_0_SELECT...right.metrics.rep`). The second column is array [min, med, max] for metric value among query runs:
```
8 [4000,4000,36000] Syste... | 2020-07-07T16:54:37 | 2020-07-07T16:54:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,048,339 | 11,971 | alexey-milovidov | Do I understand correctly, that performance test creates and fills all tables first and then runs all the tests with all substitutions?
Looks like there is a room for improvement... | 2020-07-07T18:37:04 | 2020-07-07T18:37:04 | {} | MEMBER |
ClickHouse | ClickHouse | 655,053,535 | 11,564 | bobrik | I've tried my change on a production dataset.
24h query, before and after (aligned for visual aid):
```
Read 4562944 rows, 148.05 MiB in 45.19249672 sec., 100966 rows/sec., 3.28 MiB/sec.
Read 4183040 rows, 135.78 MiB in 0.196279627 sec., 21311636 rows/sec., 691.75 MiB/sec.
``` | 2020-07-07T18:47:45 | 2020-07-07T18:47:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,054,121 | 12,179 | alexey-milovidov | > Yandex synchronization check (only for Yandex employees)
It simply does not support new tests. | 2020-07-07T18:48:59 | 2020-07-07T18:48:59 | {} | MEMBER |
ClickHouse | ClickHouse | 655,356,251 | 12,261 | jessicacheun | Good question! I test it ! Same result ! Is there anybody know it ? | 2020-07-08T07:58:39 | 2020-07-08T07:58:39 | {} | NONE |
ClickHouse | ClickHouse | 655,393,738 | 12,261 | jessicacheun | @andyyzh help | 2020-07-08T09:08:44 | 2020-07-08T09:08:44 | {} | NONE |
ClickHouse | ClickHouse | 654,562,749 | 12,171 | taiyang-li | > Could you please provide an example of how does it look?
Log will like this:
```
2020.07.07 10:27:13.964506 [ 3816 ] {} <Information> TCPHandler[372]: Processed in 0.007 sec.
2020.07.07 10:27:13.960542 [ 3816 ] {45a4d70e-6236-4930-b2d3-0908551c179b} <Trace> AccessRightsContext (default)[112]: Access granted: S... | 2020-07-07T02:35:31 | 2020-07-07T02:36:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 654,789,969 | 12,195 | qoega | Can you check speed of insert if numbers_mt is used? | 2020-07-07T11:28:10 | 2020-07-07T11:28:10 | {} | MEMBER |
ClickHouse | ClickHouse | 654,827,018 | 1,576 | SergeyMirvoda | My scenario:
from server 1
q1: ``create table x1 as select from remote``
DataGrip shows read timed out error.
Then queries like
``drop x1`` becomes immortal
after
``kill query where query_id='q1'`` (freezes)
``show process`` also freezes.
```
Cannot load data for command line suggestions: Code: 209,... | 2020-07-07T12:37:43 | 2020-07-07T12:38:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 654,829,216 | 12,195 | den-crane | @ucasFL try engine MergeTree instead of Log
```
20.6.1.3853
CREATE TABLE test (n UInt64) ENGINE=MergeTree() order by tuple()
set max_threads=16, max_insert_threads=16
insert into test select * from numbers_mt(1000000000) union all select * from numbers_mt(1000000000);
Elapsed: 19.444 sec. Processed 2.00 billi... | 2020-07-07T12:41:20 | 2020-07-07T12:43:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 654,846,313 | 12,171 | taiyang-li | > There is also another place -
>
> https://github.com/ClickHouse/ClickHouse/blob/488a8cbc9fc15d5f2fde70e6cff17521fc195c1d/src/DataStreams/InternalTextLogsRowOutputStream.cpp#L96
>
> But, is it really useful? I doubt that there are duplicated log messages on different lines
When SET send_logs_level = 'trace' ,... | 2020-07-07T13:10:34 | 2020-07-07T13:10:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 654,970,285 | 12,257 | filimonov | BTW on 20.5 the same looks better:
```
2020.07.07 16:07:20.579525 [ 1 ] {} <Error> Access(users.xml): Could not parse profile `default`: Code: 27, e.displayText() = DB::Exception: Cannot parse input: expected 'eof' before: '.0': in attempt to set the value of setting to '100000000.0', Stack trace (when copying this... | 2020-07-07T16:13:54 | 2020-07-07T16:13:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,078,059 | 11,572 | laingawbl | @den-crane whoops, you're right. I misunderstood totally. Sorry! | 2020-07-07T19:32:19 | 2020-07-07T19:42:44 | {} | NONE |
ClickHouse | ClickHouse | 655,141,766 | 12,179 | alexey-milovidov | Yes, the last patch (with `optional`) does not make sense.
We can return false (that the function is not injective) for `dictGet` if empty block was passed.
Also document this behaviour in the interface of `isInjective` - you can pass empty block to check if the function is injective with any arguments. | 2020-07-07T21:21:19 | 2020-07-07T21:21:19 | {} | MEMBER |
ClickHouse | ClickHouse | 655,173,389 | 8,212 | alexey-milovidov | @KochetovNicolai is on vacation, I will take this. | 2020-07-07T22:40:06 | 2020-07-07T22:40:06 | {} | MEMBER |
ClickHouse | ClickHouse | 655,235,806 | 12,163 | alexey-milovidov | Ok. It reproduces with `clickhouse-benchmark`. | 2020-07-08T02:03:10 | 2020-07-08T02:03:10 | {} | MEMBER |
ClickHouse | ClickHouse | 655,236,071 | 12,163 | alexey-milovidov | The crucial part is WITH TOTALS.
| 2020-07-08T02:04:04 | 2020-07-08T02:04:04 | {} | MEMBER |
ClickHouse | ClickHouse | 655,239,648 | 12,163 | alexey-milovidov | ```
$ clickhouse-benchmark <<< "SELECT finalizeAggregation(groupArrayState(x)) FROM (SELECT intDiv(number, 3) AS k, toNullable(1) AS x FROM system.numbers LIMIT 100000) GROUP BY k WITH TOTALS"Loaded 1 queries.
An error occurred while processing the query 'SELECT finalizeAggregation(groupArrayState(x)) FROM (SELECT in... | 2020-07-08T02:16:32 | 2020-07-08T02:16:32 | {} | MEMBER |
ClickHouse | ClickHouse | 655,400,845 | 12,261 | jessicacheun | > 1. AggregatingMergeTree
I tried AggregatingMergeTree either, returning the same result | 2020-07-08T09:21:42 | 2020-07-08T09:21:42 | {} | NONE |
ClickHouse | ClickHouse | 654,505,615 | 12,179 | alexey-milovidov | Ok. | 2020-07-06T23:01:46 | 2020-07-06T23:01:46 | {} | MEMBER |
ClickHouse | ClickHouse | 654,938,199 | 12,160 | azat | >Looks great! Have you tested it manually?
@abyss7 yep:
> I also tested it with the following patch:
>
> ```
> $ git di
> diff --git a/base/daemon/CMakeLists.txt b/base/daemon/CMakeLists.txt
> index 26d59a57e7..35e6ff6432 100644
> --- a/base/daemon/CMakeLists.txt
> +++ b/base/daemon/CMakeLists.txt
> @@ -9... | 2020-07-07T15:21:36 | 2020-07-07T15:21:36 | {} | MEMBER |
ClickHouse | ClickHouse | 654,979,451 | 12,263 | akuzm | Crashed the playground. | 2020-07-07T16:30:00 | 2020-07-07T16:30:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 654,994,008 | 12,195 | den-crane | @ucasFL I just created an example that demonstrates that the parallel insertion works with numbers_mt+max_insert_threads against ENGINE=MergeTree. Nothing more.
>Test is unstable, and sometimes this happened:
As expected for slow disks.
| 2020-07-07T16:55:49 | 2020-07-07T16:55:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,042,892 | 12,195 | alexey-milovidov | @qoega
> In current version it seems invalid. There could be more complex logic when we can set less threads for trivial insert queries.
Could you please elaborate? | 2020-07-07T18:26:18 | 2020-07-07T18:26:18 | {} | MEMBER |
ClickHouse | ClickHouse | 655,092,759 | 12,090 | alesapin | ` Testflows check — failed:0, passed:3, other:0 `, ok let's merge and enable it tomorrow. | 2020-07-07T19:56:07 | 2020-07-07T19:56:07 | {} | MEMBER |
ClickHouse | ClickHouse | 655,261,144 | 12,276 | alexey-milovidov | @alesapin Unit tests did not run in 4 hours. | 2020-07-08T03:31:05 | 2020-07-08T03:31:05 | {} | MEMBER |
ClickHouse | ClickHouse | 655,343,654 | 12,256 | alesapin | Performance tests are not related to changes. | 2020-07-08T07:33:38 | 2020-07-08T07:33:38 | {} | MEMBER |
ClickHouse | ClickHouse | 654,542,323 | 12,150 | dkoston | Thanks @alexey-milovidov! | 2020-07-07T01:16:05 | 2020-07-07T01:16:05 | {} | NONE |
ClickHouse | ClickHouse | 654,902,525 | 12,188 | filimonov | ```
DROP TABLE IF EXISTS test_mut;
CREATE TABLE test_mut
ENGINE = MergeTree
PARTITION BY k
ORDER BY k AS
SELECT
number % 5 AS k,
0 AS val
FROM numbers(20);
SELECT
_part,
k,
groupArray(val)
FROM test_mut
GROUP BY
_part,
k
ORDER BY k ASC
┌─_part───┬─k─┬─groupArr... | 2020-07-07T14:28:12 | 2020-07-07T14:28:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,063,219 | 11,572 | den-crane | @laingawbl OPTIMIZE DEDUPLICATE has sense only for simple MergeTree engine.
You should never use DEDUPLICATE against SummingMergeTree. It's the nonsense. | 2020-07-07T19:05:53 | 2020-07-07T19:05:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,179,546 | 11,873 | alexey-milovidov | > 01193_metadata_loading
Flaky test. | 2020-07-07T22:58:45 | 2020-07-07T22:58:45 | {} | MEMBER |
ClickHouse | ClickHouse | 655,238,157 | 12,163 | alexey-milovidov | Any complex aggregate function will suffice:
```
clickhouse-benchmark <<< "SELECT k, finalizeAggregation(uniqState(x)) FROM (WITH toNullable(number = 3 ? 3 : 1) AS d SELECT intDiv(number, 3) AS k, number % d AS x FROM system.numbers LIMIT 100000) GROUP BY k WITH TOTALS"
``` | 2020-07-08T02:11:24 | 2020-07-08T02:11:24 | {} | MEMBER |
ClickHouse | ClickHouse | 654,601,002 | 12,152 | bohutang | Could you have a review or comments Cc @alexey-milovidov | 2020-07-07T05:00:39 | 2020-07-07T05:00:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 654,660,988 | 9,092 | filimonov | @VeryaskinMax your sample is not correct (it references non-existing column).
Closing due to lack of activity.
@VeryaskinMax please reopen the issue (or open a new one) once you will have a reproducible example. | 2020-07-07T07:39:36 | 2020-07-07T07:39:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 654,669,458 | 7,673 | filimonov | The issue is still there:
```
SELECT toDateTime(CAST(NULL, 'Nullable(String)'))
Received exception from server (version 20.6.1):
Code: 41. DB::Exception: Received from localhost:9000. DB::Exception: Cannot parse datetime: Cannot parse DateTime from String.
0 rows in set. Elapsed: 0.012 sec.
``` | 2020-07-07T07:55:42 | 2020-07-07T07:55:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 654,996,222 | 12,142 | akuzm | I didn't realize we trace every allocation above 1 MB now. Just seen a query where profiling takes 25% of the query time, I'm reverting this now. Probably will have to disable CPU profiling for measured query runs as well, because the instability is just too much. Instead of that, I'll try to add both memory and CPU pr... | 2020-07-07T16:59:35 | 2020-07-07T16:59:35 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,101,534 | 12,090 | vzakaznikov | > `Testflows check — failed:0, passed:3, other:0`, ok let's merge and enable it tomorrow.
Thanks a lot! I will update our LDAP PR (https://github.com/ClickHouse/ClickHouse/pull/11234) with these changes. | 2020-07-07T20:10:01 | 2020-07-07T20:10:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 654,876,514 | 12,256 | alesapin | Waiting for tests. | 2020-07-07T13:51:13 | 2020-07-07T13:51:13 | {} | MEMBER |
ClickHouse | ClickHouse | 655,055,982 | 11,815 | alexey-milovidov | Cannot clone submodules:
https://clickhouse-test-reports.s3.yandex.net/0/9c829e3046cae0711ca822998d1a6b3e59b54edb/clone_repo/submodules_init.err.log
Need to add retries. | 2020-07-07T18:52:34 | 2020-07-07T18:52:34 | {} | MEMBER |
ClickHouse | ClickHouse | 655,201,088 | 11,189 | alexey-milovidov | This PR is incorrect. | 2020-07-08T00:00:54 | 2020-07-08T00:00:54 | {} | MEMBER |
ClickHouse | ClickHouse | 655,307,675 | 12,181 | azat | >Integration tests (asan) — failed:4, passed:729, error:0
Looks unrelated:
test_version_update_after_mutation/test.py::test_upgrade_while_mutation | FAIL
-- | --
test_adaptive_granularity/test.py::test_version_update_two_nodes | FAIL
test_storage_rabbitmq/test.py::test_rabbitmq_read_only_combo | FAIL
test_ver... | 2020-07-08T06:07:09 | 2020-07-08T06:07:09 | {} | MEMBER |
ClickHouse | ClickHouse | 654,709,758 | 12,009 | tavplubix | There is a bug:
```
$ clickhouse-local --output-format=TSVRaw -q "select 1000 from numbers(1000000)" > data.tsv
$ clickhouse-local --input-format=TSVRaw -S "n UInt64" -q "select count(n), sum(n) from table" < data.tsv
Code: 27, e.displayText() = DB::Exception: Cannot parse input: expected '\n' before: '000\n1000\n1... | 2020-07-07T09:04:54 | 2020-07-07T09:04:54 | {} | MEMBER |
ClickHouse | ClickHouse | 654,780,309 | 12,185 | shinoi2 | In my case, I have a table to save vip recharge record. like this
user-id | timestamp | amount | unit
| ---- | ---- | ---- | ---- |
a | 2019-01-01 00:00:00 | 1 | year
a | 2019-02-01 00:00:00 | 1 | month
a | 2019-02-01 00:00:00 | 1 | day
I want to calculate vip deadline
`INTERVAL 1 YEAR` same with `toInterv... | 2020-07-07T11:10:57 | 2020-07-07T11:45:05 | {} | NONE |
ClickHouse | ClickHouse | 655,184,315 | 9,577 | alexey-milovidov | The fix is flawed, the test also flawed. Nothing remains. | 2020-07-07T23:07:55 | 2020-07-07T23:07:55 | {} | MEMBER |
ClickHouse | ClickHouse | 655,237,444 | 12,163 | alexey-milovidov | And Nullable.
```
clickhouse-benchmark <<< "SELECT k, finalizeAggregation(quantilesTimingState(0.5)(x)) FROM (WITH toNullable(number = 3 ? 3 : 1) AS d SELECT intDiv(number, d) AS k, number % d AS x FROM system.numbers LIMIT 100000) GROUP BY k WITH TOTALS"
``` | 2020-07-08T02:08:43 | 2020-07-08T02:09:04 | {} | MEMBER |
ClickHouse | ClickHouse | 655,341,590 | 12,283 | filimonov | Why do you connect from SuperSet using Mysql protocol?
It have native support for clickhouse: https://github.com/apache/incubator-superset/pull/1844
About MySQL compatibility - see https://github.com/ClickHouse/ClickHouse/issues/9336 @BohuTANG is working on that, AFAIK.
| 2020-07-08T07:29:21 | 2020-07-08T07:29:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 667,692,491 | 13,201 | alexey-milovidov | > FWIW looks like now each PR may contain some failures of AST query fuzzer, @alexey-milovidov @akuzm maybe it will be better to run it constantly in background, instead of for each PR? (at least until all existing issues will be fixed)
Let's just fix all the issues.
We will also add longer runs for the new tests i... | 2020-08-02T16:05:13 | 2020-08-02T16:05:13 | {} | MEMBER |
ClickHouse | ClickHouse | 667,695,003 | 13,239 | alexey-milovidov | Waiting only for build. | 2020-08-02T16:25:52 | 2020-08-02T16:25:52 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.