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 | 860,868,447 | 10,892 | den-crane | ```
create table xx ( A DateTime, B Float64, C UInt8, D UInt8) Engine=MergeTree order by A;
insert into xx select now(), 1/(number+1), 1, 1 from numbers(1000000000) ;
alter table xx modify column B Float64 CODEC(DoubleDelta, ZSTD(22));
```
session 1 | session 2 | session 3
----------|----------|----------
optimi... | 2021-06-14T17:41:15 | 2021-06-14T17:41:15 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 860,882,050 | 21,794 | alexey-milovidov | > For query SELECT a FROM t1 FULL JOIN t2 USING (a) we will have column that contains all values conatenated from t1.a and t2.a, so a in result table have to hold all values
@vdimir Let's apply it only for `INNER JOIN`. | 2021-06-14T18:02:20 | 2021-06-14T18:02:20 | {} | MEMBER |
ClickHouse | ClickHouse | 861,152,680 | 18,227 | alexey-milovidov | This is misbehaviour of ZooKeeper or a network error. | 2021-06-15T04:07:03 | 2021-06-15T04:07:03 | {} | MEMBER |
ClickHouse | ClickHouse | 860,286,561 | 22,260 | alexey-milovidov | We will do it only if we will see the bottleneck in TCP interaction in some production use cases. | 2021-06-13T23:36:28 | 2021-06-13T23:36:28 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 860,287,205 | 21,748 | alexey-milovidov | Also there was a use case in Yandex.Metrica when we pass some hint (arbitrary random string) to the API that will allow to prefer specific API server (sticky balancing) and then `load_balancing = 'nearest_hostname'` is used to implement sticky balancing on ClickHouse side.
This is to get consistent results on subseq... | 2021-06-13T23:41:55 | 2021-06-13T23:41:55 | {} | MEMBER |
ClickHouse | ClickHouse | 860,289,475 | 21,018 | alexey-milovidov | We have recently fixed several possible issues when interaction with ZooKeeper may stuck in very low memory environments. | 2021-06-13T23:59:59 | 2021-06-13T23:59:59 | {} | MEMBER |
ClickHouse | ClickHouse | 860,572,236 | 25,011 | kssenii | @Mergifyio update | 2021-06-14T10:19:20 | 2021-06-14T10:19:20 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 860,849,639 | 10,892 | den-crane | 21.7.1.7020
```sql
create table xx ( A DateTime, B Float64, C UInt8) Engine=MergeTree order by A;
insert into xx select now(), 1/(number+1), 1 from numbers(1000000000)
alter table xx modify column B Float64 CODEC(DoubleDelta, ZSTD(22)); -- artificially slowsdown merges
```
session 1 | session 2
----------|-... | 2021-06-14T17:12:34 | 2021-06-14T17:12:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 860,894,449 | 25,067 | alexey-milovidov | No objections... | 2021-06-14T18:21:20 | 2021-06-14T18:21:20 | {} | MEMBER |
ClickHouse | ClickHouse | 861,117,046 | 23,555 | alexey-milovidov | The issue has been fixed in #24483 and the fix will be available in version 21.7. | 2021-06-15T02:12:39 | 2021-06-15T02:12:39 | {} | MEMBER |
ClickHouse | ClickHouse | 861,151,297 | 18,875 | alexey-milovidov | Does not reproduce. | 2021-06-15T04:03:02 | 2021-06-15T04:03:02 | {} | MEMBER |
ClickHouse | ClickHouse | 861,158,764 | 16,348 | alexey-milovidov | Cannot reproduce. | 2021-06-15T04:25:47 | 2021-06-15T04:25:47 | {} | MEMBER |
ClickHouse | ClickHouse | 860,286,296 | 22,260 | alexey-milovidov | Let's avoid complications. | 2021-06-13T23:34:22 | 2021-06-13T23:34:22 | {} | MEMBER |
ClickHouse | ClickHouse | 860,349,526 | 25,064 | ucasfl | > Hi, thank you for your contributuon!
>
> Please provide a better description of the feature, since right now it is not clear **why** users might want to have such a feature. Also, please fix other comments.
It comes from MySQL: query multiple MySQL databases at once. | 2021-06-14T04:14:23 | 2021-06-14T04:14:23 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 860,351,129 | 20,315 | alexey-milovidov | @vdimir The original query no longer reproduce the error but the example of Qoega is still reproducible. | 2021-06-14T04:19:29 | 2021-06-14T04:19:29 | {} | MEMBER |
ClickHouse | ClickHouse | 860,634,525 | 18,667 | den-crane | No info. No repro. | 2021-06-14T12:11:00 | 2021-06-14T12:11:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 860,927,244 | 25,272 | alexey-milovidov | It has neglible change to be enabled by default for Enums, because it will break many types of queries.
So, the setting `optimize_if_transform_strings_to_enum` remains for expert use cases. | 2021-06-14T19:12:39 | 2021-06-14T19:12:39 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 861,153,409 | 17,964 | alexey-milovidov | Does not reproduce. | 2021-06-15T04:09:24 | 2021-06-15T04:09:24 | {} | MEMBER |
ClickHouse | ClickHouse | 860,289,283 | 20,608 | alexey-milovidov | Not being able to reproduce. | 2021-06-13T23:58:10 | 2021-06-13T23:58:10 | {} | MEMBER |
ClickHouse | ClickHouse | 860,355,636 | 25,160 | alexey-milovidov | Configure two clusters.
One contains single replica and is used for SELECTs and another contains both replicas and is used for INSERTs. | 2021-06-14T04:34:52 | 2021-06-14T04:34:59 | {} | MEMBER |
ClickHouse | ClickHouse | 860,275,507 | 12,563 | alexey-milovidov | In old versions, mark cache can grow above the mark_cache_size. | 2021-06-13T21:56:45 | 2021-06-13T21:56:45 | {} | MEMBER |
ClickHouse | ClickHouse | 860,286,730 | 21,748 | alexey-milovidov | This is interesting... In our use cases we allocate additional two nodes that look on cluster and analysts connect to either one or another and create their temporary tables there. | 2021-06-13T23:38:11 | 2021-06-13T23:38:11 | {} | MEMBER |
ClickHouse | ClickHouse | 860,357,276 | 19,689 | alexey-milovidov | @vitlibar We still need the fix. | 2021-06-14T04:41:01 | 2021-06-14T04:41:01 | {} | MEMBER |
ClickHouse | ClickHouse | 861,071,531 | 24,036 | alexey-milovidov | It may happen if some third-party tool running on your server is removed the file `/var/log/clickhouse-server/clickhouse-server.log.0`.
Most likely it is an error in your infrastructure, not in ClickHouse. | 2021-06-15T00:09:09 | 2021-06-15T00:09:09 | {} | MEMBER |
ClickHouse | ClickHouse | 861,151,835 | 18,398 | alexey-milovidov | Fixed, duplicate of #19795. | 2021-06-15T04:04:42 | 2021-06-15T04:04:42 | {} | MEMBER |
ClickHouse | ClickHouse | 860,281,328 | 24,261 | alexey-milovidov | ```
../src/Common/assert_cast.h:50:12: runtime error: downcast of address 0x7f74e2267d60 which does not point to an object of type 'const DB::ColumnNullable'
0x7f74e2267d60: note: object is of type 'DB::ColumnString'
00 00 00 00 c0 9f 34 0a 00 00 00 00 01 00 00 00 00 00 00 00 00 74 8c e2 74 7f 00 00 08 74 8c e2... | 2021-06-13T22:53:56 | 2021-06-13T22:53:56 | {} | MEMBER |
ClickHouse | ClickHouse | 860,881,606 | 21,794 | alexey-milovidov | @akuzm
> transform the condition into left.key = castOrNull(<left key type>, right.key), and it just magically works. The right keys that don't fit the left type range are converted to Null, so the join condition doesn't hold for them. Maybe we can again apply this approach here?
Absolutely. We already have `acc... | 2021-06-14T18:01:37 | 2021-06-14T18:01:37 | {} | MEMBER |
ClickHouse | ClickHouse | 859,684,178 | 25,158 | yanye666 | > What level is this window? In other words, does this window serve a group of rows of the same sort key? Still serve this table?
> If it only acts on the rows of the same sort key, then the memory will not be released when the 8189 rows can be merged?
> ’we can collapse only those rows. When we collapse any two rows... | 2021-06-11T16:05:07 | 2021-06-15T07:22:21 | {} | NONE |
ClickHouse | ClickHouse | 860,280,192 | 21,267 | alexey-milovidov | > going back to direct usage of librdkafka c API
I'm totally ok with that. Less layers - more easy to understand. | 2021-06-13T22:42:00 | 2021-06-13T22:42:00 | {} | MEMBER |
ClickHouse | ClickHouse | 860,285,170 | 7,648 | alexey-milovidov | We don't have this restriction in `DDLWorker` anymore. | 2021-06-13T23:26:00 | 2021-06-13T23:26:00 | {} | MEMBER |
ClickHouse | ClickHouse | 860,339,659 | 25,139 | lopn | @abyss7 what time can you add configure? i want to use it | 2021-06-14T03:30:38 | 2021-06-14T03:30:38 | {} | NONE |
ClickHouse | ClickHouse | 860,511,343 | 25,215 | andrelec1 | I think i have to switch to `VersionedCollapsingMergeTree` Engine for my slave ...
Using timestamp for `version`, for only keeping the last value ...
i'll make some test with it ...
But i didn't find how can i realy alter ( delete ) that data properly ...
any idea is welcome . | 2021-06-14T08:47:40 | 2021-06-14T08:47:40 | {} | NONE |
ClickHouse | ClickHouse | 860,902,575 | 10,210 | alexey-milovidov | Use the `input` table function: INSERT ... SELECT ... FROM input(...).
https://clickhouse.tech/docs/en/sql-reference/table-functions/input/ | 2021-06-14T18:33:32 | 2021-06-14T18:33:42 | {
"+1": 2
} | MEMBER |
ClickHouse | ClickHouse | 860,903,670 | 7,666 | alexey-milovidov | Workaround:
SELECT any(version()), any(uptime()) FROM cluster(segmented, system, one) **GROUP BY hostName()** | 2021-06-14T18:35:25 | 2021-06-14T18:35:25 | {} | MEMBER |
ClickHouse | ClickHouse | 861,093,216 | 11,098 | alexey-milovidov | Most likely it is caused by misconfiguration. Check server logs that replica is being able to connect to another replica and download data. If not, you will see error messages with explanation. | 2021-06-15T01:01:10 | 2021-06-15T01:01:10 | {} | MEMBER |
ClickHouse | ClickHouse | 861,107,391 | 11,124 | alexey-milovidov | Amesaru did not try to do this task. | 2021-06-15T01:46:05 | 2021-06-15T01:46:05 | {} | MEMBER |
ClickHouse | ClickHouse | 860,313,946 | 25,259 | alexey-milovidov | `SELECT length(topK(5592405)(tuple(number))) FROM numbers(100) GROUP BY number` | 2021-06-14T01:41:56 | 2021-06-14T01:41:56 | {} | MEMBER |
ClickHouse | ClickHouse | 860,326,086 | 23,645 | alexey-milovidov | We have removed throwing from MemoryTracker in operator new due to per-query/user/server memory limits.
It can still throw bad_alloc, but in that case, std::terminate is totally Ok. | 2021-06-14T02:32:06 | 2021-06-14T02:32:06 | {} | MEMBER |
ClickHouse | ClickHouse | 860,569,339 | 23,864 | azat | @alexey-milovidov any thoughts about this? | 2021-06-14T10:14:19 | 2021-06-14T10:14:19 | {} | MEMBER |
ClickHouse | ClickHouse | 860,896,768 | 25,267 | alexey-milovidov | Ok.
I'm wondering, how did you find it?
We have multiple builds in our CI (gcc, clang, on x64, arm, linux, freebsd...) and all of them are green. | 2021-06-14T18:24:47 | 2021-06-14T18:24:47 | {} | MEMBER |
ClickHouse | ClickHouse | 861,107,765 | 11,154 | alexey-milovidov | @qoega This usability issue is still waiting your fix. | 2021-06-15T01:47:29 | 2021-06-15T01:47:29 | {} | MEMBER |
ClickHouse | ClickHouse | 860,285,368 | 11,731 | alexey-milovidov | Yes, this is not easy to do. | 2021-06-13T23:27:15 | 2021-06-13T23:27:15 | {} | MEMBER |
ClickHouse | ClickHouse | 860,302,732 | 22,069 | alexey-milovidov | Looks like it really needs this amount of memory to read data from table. | 2021-06-14T01:02:45 | 2021-06-14T01:02:45 | {} | MEMBER |
ClickHouse | ClickHouse | 860,327,243 | 21,592 | alexey-milovidov | Does not reproduce. | 2021-06-14T02:37:44 | 2021-06-14T02:37:44 | {} | MEMBER |
ClickHouse | ClickHouse | 860,572,595 | 25,011 | mergify[bot] | **Command `update`: success**
> **Branch has been successfully updated**
>
| 2021-06-14T10:19:58 | 2021-06-14T10:19:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 860,696,309 | 10,892 | SaltTan | > @den-crane It always works instantly in recent ClickHouse releases, no locking is performed.
@alexey-milovidov Metadata changes are applied instantly.
But actual dropping is a mutation, and it gets queued and waits for running merges. | 2021-06-14T13:44:22 | 2021-06-14T13:44:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 860,898,162 | 25,263 | alexey-milovidov | @filimonov
>
test_storage_kafka/test.py::test_kafka_consumer_failover | FAIL
-- | --
| 2021-06-14T18:26:48 | 2021-06-14T18:26:48 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 861,081,686 | 23,700 | alexey-milovidov | The issue is fixed in #24157.
The fix is backported to versions 21.5 and 21.6. | 2021-06-15T00:34:35 | 2021-06-15T00:34:35 | {} | MEMBER |
ClickHouse | ClickHouse | 861,141,868 | 20,549 | alexey-milovidov | ```
select rand() % 2 ? materialize(1)::Decimal(18, 10) : 1
``` | 2021-06-15T03:30:15 | 2021-06-15T03:30:15 | {} | MEMBER |
ClickHouse | ClickHouse | 861,158,483 | 16,674 | alexey-milovidov | @nikitamikhaylov maybe reassign to @Avogar? | 2021-06-15T04:24:49 | 2021-06-15T04:24:49 | {} | MEMBER |
ClickHouse | ClickHouse | 860,285,520 | 12,478 | alexey-milovidov | Duplicate of #1270. | 2021-06-13T23:28:20 | 2021-06-13T23:28:20 | {} | MEMBER |
ClickHouse | ClickHouse | 860,285,734 | 12,489 | alexey-milovidov | No. It is in "discussion" stage. | 2021-06-13T23:29:47 | 2021-06-13T23:29:47 | {} | MEMBER |
ClickHouse | ClickHouse | 860,650,498 | 25,239 | den-crane | You can set another value for ` <path>/var/lib/clickhouse/</path>` /etc/clickhouse-server/config.xml
Or you can change default policy.
```
cat /etc/clickhouse-server/config.d/storage.xml
<?xml version="1.0" ?>
<yandex>
<storage_configuration>
<disks>
<disk1>
... | 2021-06-14T12:37:56 | 2021-06-14T12:37:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 861,070,369 | 23,779 | alexey-milovidov | The issue is fixed in #24157.
The fix is backported to versions 21.5 and 21.6. | 2021-06-15T00:06:25 | 2021-06-15T00:06:25 | {} | MEMBER |
ClickHouse | ClickHouse | 861,088,769 | 22,438 | alexey-milovidov | Fixed in #24483. | 2021-06-15T00:51:11 | 2021-06-15T00:51:11 | {} | MEMBER |
ClickHouse | ClickHouse | 861,125,260 | 10,892 | yingshikong186 | > @yingshikong186 What issues do you have with this version?
i have resolve this problem with
` KILL MUTATION WHERE database = 'default' AND table = 'table' `。
This problem may be due to the wrong sql
| 2021-06-15T02:38:32 | 2021-06-15T02:38:32 | {} | NONE |
ClickHouse | ClickHouse | 861,228,465 | 24,867 | bharatnc | @nikitamikhaylov Builds look ok to me. The failures I see seem unrelated.
Let me know if the changes and tests that I added look good to you.
| 2021-06-15T06:51:43 | 2021-06-15T06:51:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 860,289,032 | 17,185 | alexey-milovidov | We have implemented window functions. | 2021-06-13T23:56:19 | 2021-06-13T23:56:19 | {} | MEMBER |
ClickHouse | ClickHouse | 860,357,845 | 24,158 | alexey-milovidov | @azat Could you please check, maybe it is somehow related to squashing of AggregateFunction columns or LowCardinality. | 2021-06-14T04:43:11 | 2021-06-14T04:43:11 | {} | MEMBER |
ClickHouse | ClickHouse | 861,086,853 | 25,230 | alexey-milovidov | The comment just warns that the destructor of ThreadFromGlobalPool won't be called in case of exception.
Although it does not have any significant logic, only an assertion:
```
if (joinable())
abort();
``` | 2021-06-15T00:46:43 | 2021-06-15T00:46:43 | {} | MEMBER |
ClickHouse | ClickHouse | 861,154,707 | 18,971 | alexey-milovidov | It works as expected. | 2021-06-15T04:13:27 | 2021-06-15T04:13:27 | {} | MEMBER |
ClickHouse | ClickHouse | 861,160,438 | 12,264 | alexey-milovidov | CC @zhang2014 | 2021-06-15T04:30:35 | 2021-06-15T04:30:35 | {} | MEMBER |
ClickHouse | ClickHouse | 860,284,095 | 4,226 | alexey-milovidov | The demand for this feature is still low. Please add more likes if you want it to be implemented. | 2021-06-13T23:17:17 | 2021-06-13T23:17:17 | {
"+1": 2
} | MEMBER |
ClickHouse | ClickHouse | 860,289,364 | 21,594 | alexey-milovidov | No info provided. | 2021-06-13T23:58:58 | 2021-06-13T23:58:58 | {} | MEMBER |
ClickHouse | ClickHouse | 860,290,073 | 23,730 | alexey-milovidov | That's true - connections between servers in a cluster are long living. | 2021-06-14T00:02:50 | 2021-06-14T00:02:50 | {} | MEMBER |
ClickHouse | ClickHouse | 860,645,784 | 25,261 | den-crane | 1,2. CH keeps only recent mutations
Merge_tree setting
```finished_mutations_to_keep │ 100 │ 0 │ How many records about mutations that are done to keep. If zero, then keep all of them.```
https://clickhouse.tech/docs/en/operations/settings/merge-tree-settings/#merge-tree-se... | 2021-06-14T12:30:20 | 2021-06-14T12:30:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 861,047,892 | 24,431 | vzakaznikov | @traceon, see https://github.com/ClickHouse/ClickHouse/pull/25278/commits/33f270c2dceb4eab3522a61ef76de1f5d6126edf for the new test. | 2021-06-14T23:01:45 | 2021-06-14T23:01:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 861,089,436 | 1,158 | alexey-milovidov | `quantileTDigest` accuracy has been improved.
Although there is no guarantee that it gives better accuracy than `quantile`. | 2021-06-15T00:52:40 | 2021-06-15T00:52:40 | {} | MEMBER |
ClickHouse | ClickHouse | 861,094,720 | 3,575 | alexey-milovidov | @GinGin3203 Yes, the task is easy to implement.
We have no current plans to do it. | 2021-06-15T01:05:28 | 2021-06-15T01:05:50 | {} | MEMBER |
ClickHouse | ClickHouse | 859,564,789 | 25,158 | yanye666 | > Hi!
> It is expected behaviour (or implementation detail, if you wish). It would be very inexpensive to merge all the rows in case you have, like 1 billion same rows with +1 sign, and then 1 billion same rows with -1 sign. Implementation uses a window, it's size about 8192 (minus 3, probably). Only rows are in the s... | 2021-06-11T13:00:40 | 2021-06-15T07:22:54 | {} | NONE |
ClickHouse | ClickHouse | 861,253,901 | 25,215 | shacuros | Hey,
first of all, i'd suggest to read more about the clickhouse topology logic or watch some tutorials about it (i can advice to altinity youtube channel for that) Clickhouse does not have a master/slave topology. It just has shards and replicas. So if i interprete your setting correctly you have two shards and eac... | 2021-06-15T07:26:20 | 2021-06-15T07:26:20 | {
"hooray": 1
} | NONE |
ClickHouse | ClickHouse | 860,289,639 | 22,385 | alexey-milovidov | We have removed this "feature". | 2021-06-14T00:00:46 | 2021-06-14T00:00:46 | {} | MEMBER |
ClickHouse | ClickHouse | 860,875,127 | 5,986 | alexey-milovidov | CC @azat
Yes, it makes sense. If we apply GROUP BY sharding key, we can process it independently.
(if the setting to trust sharding key is enabled)
AFAIK it is already implemented, but I'm not sure if it works with subqueries. | 2021-06-14T17:51:55 | 2021-06-14T17:51:55 | {} | MEMBER |
ClickHouse | ClickHouse | 860,897,816 | 25,245 | alexey-milovidov | test_cluster_copier is broken in 21.6. | 2021-06-14T18:26:15 | 2021-06-14T18:26:15 | {} | MEMBER |
ClickHouse | ClickHouse | 861,159,541 | 14,451 | alexey-milovidov | The behaviour is correct. | 2021-06-15T04:28:04 | 2021-06-15T04:28:04 | {} | MEMBER |
ClickHouse | ClickHouse | 861,163,065 | 25,265 | alexey-milovidov | Arcadia is fixed manually. | 2021-06-15T04:38:26 | 2021-06-15T04:38:26 | {} | MEMBER |
ClickHouse | ClickHouse | 860,283,903 | 2,358 | alexey-milovidov | Жалко символов. Хотя для совместимости с MySQL можно сделать. | 2021-06-13T23:15:41 | 2021-06-13T23:15:41 | {} | MEMBER |
ClickHouse | ClickHouse | 860,350,477 | 21,730 | alexey-milovidov | ```
CREATE TABLE l
(
`x` UInt32,
`lc` String
)
ENGINE = Memory
Query id: b5909ce0-ab0a-477f-81a6-cdab2d7e47b6
Ok.
0 rows in set. Elapsed: 0.001 sec.
CREATE TABLE r
(
`x` UInt32,
`lc` String
)
ENGINE = Memory
Query id: 400fb36f-7dc0-41b5-b653-84a1a4516679
Ok.
0 rows in ... | 2021-06-14T04:17:26 | 2021-06-14T04:17:26 | {} | MEMBER |
ClickHouse | ClickHouse | 860,454,446 | 24,413 | kssenii | Will wait for update in https://github.com/jtv/libpqxx/issues/455 instead of adding a lot of *nolint*. | 2021-06-14T07:31:29 | 2021-06-14T07:31:29 | {} | MEMBER |
ClickHouse | ClickHouse | 860,923,837 | 10,892 | alexey-milovidov | Maybe we forgot to remove some locks from the code :cry: | 2021-06-14T19:06:45 | 2021-06-14T19:06:45 | {} | MEMBER |
ClickHouse | ClickHouse | 860,969,317 | 21,191 | filimonov | > It would be great if you provide exact logs and errors about inability to build MySQL in your particular configuration.
Have no idea (i have no Mac) . It is simply disabled in CI builds
https://github.com/ClickHouse/ClickHouse/blob/e5f49477a6873e43305341e00b9111e870a7fedf/cmake/find/mysqlclient.cmake#L1 | 2021-06-14T20:26:21 | 2021-06-14T20:27:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 861,019,346 | 21,191 | alexey-milovidov | Let's wait if some mysterious person will implement it. | 2021-06-14T21:51:08 | 2021-06-14T21:51:08 | {} | MEMBER |
ClickHouse | ClickHouse | 861,157,468 | 17,367 | alexey-milovidov | It works perfectly:
```
WITH arrayJoin(['a', 'b']) AS z
SELECT
z,
sumMergeForEach(x) AS x
FROM
(
SELECT sumStateForEach([1., 1.1, 1.1300175]) AS x
FROM remote('127.0.0.{1,2}', system.one)
)
GROUP BY z
``` | 2021-06-15T04:21:42 | 2021-06-15T04:21:42 | {} | MEMBER |
ClickHouse | ClickHouse | 860,287,442 | 20,277 | alexey-milovidov | We have two negative votes and two positive. | 2021-06-13T23:43:50 | 2021-06-13T23:44:08 | {} | MEMBER |
ClickHouse | ClickHouse | 860,353,088 | 24,427 | alexey-milovidov | This fix is very appreciated. | 2021-06-14T04:26:09 | 2021-06-14T04:26:09 | {} | MEMBER |
ClickHouse | ClickHouse | 860,353,300 | 25,067 | alexey-milovidov | I'm ok for this change if we bury the code as deep as possible... | 2021-06-14T04:26:53 | 2021-06-14T04:26:53 | {} | MEMBER |
ClickHouse | ClickHouse | 860,880,336 | 21,794 | alexey-milovidov | @vdimir
We don't need "least supertype" when doing JOIN.
What we actually need is the "most subtype". That always exists.
We already applying this method for `IN (subquery)`.
We are converting all the values with `accurateCase` to the most subtype and throwing off the values that are unconvertible.
Ask @kit... | 2021-06-14T17:59:48 | 2021-06-14T18:02:59 | {} | MEMBER |
ClickHouse | ClickHouse | 860,904,709 | 7,666 | alexey-milovidov | The place where the expression in SQL query is executed is not specified for distributed queries.
That's why I remove the "bug" label. | 2021-06-14T18:37:05 | 2021-06-14T18:37:05 | {} | MEMBER |
ClickHouse | ClickHouse | 860,904,706 | 25,267 | matwey | I am just trying to use gcc 11 for build. Every next gcc version is stricter about such things.
I remember that gcc 10 has found some similar missed `#include`. | 2021-06-14T18:37:05 | 2021-06-14T18:37:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 860,284,252 | 7,127 | alexey-milovidov | Let's keep it in mind... | 2021-06-13T23:18:34 | 2021-06-13T23:18:34 | {} | MEMBER |
ClickHouse | ClickHouse | 860,358,375 | 21,414 | alexey-milovidov | Need to merge with master. | 2021-06-14T04:45:12 | 2021-06-14T04:45:12 | {} | MEMBER |
ClickHouse | ClickHouse | 860,885,698 | 25,272 | UnamedRus | For me it looks like the same as https://github.com/ClickHouse/ClickHouse/pull/12515
Btw toTypeName as well as dumpColumnStructure disable this optimization for some reason.
```
SELECT toTypeName(transform(number, [2, 4, 6], CAST(['google', 'yandex', 'yahoo'], 'Array(Enum8(\'google\' = 1, \'other\' = 2, \'yahoo\... | 2021-06-14T18:07:50 | 2021-06-14T18:24:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 860,568,924 | 25,067 | azat | >Then let's make it support arbitrary number of components (simple std::vector)
I made VersionNumber for arbitrary numbers, and now I think it is OK to keep it in `Common/VersionNumber`. Objections? | 2021-06-14T10:13:34 | 2021-06-14T20:06:09 | {} | MEMBER |
ClickHouse | ClickHouse | 861,019,183 | 21,191 | alexey-milovidov | If you have no Mac, then you don't need this task. | 2021-06-14T21:50:47 | 2021-06-14T21:50:47 | {} | MEMBER |
ClickHouse | ClickHouse | 861,126,612 | 10,892 | yingshikong186 | > @yingshikong186 `alter table drop column` can be blocked for hours by a running merge.
thank you。I execute sql "ALTER TABLE XX ON CLUSTER YYY DELETE WHERE COLUMN = 'ZZZZ' "， but the column value is illegal | 2021-06-15T02:42:43 | 2021-06-15T02:42:43 | {} | NONE |
ClickHouse | ClickHouse | 861,155,180 | 19,027 | alexey-milovidov | Duplicate #18247 | 2021-06-15T04:14:54 | 2021-06-15T04:14:54 | {} | MEMBER |
ClickHouse | ClickHouse | 849,909,583 | 24,584 | qoega | SQLancer fail is insignificant:
>Cyclic aliases: default expression and column type are incompatible
I'll add new error message to skip list tomorrow | 2021-05-27T20:11:30 | 2021-05-27T20:11:30 | {} | MEMBER |
ClickHouse | ClickHouse | 850,443,118 | 24,605 | den-crane | `CST-8` Numeric TZ are not suppoted.
```
config.xml
<timezone>Asia/Shanghai</timezone>
restart
SELECT timeZone()
┌─timeZone()────┐
│ Asia/Shanghai │
└───────────────┘
``` | 2021-05-28T14:05:02 | 2021-05-28T14:05:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 850,922,046 | 24,749 | wangzhen11aaa | > try this
>
> ```
> cat /etc/clickhouse-server/users.d/min_bytes_to_use_mmap_io.xml
> <?xml version="1.0"?>
> <yandex>
> <profiles>
> <default>
> <min_bytes_to_use_mmap_io>0</min_bytes_to_use_mmap_io>
> </default>
> </profiles>
> </yandex>
> ```
>
> > 21.4.1.1
>
> ... | 2021-05-30T01:03:54 | 2021-05-30T01:03:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 849,910,980 | 24,399 | alexey-milovidov | @mergifyio update | 2021-05-27T20:13:55 | 2021-05-27T20:13:55 | {
"+1": 1
} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.