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 | 815,859,103 | 22,859 | alesapin | Also after this operation, all parts will have confusing names (`whatever_` instead of `all_`). | 2021-04-08T14:13:20 | 2021-04-08T14:13:20 | {} | MEMBER |
ClickHouse | ClickHouse | 815,909,263 | 22,856 | benbiti | oh，my mistake。 i update the amqp locally。3x @kssenii | 2021-04-08T15:18:32 | 2021-04-08T15:18:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,930,679 | 14,762 | den-crane | No info. | 2021-04-08T15:47:30 | 2021-04-08T15:47:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,019,539 | 22,763 | KochetovNicolai | ```
The query formatting is broken.
Changed settings: optimize_skip_unused_shards = '1', force_optimize_skip_unused_shards = '1', optimize_skip_unused_shards_nesting = '1', force_optimize_skip_unused_shards_nesting = '2', any_join_distinct_right_table_keys = '1', optimize_duplicate_order_by_and_distinct = '1'
Got th... | 2021-04-08T17:50:14 | 2021-04-08T17:50:14 | {} | MEMBER |
ClickHouse | ClickHouse | 815,066,920 | 17,708 | nrolf | I found this issue on ovio.org and would love to contribute! | 2021-04-07T16:50:24 | 2021-04-07T16:50:24 | {
"heart": 1
} | NONE |
ClickHouse | ClickHouse | 815,173,223 | 21,918 | ilejn | @vdimir could you clarify what went wrong, please.
Is it possible to have the thing in 20.8? May I assist? | 2021-04-07T19:36:21 | 2021-04-07T19:36:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,666,940 | 17,623 | dima898 | Hi. Any ideas on when to expect Window Functions to become stable and generally available? | 2021-04-08T10:59:50 | 2021-04-08T10:59:50 | {} | NONE |
ClickHouse | ClickHouse | 815,131,086 | 21,639 | buyology | Really nice work 👍
> ```shell
> cmake .. -DCMAKE_C_COMPILER=`brew --prefix llvm`/bin/clang -DCMAKE_CXX_COMPILER=`brew --prefix llvm`/bin/clang++ -DCMAKE_PREFIX_PATH=`brew --prefix llvm` -DENABLE_JEMALLOC=OFF
> ```
Adding a data point: ✅ Tested compiling with the above, and can confirm that I got it to work.
| 2021-04-07T18:29:48 | 2021-04-07T18:29:48 | {} | NONE |
ClickHouse | ClickHouse | 815,555,760 | 22,760 | cfroystad | As far as I can see, non of the test failures are related to the changes in this PR. Please correct me if I'm wrong 🙂 | 2021-04-08T08:16:00 | 2021-04-08T08:16:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,556,487 | 22,696 | alexey-milovidov | @den-crane See the `javaHashUTF16LE` function. | 2021-04-08T08:16:56 | 2021-04-08T08:16:56 | {} | MEMBER |
ClickHouse | ClickHouse | 815,571,105 | 21,888 | Avogar | @rf can you give an example that produce incorrect result? I tried to create such example, but I couldn't.
> This will produce incorrect deltasums when compared to using runningDifference against the raw data.
It seems like runningDifference can produce wrong result, because with each new block, the runningDiffer... | 2021-04-08T08:37:16 | 2021-04-08T08:41:27 | {} | MEMBER |
ClickHouse | ClickHouse | 816,025,861 | 21,888 | rf | Oh sure, it works fine when used outside of MVs. But when used in a MV:
```
create table default.metrics
(
timestamp DateTime Codec (DoubleDelta, LZ4),
name String,
value Float64
) ENGINE = MergeTree()
PARTITION BY toYYYYMM(timestamp)
ORDER BY (name, timestamp);
create table default.m... | 2021-04-08T18:00:42 | 2021-04-08T18:00:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,344,530 | 22,356 | akuzm | Somehow it still works badly for GCC, only 60% cache hits when it should be the same 99% as for clang. I'll leave it for the next time because I'm tired already. | 2021-04-07T23:53:33 | 2021-04-07T23:53:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,783,891 | 22,856 | alexey-milovidov | @benbiti I don't understand why do we need this change?
Everything is build and works correctly. | 2021-04-08T12:44:18 | 2021-04-08T12:44:38 | {} | MEMBER |
ClickHouse | ClickHouse | 815,800,485 | 22,856 | benbiti | @alexey-milovidov
in this commit https://github.com/ClickHouse-Extras/AMQP-CPP/commit/4546111cc5f3ee6a73f5aa7c82199d6fd4dfd3b3
file "deferredpublisher.cpp" has been changed to deferredrecall.cpp。
Maybe we should use "clickhouse" branch in gitmodule for AMQP-CPP too. | 2021-04-08T12:55:54 | 2021-04-08T12:59:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,898,702 | 21,639 | changvvb | Yes, I will merge it after you complete the job. | 2021-04-08T15:04:37 | 2021-04-08T15:04:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,113,917 | 22,765 | filimonov | I guess it's some other corner case for https://github.com/ClickHouse/ClickHouse/pull/14845 | 2021-04-07T18:02:37 | 2021-04-07T18:02:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,163,885 | 22,696 | den-crane | @EwainCai treats bytes as singed integers, so 128 became -128 and further summation is incorrect.
I suspect Clickhouse implementation of javaHash has the same issue but the opposite. Java treats bytes as singed. | 2021-04-07T19:19:59 | 2021-04-07T20:06:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,455,687 | 21,888 | rf | Ah, I was a little hasty in writing that example. Here is an example that is based on what I actually tested:
```
create table default.metrics_stat
(
bucket DateTime Codec (DoubleDelta, LZ4),
machine_id UUID,
name String,
s_avg AggregateFunction(avg, Float64),
s_count ... | 2021-04-08T05:22:18 | 2021-04-08T05:26:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,574,843 | 22,796 | alesapin | already fixed | 2021-04-08T08:42:41 | 2021-04-08T08:42:41 | {} | MEMBER |
ClickHouse | ClickHouse | 815,598,705 | 22,325 | vdimir | > It spams in err.log which is monitored by many setups.
>
> Especially important when you push logs to some external log monitoring system, where you pay per logs amount.
Ok, I'll backport it | 2021-04-08T09:16:03 | 2021-04-08T09:16:03 | {} | MEMBER |
ClickHouse | ClickHouse | 815,776,168 | 21,639 | changvvb | > do you experience any specific problem with it?
No. I just execute `./configure` at `contrib/openldap` and copy the generated files to `contrib/openldap-cmake`.
Maybe I lose some parameters of `./configure`, I guess. | 2021-04-08T12:39:04 | 2021-04-08T12:39:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,995,062 | 22,842 | alexey-milovidov | No, we don't close ports on startup. | 2021-04-08T17:12:44 | 2021-04-08T17:12:44 | {} | MEMBER |
ClickHouse | ClickHouse | 815,354,464 | 22,012 | nikitamikhaylov | Now we can write query like this (with two asterisk)
```sql
SELECT * FROM s3('https://s3.mds.yandex.net/clickhouse-test-reports/*/*/functional_stateless_tests_(ubsan)/test_results.tsv', 'xxx', 'yyy', 'LineAsString', 'line String') limit 100
```
or
```sql
SELECT * FROM s3Distributed('test_cluster_two_shards', '... | 2021-04-08T00:15:11 | 2021-04-08T00:15:11 | {} | MEMBER |
ClickHouse | ClickHouse | 815,012,195 | 19,766 | ghost | @vdimir
> Also order of arguments looks confusing, because we have base_cond between timestamp_column and event_column. Maybe (timestamp, event_column, base_cond, first_cond, second_cond, ...) will be better?
That's better, I will fix the order of arguments.
> Is it true that in most cases base_cond and first... | 2021-04-07T15:34:11 | 2021-04-08T05:55:56 | {} | NONE |
ClickHouse | ClickHouse | 815,574,482 | 22,800 | alesapin | For some reason, it's not required. | 2021-04-08T08:42:06 | 2021-04-08T08:42:06 | {} | MEMBER |
ClickHouse | ClickHouse | 815,814,444 | 22,763 | KochetovNicolai | queries
```
SELECT c0 + -1, sum(intDivOrZero(intDivOrZero(NULL, NULL), '2'), intDivOrZero(10000000000., intDivOrZero(intDivOrZero(intDivOrZero(NULL, NULL), 10), NULL))) FROM t_having GROUP BY c0 = 2, c0 = 10, intDivOrZero(intDivOrZero(intDivOrZero(NULL, NULL), NULL), NULL), c0 HAVING c0 = 2 SETTINGS enable_optimize_... | 2021-04-08T13:14:36 | 2021-04-08T13:15:26 | {} | MEMBER |
ClickHouse | ClickHouse | 815,883,092 | 22,834 | nvartolomei | If I understand correctly CH's memory profiler is working only on query level. If we could use it for global server profiling then this wouldn't be really needed. | 2021-04-08T14:45:23 | 2021-04-08T14:45:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,891,612 | 22,834 | akuzm | > @akuzm what do you think about performance results? Any reasons not to have profiling available in default build?
The main problem we had with always-on embedded profiler was unstable query run times, which lead to false positives. Maybe it's less bad for embedded memory profiling or profiling with `jemalloc`. I'l... | 2021-04-08T14:55:45 | 2021-04-08T14:55:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,904,363 | 22,834 | nvartolomei | @akuzm the goal is not to have always on profiling but instead have the jemalloc built with profiling support to allow end user to enable when needed without having to recompile clickhouse.
I didn't enable profiling in this PR, just changed the a build flag to include build jemalloc with profiling support. | 2021-04-08T15:12:15 | 2021-04-08T15:12:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,112,233 | 22,405 | Jirennor | @den-crane could you help me out with the follow up questions I posted above? | 2021-04-07T18:00:06 | 2021-04-07T18:00:06 | {} | NONE |
ClickHouse | ClickHouse | 815,532,520 | 22,763 | alexey-milovidov | Don't understand how this change breaks everything... | 2021-04-08T07:43:01 | 2021-04-08T07:43:01 | {} | MEMBER |
ClickHouse | ClickHouse | 815,547,578 | 22,706 | songenjie | @nikitamikhaylov @alexey-milovidov PTAL | 2021-04-08T08:04:23 | 2021-04-08T08:04:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,570,381 | 14,113 | alexey-milovidov | @alesapin It would be better to also put explanation and example in config.xml. | 2021-04-08T08:36:19 | 2021-04-08T08:36:19 | {} | MEMBER |
ClickHouse | ClickHouse | 815,594,260 | 14,113 | johnskopis | thanks for finally merging this one! | 2021-04-08T09:09:35 | 2021-04-08T09:09:35 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 815,640,032 | 21,639 | changvvb | @traceon I have merge master to this PR. As for ldap, maybe I need some hours to hack it. | 2021-04-08T10:15:58 | 2021-04-08T10:15:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,909,364 | 22,834 | akuzm | > I didn't enable profiling in this PR, just changed the a build flag to include build jemalloc with profiling support.
I don't see any changes in performance, so probably we can merge this. We would prefer people using the builtin profiler, but another one doesn't hurt. | 2021-04-08T15:18:40 | 2021-04-08T15:18:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,089,480 | 22,325 | filimonov | It spams in err.log which is monitored by many setups.
Especially important when you push logs to some external log monitoring system, where you pay per logs amount. | 2021-04-07T17:24:37 | 2021-04-07T17:24:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,110,254 | 22,727 | kitaisreal | Related to first issue, this is expected behaviour because dictionaries does not support nullable keys, and when nullable keys are passed into function we use default implementation.
Default implementation return null values if value in input column was null.
Related to second issues, I updated exceptions in scope of... | 2021-04-07T17:57:32 | 2021-04-07T17:57:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,141,242 | 22,405 | den-crane | @Jirennor I don't really know.
And TTL GROUP BY is not production ready.
I suspect the you are corrupting primary index if you clearing leading columns:
you had
21 , 1
21 , 2
22, 1
22, 2
if you do min(toStartOfHour(Timeslot)) for leading column, you get
0 , 1
0 , 2
0, 1
0, 2
Order of the rows ... | 2021-04-07T18:45:55 | 2021-04-07T18:45:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,538,350 | 22,823 | Avogar | > @Avogar Isn't it manually backported?
No, PR had pr-must-backport label but backport was created only yesterday, I didn't create manual backport. I will fix build failures. | 2021-04-08T07:51:27 | 2021-04-08T07:51:55 | {} | MEMBER |
ClickHouse | ClickHouse | 815,545,437 | 14,113 | alesapin | No related failures. Let's merge. | 2021-04-08T08:01:20 | 2021-04-08T08:01:20 | {} | MEMBER |
ClickHouse | ClickHouse | 815,881,729 | 22,834 | nvartolomei | @akuzm what do you think about performance results? Any reasons not to have profiling available in default build? | 2021-04-08T14:43:36 | 2021-04-08T14:43:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,886,530 | 21,639 | traceon | Let me prepare a new PR to your branch. | 2021-04-08T14:49:30 | 2021-04-08T14:49:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,911,613 | 21,639 | traceon | Great, so now AppleClang with Xcode or Ninja should work too, just as described in the docs. | 2021-04-08T15:21:40 | 2021-04-08T15:21:40 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,974,518 | 22,862 | alexey-milovidov | The simplest is just to disable `memcpy-bench` when shared linking is used. | 2021-04-08T16:41:22 | 2021-04-08T16:41:22 | {} | MEMBER |
ClickHouse | ClickHouse | 815,078,052 | 22,696 | EwainCai | > Though
>
> ```
> String str3 = "aaabbbbbcccc";
> System.out.println(str3.hashCode());
> --
> 1478625409
>
> ┌─javaHash('aaabbbbbcccc')─┐
> │ 1478625409 │
> └──────────────────────────┘
>
>
> String str3 = "hűtő";
> System.out.println(str3.hashCode());
> --
> 3456806
>
> ┌─javaHash... | 2021-04-07T17:07:02 | 2021-04-07T17:07:02 | {} | NONE |
ClickHouse | ClickHouse | 815,158,176 | 22,727 | UnamedRus | > Default implementation return null values if value in input column was null.
For me it's would be more convenient and expected if dictionary would return Default value in this case. (And dictGetOrNull would return Null)
Our client face this behavior while running mutations and they stuck because of Null to non-... | 2021-04-07T19:10:53 | 2021-04-07T19:10:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,200,501 | 22,826 | alesapin | Required for tmpfs. | 2021-04-07T20:01:08 | 2021-04-07T20:01:08 | {} | MEMBER |
ClickHouse | ClickHouse | 815,494,709 | 22,765 | hexiaoting | It seems work on fresh master.
```
dell123 :) select version();
SELECT version()
Query id: dac70684-3000-4e81-a4ae-b5d22918c7d7
┌─version()─┐
│ 21.5.1.1 │
└───────────┘
1 rows in set. Elapsed: 0.015 sec.
dell123 :) CREATE TABLE xxx_arr
:-] (
:-] `key` UInt32,
:-] `arr` ALIAS [1, 2],
:-] ... | 2021-04-08T06:43:40 | 2021-04-08T06:43:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,533,603 | 22,826 | alexey-milovidov | > SQLancer test — SQLancer test run. See report
Non-obvious report, @qoega. | 2021-04-08T07:44:37 | 2021-04-08T07:44:37 | {} | MEMBER |
ClickHouse | ClickHouse | 815,865,453 | 21,639 | traceon | Ok, so then let's use the copy-pasted one (I reverted to it in https://github.com/changvvb/ClickHouse/pull/1 ) - it works for me, everything compiles.
I'll try to separately look into GCC build (Clang and AppleClang even with Xcode IDE ~~after some zlib-ng patch that hopefully will be merged soon~~ workaround merged... | 2021-04-08T14:21:34 | 2021-04-08T14:40:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,082,754 | 22,696 | alexey-milovidov | Also check unicode normalization/decomposition. | 2021-04-07T17:14:47 | 2021-04-07T17:14:47 | {} | MEMBER |
ClickHouse | ClickHouse | 815,120,231 | 22,765 | filimonov | /cc @alesapin | 2021-04-07T18:12:47 | 2021-04-07T18:12:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,469,052 | 21,058 | windylcx | > Try to add `/usr/local/opt/llvm/bin/` into your PATH
i try this,doesn't work | 2021-04-08T05:53:14 | 2021-04-08T05:53:14 | {} | NONE |
ClickHouse | ClickHouse | 815,538,391 | 22,619 | alexey-milovidov | @olgarev @KochetovNicolai You forgot about the bonus:
https://github.com/ClickHouse/ClickHouse/commit/ac3de6626af6772297c2f5df653854b98635605d | 2021-04-08T07:51:30 | 2021-04-08T07:51:30 | {} | MEMBER |
ClickHouse | ClickHouse | 815,976,246 | 22,862 | alexey-milovidov | PS. The usage of `memcpy-bench` is non-obvious, and it's very easy to drive misleading conclusions. Maybe I have to write a blog post about it, it's really complicated. | 2021-04-08T16:44:13 | 2021-04-08T16:44:13 | {} | MEMBER |
ClickHouse | ClickHouse | 816,025,492 | 22,763 | alexey-milovidov | This is what @akuzm is going to fix in #22703 (not a bug) | 2021-04-08T18:00:08 | 2021-04-08T18:00:18 | {} | MEMBER |
ClickHouse | ClickHouse | 815,051,886 | 22,619 | olgarev | @KochetovNicolai Thanks! | 2021-04-07T16:28:14 | 2021-04-07T16:28:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,244,796 | 21,639 | mishfish | @buyology @changvvb
Thank you a lot, I finally compile it:)
Sorry for stupid questions:) | 2021-04-07T20:36:09 | 2021-04-07T20:36:09 | {
"hooray": 2
} | NONE |
ClickHouse | ClickHouse | 815,488,363 | 22,828 | hexiaoting | So create table with `order by id` is only used for storage , for index, It has nothing to do with select query, right?
| 2021-04-08T06:31:43 | 2021-04-08T06:32:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,557,474 | 16,798 | zhanglistar | > @litaoyx, thanks for sharing your solution. I read through your blog and was able to add RoaringBitmap support. One thing I want to clarify is that we don't need a base64 encoded string, since we can insert serialized value into AggregateFunction column directly. I guess the error you met earlier is related to the `T... | 2021-04-08T08:18:16 | 2021-04-08T08:18:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,601,373 | 21,918 | vdimir | > @vdimir could you clarify what went wrong, please.
> Is it possible to have the thing in 20.8? May I assist?
We are highly recommend to update to 21.3.
But since it's backported to 21.1-21.3, I'll backport to 20.8 too. | 2021-04-08T09:19:54 | 2021-04-08T09:19:54 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 815,411,963 | 22,706 | songenjie | > it's ready.
3ks, it's ready | 2021-04-08T03:07:47 | 2021-04-08T03:07:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,445,740 | 22,828 | ucasfl | It's not a bug, because the order of returning blocks is uncentain. Only if with `order by id` in select sql, it will always return same result. | 2021-04-08T04:56:20 | 2021-04-08T04:56:44 | {
"+1": 1
} | COLLABORATOR |
ClickHouse | ClickHouse | 815,573,727 | 22,698 | alesapin | No new failures | 2021-04-08T08:41:00 | 2021-04-08T08:41:00 | {} | MEMBER |
ClickHouse | ClickHouse | 815,614,421 | 21,107 | petuhovskiy | Status update:
1. AggregatingMemory is wokring now, but Aggregator state is not shared between inserts, and it is working the same way the MaterializedView does.
2. AggregatingMemory is currently a special engine, which accepts non-aggregated rows, executes SELECT GROUP BY on them, and inserts aggregated data to a me... | 2021-04-08T09:38:33 | 2021-04-08T09:38:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,646,141 | 22,841 | alexey-milovidov | > P.S: It seems that the watchdog process has a typo in its name.
It's intentional.
```
setThreadName("clckhouse-watch"); /// 15 characters
```
As alternative, we can change to `clickhouse-wtch` if you like. But then it can be confused with the word "witch". | 2021-04-08T10:25:30 | 2021-04-08T10:28:56 | {} | MEMBER |
ClickHouse | ClickHouse | 815,052,355 | 21,888 | Avogar | @rf I've just noticed, are you sure that you use materialized view with ENGINE=AggregatingMergeTree() correctly? You wrote such query:
```
create materialized view metrics_test_1m
ENGINE=AggregatingMergeTree()
PARTITION BY toYYYYMM(bucket) ORDER BY (name, bucket)
POPULATE
AS SELECT toStartOfMinute(timestamp) as b... | 2021-04-07T16:28:53 | 2021-04-07T16:33:32 | {} | MEMBER |
ClickHouse | ClickHouse | 815,098,570 | 22,490 | filimonov | BTW maybe your English / translate is just too poor. Execute vs 'have executed'?
And answer is still the same. Every replica has its own queue, feed from the main queue. Other replicas will have that task in queue and will be not able to execute that. Task will disappear from replica queue once it will be executed o... | 2021-04-07T17:38:55 | 2021-04-07T17:45:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,575,441 | 12,656 | alesapin | Hi, @zhang2014. Do you have any plans to continue on this? Maybe we can close it temporarily? | 2021-04-08T08:43:28 | 2021-04-08T08:43:28 | {} | MEMBER |
ClickHouse | ClickHouse | 815,895,500 | 22,834 | alexey-milovidov | > If I understand correctly CH's memory profiler is working only on query level.
We have global profiler as well, see `memory_profiler_*` in config.xml. | 2021-04-08T15:00:34 | 2021-04-08T15:00:34 | {} | MEMBER |
ClickHouse | ClickHouse | 815,904,868 | 21,639 | traceon | @changvvb could you please merge this: https://github.com/changvvb/ClickHouse/pull/2 | 2021-04-08T15:12:56 | 2021-04-08T15:12:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,078,153 | 22,299 | alexey-milovidov | That's awesome! | 2021-04-07T17:07:13 | 2021-04-07T17:07:13 | {} | MEMBER |
ClickHouse | ClickHouse | 815,153,834 | 21,639 | mishfish | Unfortunately still cannot build it:)
This time I get
`[2476/5628] Building ASM object contrib/boost-cmake/CMakeFiles/_boost_context.dir/__/boost/libs/context/src/asm/jump_arm64_aapcs_elf_gas.S.o
FAILED: contrib/boost-cmake/CMakeFiles/_boost_context.dir/__/boost/libs/context/src/asm/jump_arm64_aapcs_elf_gas.S.o
... | 2021-04-07T19:03:58 | 2021-04-07T19:03:58 | {} | NONE |
ClickHouse | ClickHouse | 815,283,279 | 22,514 | gyuton | Internal documentation ticket: DOCSUP-8347. | 2021-04-07T21:42:13 | 2021-04-07T21:42:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,423,330 | 21,574 | marising | @Sallery-X
Okay, I will sort out the code and test cases, and then submit a PR.
| 2021-04-08T03:43:53 | 2021-04-08T03:43:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,534,922 | 22,823 | alexey-milovidov | @Avogar Isn't it manually backported? | 2021-04-08T07:46:27 | 2021-04-08T07:46:27 | {} | MEMBER |
ClickHouse | ClickHouse | 815,550,982 | 21,639 | traceon | @changvvb This commit doesn't look good: https://github.com/ClickHouse/ClickHouse/pull/21639/commits/21a2eb32f26496692cef118c8a04a0e20fbc1194
It doesn't follow the same feature check pattern as for other platforms. | 2021-04-08T08:09:09 | 2021-04-08T08:09:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,835,033 | 22,843 | den-crane | It's the expected behavior.
Each column use 2 MB RAM on Insert. 5000*2 = 10GB.
Clickhouse documentation suggests to have up to 1000 columns | 2021-04-08T13:42:09 | 2021-04-08T13:42:09 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,166,397 | 21,639 | buyology | > Unfortunately still cannot build it:)
>
> This time I get
>
> `
> […] ../contrib/boost/libs/context/src/asm/make_arm64_aapcs_elf_gas.S:58:1: error: unknown directive .type make_fcontext, %function ^ ../contrib/boost/libs/context/src/asm/make_arm64_aapcs_elf_gas.S:83:1: error: unknown directive .size make_fcont... | 2021-04-07T19:24:27 | 2021-04-07T19:24:27 | {} | NONE |
ClickHouse | ClickHouse | 815,187,125 | 21,639 | mishfish | > > Unfortunately still cannot build it:)
> > This time I get
> > `[…] ../contrib/boost/libs/context/src/asm/make_arm64_aapcs_elf_gas.S:58:1: error: unknown directive .type make_fcontext, %function ^ ../contrib/boost/libs/context/src/asm/make_arm64_aapcs_elf_gas.S:83:1: error: unknown directive .size make_fcontext,.-... | 2021-04-07T19:54:41 | 2021-04-07T19:54:41 | {} | NONE |
ClickHouse | ClickHouse | 815,468,211 | 21,058 | windylcx | > @windylcx hi,I have also this question #21867 ，Excuse me, did you solve the problem
sorry,not yet | 2021-04-08T05:51:20 | 2021-04-08T05:51:20 | {} | NONE |
ClickHouse | ClickHouse | 815,517,135 | 22,809 | alexey-milovidov | The fix is relevant but it's minor and interfere with other tests. | 2021-04-08T07:20:20 | 2021-04-08T07:20:20 | {} | MEMBER |
ClickHouse | ClickHouse | 815,531,496 | 22,821 | alexey-milovidov | Looks like the behaviour did not change, only exception messages and minor refactoring. | 2021-04-08T07:41:33 | 2021-04-08T07:41:33 | {} | MEMBER |
ClickHouse | ClickHouse | 815,562,074 | 22,827 | alexey-milovidov | In your examples you have Unicode string literals with some characters.
How Unicode string are mapped to bytes is dependend on the charset (UTF-8, UTF-16, UTF-32, UCS-2...),
which is unspecified in your example.
Also Unicode string can be represented by different sequences of code points, examples:
`ё` and `е... | 2021-04-08T08:24:50 | 2021-04-08T08:25:00 | {} | MEMBER |
ClickHouse | ClickHouse | 815,642,973 | 22,824 | alexey-milovidov | СС @Enmk | 2021-04-08T10:20:28 | 2021-04-08T10:20:28 | {} | MEMBER |
ClickHouse | ClickHouse | 815,878,534 | 22,856 | kssenii | > in this commit ClickHouse-Extras/AMQP-CPP@4546111
file "deferredpublisher.cpp" has been changed to deferredrecall.cpp。
But submodule was not updated in master up to this commit. It is still the version before this commit.
| 2021-04-08T14:39:19 | 2021-04-08T14:39:19 | {} | MEMBER |
ClickHouse | ClickHouse | 815,885,880 | 21,639 | traceon | No, I am trying to commit those changes in your branch actually, so they will be reverted there. If you could merge my changes, that would be great. | 2021-04-08T14:48:39 | 2021-04-08T14:48:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,996,215 | 22,863 | alexey-milovidov | Also need a test. | 2021-04-08T17:14:29 | 2021-04-08T17:14:29 | {} | MEMBER |
ClickHouse | ClickHouse | 815,053,042 | 20,539 | myrrc | A word about the IPC between the tester script and the server -- currently we have 3 main options:
1) Utilizing the tester's --testname option which executes a SELECT testname query before running the test.
The issue here is that we need to interfere with the query process pipeline -- either write a hook or handle th... | 2021-04-07T16:29:54 | 2021-04-07T16:29:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,606,285 | 21,639 | traceon | @changvvb could you please sync with the master? | 2021-04-08T09:27:03 | 2021-04-08T09:27:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,647,703 | 22,841 | alexey-milovidov | Ok. The usage of transparent huge pages is controversial.
- on old Linux kernel it leads to significant performance drop, but only after some runtime under load;
- in newer Linux kernel user reports that it improves performance without drawbacks;
- and you have added another consideration...
Let's edit the docs. | 2021-04-08T10:27:57 | 2021-04-08T10:27:57 | {} | MEMBER |
ClickHouse | ClickHouse | 815,659,168 | 21,639 | traceon | @changvvb do you experience any specific problem with it? I believe, the same set of config header files should work for both x86_64 and arm64. | 2021-04-08T10:46:55 | 2021-04-08T10:46:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,881,086 | 21,639 | changvvb | > Ok, so then let's use the copy-pasted one (I reverted to it in [changvvb#1](https://github.com/changvvb/ClickHouse/pull/1) ) - it works for me, everything compiles.
OK, should I also revert that commit in this PR? (reverted) | 2021-04-08T14:42:45 | 2021-04-08T14:49:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,051,222 | 22,843 | filimonov | Duplicate of https://github.com/ClickHouse/ClickHouse/issues/6943 | 2021-04-08T18:38:15 | 2021-04-08T18:38:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 811,622,148 | 22,424 | ucasfl | Sorry, this a mistake. | 2021-04-01T04:05:04 | 2021-04-01T04:05:04 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 811,641,867 | 21,781 | alexey-milovidov | Also it's possible to extract the relevant data from `system.query_log`.
It has `tables` field, `ProfileEvents` and `query_kind`. | 2021-04-01T05:08:46 | 2021-04-01T05:08:46 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 811,642,042 | 19,766 | ghost | @vdimir
I modified the core algorithm because the core algorithm had the [exceptional case](https://github.com/ClickHouse/ClickHouse/pull/19766#issuecomment-795175943).
The changes from the last version:
- Deleted an algorithm like `Boyer-Moore`. The current version of the algorithm uses just simple for-loop in... | 2021-04-01T05:09:15 | 2021-04-01T05:15:17 | {
"rocket": 1
} | NONE |
ClickHouse | ClickHouse | 811,758,964 | 21,437 | alexey-milovidov | @tangxueming It is highly recommended to upgrade to version 21.3 LTS. | 2021-04-01T08:52:05 | 2021-04-01T08:52:05 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.