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 | 698,732,460 | 14,002 | alexey-milovidov | You have to install `llvm-objcopy`, it should be used instead of `objcopy`. | 2020-09-25T05:44:04 | 2020-09-25T05:44:14 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 698,764,531 | 15,063 | alesapin | @azat `01370_client_autocomplete_word_break_characters` failure seems to be related to changes? | 2020-09-25T07:12:20 | 2020-09-25T07:12:20 | {} | MEMBER |
ClickHouse | ClickHouse | 698,971,359 | 11,407 | den-crane | Wrong syntax. CH does not support ` .... in Array`
```
with (select groupArray(id) from bbb) as ids select * from aaa where id in ids;
DB::Exception: Types of column 1 in section IN don't match: UInt16 on the left, Array(UInt16) on the right.
```
20.10.1
```
with (select groupArray(id) from bbb) as id... | 2020-09-25T14:45:20 | 2020-09-25T14:45:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 698,979,293 | 15,301 | alexey-milovidov | Or: install `libc6-dbg` then decode the address `0x00007ffff7a48960`. | 2020-09-25T14:59:24 | 2020-09-25T14:59:24 | {} | MEMBER |
ClickHouse | ClickHouse | 698,707,921 | 15,266 | alexey-milovidov | Waiting only builds. | 2020-09-25T04:05:04 | 2020-09-25T04:05:04 | {} | MEMBER |
ClickHouse | ClickHouse | 698,753,038 | 15,271 | ZonghaoWang | is here?
I just think it can accerate the groupAgg speed simply, but I donot benchmark them | 2020-09-25T06:42:42 | 2020-09-25T06:42:42 | {} | NONE |
ClickHouse | ClickHouse | 698,816,098 | 15,272 | hexiaoting | Yes, the error still exists.
My gcc version:
```
c++ (Gentoo 10.2.0-r1 p2) 10.2.0
``` | 2020-09-25T09:06:22 | 2020-09-25T09:06:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 698,858,654 | 15,242 | 4ertus2 | It's something really strange with test. | 2020-09-25T10:40:46 | 2020-09-25T10:40:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 698,693,807 | 15,191 | fastio | > Seems like test failures related to changes.
@alesapin Thanks. I want to run the integration tests in my test env. Does it must run on Ubuntu ?
> Running natively
>Prerequisites:
>Ubuntu 14.04 (Trusty) or higher.
>docker. Minimum required API version: 1.25, check with docker version.
https://github.co... | 2020-09-25T03:03:05 | 2020-09-25T03:03:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 698,707,978 | 15,266 | alexey-milovidov | Yes, it started to work. | 2020-09-25T04:05:19 | 2020-09-25T04:05:19 | {} | MEMBER |
ClickHouse | ClickHouse | 698,880,036 | 15,229 | simPod | I dropped the buffer table in order to tweak it's min and max conditions and it has all got processed 🤔 Thanks for help, will definitely try to increase the pool size when similar issue happens. | 2020-09-25T11:35:34 | 2020-09-25T11:35:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 698,956,797 | 15,278 | den-crane | There is no easy way to change this list `(value_a,value_b)`.
Usually it is excessive section and you can create just SummingMergeTree() then Summing will sum all columns except columns in ORDERBY list.
Options how to change it:
1. detach table XXX;
vi /var/lib/clickhouse/metadata/{db}/{table}.sql # edit file... | 2020-09-25T14:19:18 | 2020-09-25T14:19:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 698,969,607 | 15,303 | alexey-milovidov | There are some files that we don't compress, like `.idx`, `.mrk`.
But we still calculate hash of the data in these files, accumulate it and present in the `hash_of_uncompressed_files` field.
The purpose is the same as for other hashes - to compare them to find possible difference. | 2020-09-25T14:42:14 | 2020-09-25T14:42:14 | {} | MEMBER |
ClickHouse | ClickHouse | 698,978,262 | 15,301 | alexey-milovidov | `#11 0x00007ffff7a48960 in ?? ()`
It can be a bug in `glibc` - some functions do not have proper call frame information.
Could you please reproduce this issue with `libc6-dbg` installed? | 2020-09-25T14:57:27 | 2020-09-25T14:58:43 | {} | MEMBER |
ClickHouse | ClickHouse | 699,046,135 | 15,300 | alesapin | test_multiple_disks Flaky? Restarted tests. | 2020-09-25T17:08:36 | 2020-09-25T17:08:36 | {} | MEMBER |
ClickHouse | ClickHouse | 699,064,422 | 15,144 | azat | >https://clickhouse-test-reports.s3.yandex.net/15144/26abe8cb30819eea1c1f3383cc99620ab2e9da9a/performance_comparison/report.html#fail1
>disabling percpu arenas doesn't really influence the performance
In general it should not, the purpose of using percpu arena (#11084) is to avoid live leaks with big thread pool, s... | 2020-09-25T17:46:29 | 2020-09-25T17:46:29 | {} | MEMBER |
ClickHouse | ClickHouse | 699,136,793 | 15,315 | andrey-potyomkin | > What is it? Such syntax is not supported.
Oh, sorry. I checked again SQL query from my previous message and it's written with mistake. In my original SQL used syntax groupArray((date_time, column_1, column_2, column_3)) and it's working.
```
SELECT groupArray((date, column_1, column_2, column_3))
FROM db.tes... | 2020-09-25T20:22:29 | 2020-09-25T20:22:29 | {} | NONE |
ClickHouse | ClickHouse | 698,334,514 | 15,174 | alexey-milovidov | @l1t1
Here is the benchmark:
https://clickhouse.tech/benchmark/hardware/
AWS `a...`, `c6`, `m6` and `Huawei` are AArch64 servers.
There are multiple variants of server AArch64 CPU, this is incomplete list:
- Amazon Graviton 1 and 2;
- Huawei Taishan;
- Cavium ThunderX 1, 2;
- APM X-Gene;
The results w... | 2020-09-24T13:11:12 | 2020-09-25T01:41:11 | {} | MEMBER |
ClickHouse | ClickHouse | 698,818,441 | 15,272 | hexiaoting | @alexey-milovidov if this is a bug, I want to try to fix it. | 2020-09-25T09:11:11 | 2020-09-25T09:11:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 698,849,733 | 5,183 | John-belt | Same here tons of these messages, running 20.6.3. what can be possible reason for it??? and can we move it to trace logs. | 2020-09-25T10:18:45 | 2020-09-25T10:18:45 | {} | NONE |
ClickHouse | ClickHouse | 698,963,214 | 15,294 | den-crane | duplicate https://github.com/ClickHouse/ClickHouse/issues/11621
already fixed: https://github.com/ClickHouse/ClickHouse/pull/11767 | 2020-09-25T14:30:56 | 2020-09-25T14:30:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 699,035,113 | 15,235 | nikitamikhaylov | I've added some tests. Everything works fine with existing partition key. Will try to make copier work with absent partition key in destination table. (Or will add human readable error.) | 2020-09-25T16:46:42 | 2020-09-25T16:46:42 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 699,079,355 | 15,007 | azat | > Integration tests (asan) — No tests found for some reason!
By some reason it uses version of some scripts from upstream/master (I guess not all images are versioned?)
```
conftest.py:1: in <module>
from helpers.test_tools import TSV
/usr/lib/python3.6/ast.py:35: in parse
return compile(source, filen... | 2020-09-25T18:16:39 | 2020-09-25T18:28:47 | {} | MEMBER |
ClickHouse | ClickHouse | 698,695,700 | 15,257 | alexey-milovidov | E.g. a preimage of a point for `toDate` function is a whole day. | 2020-09-25T03:10:32 | 2020-09-25T03:10:32 | {} | MEMBER |
ClickHouse | ClickHouse | 698,816,861 | 15,283 | akuzm | I tried with today's master and it still fails, so probably different. It's simple to check locally if you run `clickhouse-client --query-fuzzer-runs=1000` and enter the failing query.
```
base before fuzz: SELECT * REPLACE(i + 1048575 AS i) APPLY(sum) FROM columns_transformers
base after fuzz: SELECT * REPLACE(i ... | 2020-09-25T09:07:54 | 2020-09-25T09:07:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 698,628,721 | 13,997 | UnamedRus | Ok,
You sounds reasonable.
> It's the way how push down predicate works now.
So conditions in WHERE is being executed only once?
One more optimization? that related to JOIN (or multiple joins)
SELECT key, col_b, col_c FROM table(key, col_b, col_c) INNER JOIN table_2(key, col_e) ON key WHERE col_e = 'value... | 2020-09-24T22:53:41 | 2020-09-25T09:14:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 698,976,284 | 15,303 | alexey-milovidov | It is computed with `HashingWriteBuffer`, non-cryptographic hash function `CityHash_v1_0_2::CityHash128WithSeed` is used. | 2020-09-25T14:53:56 | 2020-09-25T14:53:56 | {} | MEMBER |
ClickHouse | ClickHouse | 698,670,216 | 14,476 | alexey-milovidov | ```
milovidov-desktop :) SELECT (SELECT groupArray(a) FROM (SELECT [1, 2] AS a)) AS aaa
SELECT
(
SELECT groupArray(a)
FROM
(
SELECT [1, 2] AS a
)
) AS aaa
┌─aaa─────┐
│ [[1,2]] │
└─────────┘
1 rows in set. Elapsed: 0.017 sec.
milovidov-desktop... | 2020-09-25T01:27:24 | 2020-09-25T01:27:24 | {} | MEMBER |
ClickHouse | ClickHouse | 698,685,715 | 15,257 | amosbird | Rewritting those predicates seems too specific. Perhaps we can introduce dedicated functions such as isDate, isStartOfHour, etc. | 2020-09-25T02:31:56 | 2020-09-25T02:31:56 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 698,766,178 | 15,110 | alesapin | All failed tests not related to changes.
`Yandex synchronization check` -- suspicious, but should be already fixed in `arcadia`. | 2020-09-25T07:16:23 | 2020-09-25T07:16:23 | {} | MEMBER |
ClickHouse | ClickHouse | 698,800,775 | 15,272 | ZonghaoWang | gcc-10 also error at here | 2020-09-25T08:34:16 | 2020-09-25T08:34:16 | {} | NONE |
ClickHouse | ClickHouse | 698,874,253 | 11,535 | c3mb0 | this is still present in 20.9
a single alias in with works fine, 2+ aliases give the unknown column error | 2020-09-25T11:20:36 | 2020-09-25T11:20:36 | {} | NONE |
ClickHouse | ClickHouse | 698,950,146 | 15,296 | den-crane | The problem here is how to change SummingMergeTree in **old syntax format**. You сannot to alter it if you need to add new DIMENSION column, you can add only METRIC(measure)
```
detach table consumer; -- at this moment consume has stopped.
drop table queue;
create table queue (.....) ENGINE = Kafka;
alter... | 2020-09-25T14:07:39 | 2020-09-25T14:13:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 699,082,465 | 15,317 | alexey-milovidov | @detailyang You can also generate the list of columns by
```
EXPLAIN SYNTAX SELECT COLUMNS('tags_') FROM table
```
And then copy-paste the generated list. | 2020-09-25T18:23:07 | 2020-09-25T18:23:07 | {} | MEMBER |
ClickHouse | ClickHouse | 698,691,182 | 15,066 | flyfoxCI | > 1. Do see any difference if you add `identity` (for a test only) ` toYYYYMMDDhhmmss(identity(date_time))=20200920214000`
there is no difference when add identity | 2020-09-25T02:52:48 | 2020-09-25T02:52:48 | {} | NONE |
ClickHouse | ClickHouse | 698,808,435 | 15,283 | amosbird | Heh, I've mis-cloned AST twice... | 2020-09-25T08:50:44 | 2020-09-25T08:50:44 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 698,846,126 | 15,284 | alesapin | Fixed in #15258 | 2020-09-25T10:10:07 | 2020-09-25T10:10:07 | {} | MEMBER |
ClickHouse | ClickHouse | 698,852,588 | 14,414 | tavplubix | MemorySanitizer - use-of-uninitialized-value when logging an exception
https://clickhouse-test-reports.s3.yandex.net/15003/fb826557b7ce3a3015fe00df78609ed42d9c5412/stress_test_(memory)/stderr.log | 2020-09-25T10:25:59 | 2020-09-25T10:25:59 | {} | MEMBER |
ClickHouse | ClickHouse | 698,892,054 | 15,243 | tavplubix | Functional stateless tests (release) — Timeout - known issue in CI infrastructure
AST fuzzer - Fixed in #15302
Functional stateless tests (release, DatabaseAtomic) — fail: 1 - wrong skip_list, already fixed
Stress test (memory) - use-of-uninitialized-value - known issue, not related
Stress test (thread) - data race... | 2020-09-25T12:07:42 | 2020-09-25T12:07:42 | {} | MEMBER |
ClickHouse | ClickHouse | 698,964,912 | 15,270 | den-crane | @BernieJiangDL for now clickhouse touches every part for any mutation (alter update/delete).
This behavior is by design. It needs to do consistent mutations.
Mutations has been implemented for rare admin operations (twice a year) not for daily business tasks. | 2020-09-25T14:33:52 | 2020-09-25T14:37:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 698,981,231 | 14,403 | shantam04 | @den-crane Sorry for the late response, I checked both of the queries and all our inactive parts have refcount=1 and the mutation count is zero.
This looks to me that part removal itself is slow is there a way to speed this up.
Also, I checked by when are parts getting removed ad found the removal time for the old pa... | 2020-09-25T15:02:54 | 2020-09-25T15:21:31 | {} | NONE |
ClickHouse | ClickHouse | 699,016,636 | 15,173 | simo-lineate | @den-crane Thanks for the update. You're suggesting to upgrade to 20.7+. However, 2.3 is marked as the current LTS in your docs. Are the docs a little behind? My team is trying to update to the most stable version in production, not necessarily the latest version. What version would you suggest to upgrade to as of this... | 2020-09-25T16:08:46 | 2020-09-25T16:08:46 | {} | NONE |
ClickHouse | ClickHouse | 699,084,076 | 15,314 | alexey-milovidov | It's better to make ORDER BY `organization_id`, `request_time`, not PARTITION BY. | 2020-09-25T18:26:42 | 2020-09-25T18:26:42 | {} | MEMBER |
ClickHouse | ClickHouse | 699,287,476 | 15,317 | detailyang | @alexey-milovidov thanks. Gotcha, I will try the SQL or implement COLUMNS in ORDER BY :) | 2020-09-26T03:15:43 | 2020-09-26T03:15:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 698,689,082 | 15,257 | alexey-milovidov | Generic solution is - function inversion.
We should introduce the following methods:
```
bool IFunctionBase::hasInformationAboutPreimage() const;
Range IFunctionBase::getPreimageForRange(const IDataType & /*type*/, const Field & /*left*/, const Field & /*right*/) const;
```
Then we will be able to rewrite exp... | 2020-09-25T02:45:10 | 2020-09-25T02:46:57 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 698,691,967 | 15,257 | amosbird | > Then we will be able to rewrite expressions with comparisons of f(x) with constants to comparisons of x with another constants.
But toDate and toStartOfHour is noninversive. | 2020-09-25T02:56:07 | 2020-09-25T02:56:07 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 698,760,231 | 15,244 | qoega | ASTFuzzer not related
https://github.com/ClickHouse/ClickHouse/issues/15283
ClickHouse Stress Test (thread) known issue with openssl. Not related
ClickHouse Stress Test (memory) not related
https://github.com/ClickHouse/ClickHouse/issues/15284 | 2020-09-25T07:01:20 | 2020-09-25T07:01:20 | {} | MEMBER |
ClickHouse | ClickHouse | 698,834,840 | 15,295 | winter7 | The reason I ask this question is that I want to use Kafka Engine. I wonder if I should temporarily add some ClickHouse instances with kafka engine table as consumers when Kafka traffic surges and convert data by materialized View to the original (remote) Clickhouse instances. Is this a good way to deal with the sudden... | 2020-09-25T09:45:50 | 2020-09-25T09:45:50 | {} | NONE |
ClickHouse | ClickHouse | 698,961,869 | 15,294 | den-crane | ```
20.10.1.4704
┌─column─┬─d─────────┐
│ test │ [1,2,3,4] │
└────────┴───────────┘
``` | 2020-09-25T14:28:29 | 2020-09-25T14:28:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 699,048,067 | 15,286 | alexey-milovidov | Waiting only for builds. | 2020-09-25T17:12:35 | 2020-09-25T17:12:35 | {} | MEMBER |
ClickHouse | ClickHouse | 699,125,936 | 15,320 | 4ertus2 | I've tried to make plus and minus operations inplace over lhs. It does not improve performance even for multiply operation where we have several `res = plus(res, some)`. Also I've tried to specialize operations with small (<= 8 bytes) integers that would have 1 element to plus/minus and simple overflow case. It also ha... | 2020-09-25T19:57:51 | 2020-09-25T20:02:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 698,670,551 | 15,211 | chang0391 | @zhang2014
Good idea, thanks for your effort, and hope this function can be realized soon | 2020-09-25T01:28:46 | 2020-09-25T01:46:05 | {} | NONE |
ClickHouse | ClickHouse | 698,718,276 | 15,270 | filimonov | 2020071011_0_55_8_23697
2020071011_0_55_8_23698
Last digit changed. That means some mutation (alter update/delete/modify column etc) was applied to a part.
When mutating data clickhouse touches every part in the table (even if it was noop mutation for that part).
Regarding loading time - what are your observa... | 2020-09-25T04:50:08 | 2020-09-25T04:50:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 698,761,961 | 14,414 | qoega | MemorySanitizer src/Common/hex.h:57:9 in void writeHexUIntImpl<unsigned long>
https://github.com/ClickHouse/ClickHouse/issues/15284 | 2020-09-25T07:05:47 | 2020-09-25T07:06:03 | {} | MEMBER |
ClickHouse | ClickHouse | 698,763,144 | 15,220 | alesapin | Test failure not related to changes (fixing in master). | 2020-09-25T07:08:59 | 2020-09-25T07:08:59 | {} | MEMBER |
ClickHouse | ClickHouse | 698,774,811 | 15,257 | UnamedRus | intDiv can be rewritten in that manner too, so it can be done for all monotonic functions? | 2020-09-25T07:37:03 | 2020-09-25T07:37:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 698,820,400 | 15,272 | ZonghaoWang | > Yes, the error still exists.
> My gcc version:
>
> ```
> c++ (Gentoo 10.2.0-r1 p2) 10.2.0
> ```
the error file in test dir, I think you may be more interest at `ninja clickhouse` | 2020-09-25T09:15:16 | 2020-09-25T09:15:16 | {} | NONE |
ClickHouse | ClickHouse | 698,853,378 | 14,414 | tavplubix | > Stress tests ThreadSanitizer: data race /build/obj-x86_64-linux-gnu/../src/Core/SettingsFields.h:43:52 in DB::SettingFieldNumber::operator=(unsigned long) https://clickhouse-test-reports.s3.yandex.net/14265/494e59e7c12e0887a32b2564463a84aea500118b/stress_test_(thread)/stderr.log
Fixed in #15243 | 2020-09-25T10:27:52 | 2020-09-25T10:27:52 | {} | MEMBER |
ClickHouse | ClickHouse | 699,034,612 | 15,315 | den-crane | MaterializedView is an insert trigger.
It gets data from **INSERT**.
It **never** reads (selects) from table.
>groupArray(date_time, column_1, column_2, column_3)
What is it? Such syntax is not supported. Can you show `desc data` | 2020-09-25T16:45:38 | 2020-09-25T16:46:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 699,067,842 | 15,317 | detailyang | @alexey-milovidov Hello. So the only solution is to support COLUMNS in ORDER BY? | 2020-09-25T17:53:42 | 2020-09-25T17:53:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 698,731,740 | 12,513 | filimonov | @zhang2014 @4ertus2 can you take a look on that? | 2020-09-25T05:41:36 | 2020-09-25T05:41:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 698,854,165 | 15,003 | tavplubix | Functional stateless tests (release, DatabaseAtomic) — fail: 3 - wrong skip_list (now it's DatabaseOrdinary)
Stress test (memory) - use-of-uninitialized-value when logging an exception - not related, reported to CI Fest
Yandex synchronization check - wrong skip_list (Ordinary engine is used for functional tests in Ar... | 2020-09-25T10:29:47 | 2020-09-25T10:29:47 | {} | MEMBER |
ClickHouse | ClickHouse | 699,000,608 | 15,291 | alesapin | Ok, seems like it works. | 2020-09-25T15:38:43 | 2020-09-25T15:38:43 | {} | MEMBER |
ClickHouse | ClickHouse | 699,051,356 | 15,309 | den-crane | related https://github.com/ClickHouse/ClickHouse/issues/5327 | 2020-09-25T17:19:25 | 2020-09-25T17:19:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 698,685,166 | 15,263 | alexey-milovidov | > Docs check — Docs check passed | 2020-09-25T02:29:38 | 2020-09-25T02:29:38 | {} | MEMBER |
ClickHouse | ClickHouse | 698,764,145 | 15,191 | alesapin | > > Seems like test failures related to changes.
>
> @alesapin Thanks. I want to run the integration tests in my test env. Does it must run on Ubuntu ?
>
> > Running natively
> > Prerequisites:
>
> > Ubuntu 14.04 (Trusty) or higher.
> > docker. Minimum required API version: 1.25, check with docker version.
... | 2020-09-25T07:11:25 | 2020-09-25T07:11:25 | {} | MEMBER |
ClickHouse | ClickHouse | 698,772,679 | 15,270 | BernieJiangDL | > 2020071011_0_55_8_23697
> 2020071011_0_55_8_23698
>
> Last digit changed. That means some mutation (alter update/delete/modify column etc) was applied to a part.
>
> When mutating data clickhouse touches every part in the table (even if it was noop mutation for that part).
>
> Regarding loading time - what ... | 2020-09-25T07:31:44 | 2020-09-25T07:31:44 | {} | NONE |
ClickHouse | ClickHouse | 698,996,451 | 14,509 | den-crane | this PR has broke some valid cases
```
SELECT toStartOfDay(yesterday())
Received exception from server (version 20.10.1):
Code: 43. DB::Exception: Received from localhost:9000. DB::Exception: Function toStartOfDay supports a 2nd argument (optional) that must be non-empty and be a valid time zone.
SELECT ... | 2020-09-25T15:31:02 | 2020-09-25T15:31:42 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 699,029,044 | 14,414 | qoega | 01455_shard_leaf_max_rows_bytes_to_read
https://clickhouse-test-reports.s3.yandex.net/0/651518566a527f8a0c225f1d8a130712c8b810ee/functional_stateless_tests_(address).html | 2020-09-25T16:33:42 | 2020-09-25T16:33:42 | {} | MEMBER |
ClickHouse | ClickHouse | 698,701,294 | 15,257 | alexey-milovidov | Yes. `f(x) = c` can be replaced to `x >= c1 AND x < c2` (some range, maybe open or closed from the both sides, may be unbounded from one side...) | 2020-09-25T03:35:25 | 2020-09-25T03:35:25 | {
"+1": 4
} | MEMBER |
ClickHouse | ClickHouse | 698,721,035 | 15,272 | hexiaoting | Ok, I will try gcc-10 later. | 2020-09-25T05:02:15 | 2020-09-25T05:02:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 698,810,809 | 15,283 | akuzm | > Heh, I've mis-cloned AST twice...
You think it's something different? I thought it was the same and wanted to close the task, but apparently forgot to press the button. | 2020-09-25T08:55:52 | 2020-09-25T08:55:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 698,892,497 | 15,293 | alesapin | Fix for race found by stress test https://github.com/ClickHouse/ClickHouse/pull/15304 | 2020-09-25T12:08:47 | 2020-09-25T12:08:47 | {} | MEMBER |
ClickHouse | ClickHouse | 698,960,325 | 15,295 | den-crane | MaterializedView is an insert trigger.
It gets data from **INSERT**.
It **never** reads (selects) from table.
You must create MaterializedView at a server where inserts happen. You can re-route MaterializedView output to remote server. `create table xxx as remote(); create Materialized View .... to xxx as se... | 2020-09-25T14:25:39 | 2020-09-25T14:25:39 | {
"+1": 3
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 698,998,047 | 15,301 | filimonov | > Could you please reproduce this issue
It reproduces once per few weeks...
Regarding lack of call frame info - there are 2 possible reasons:
1)
```
BFD: warning: /tmp/core.ParalInputsProc.103943 is truncated: expected core file size >= 364188540928, found: 123735351296
```
2) i'm analyzing core dump offlin... | 2020-09-25T15:33:53 | 2020-09-25T15:37:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 699,028,586 | 14,414 | qoega | 01249_bad_arguments_for_bloom_filter
https://clickhouse-test-reports.s3.yandex.net/0/788b62b931c28b63ae47f1adce4ce5ca80cd3c68/functional_stateless_tests_(release,_databaseatomic).html | 2020-09-25T16:32:43 | 2020-09-25T16:32:43 | {} | MEMBER |
ClickHouse | ClickHouse | 698,667,051 | 15,212 | dependabot-preview[bot] | OK, I won't notify you again about this release, but will get in touch when a new version is available. If you'd rather skip all updates until the next major or minor version, let me know by commenting `@dependabot ignore this major version` or `@dependabot ignore this minor version`.
If you change your mind, just re-... | 2020-09-25T01:13:34 | 2020-09-25T01:13:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 698,779,360 | 15,257 | amosbird | Inversing intDiv is hard because of potential arithmetic overflow. | 2020-09-25T07:47:13 | 2020-09-25T07:47:13 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 698,878,695 | 15,300 | alesapin | Fixes race condition https://clickhouse-test-reports.s3.yandex.net/0/9bf27fb42af50b639373b18e2a4d42707b81b27f/functional_stateful_tests_(thread).html. The solution is quite ugly. Another way is to copy global context into system logs. In this case, we will have a partially initialized copy. I think the current solution... | 2020-09-25T11:32:01 | 2020-09-25T11:32:01 | {} | MEMBER |
ClickHouse | ClickHouse | 699,087,798 | 15,236 | alesapin | Wow
| 2020-09-25T18:34:35 | 2020-09-25T18:34:56 | {} | MEMBER |
ClickHouse | ClickHouse | 698,695,150 | 15,257 | alexey-milovidov | One to one mapping is not required. For example, a preimage of a point can be interval. As in examples in this task. | 2020-09-25T03:08:16 | 2020-09-25T03:08:16 | {} | MEMBER |
ClickHouse | ClickHouse | 698,615,777 | 5,183 | jacobrandall | Seeing a ton of these messages in all our clusters (a few every second per node). Would it make sense to change these to trace log level instead of error? (also running `20.3.4.10`)
```
2020.09.25 03:32:09.024210 [ 327 ] {} <Error> ServerErrorHandler: Poco::Exception. Code: 1000, e.code() = 104, e.displayText() = C... | 2020-09-24T22:12:56 | 2020-09-25T03:32:54 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 698,972,098 | 15,303 | alexey-milovidov | If table is replicated, hashes will be used to ensure that data on replicas will be byte-identical after replication. It protects from bit-flip errors in HDD (bit rot), bit flips in RAM on the server, RAM on the network controller, RAM on the network switches, data corruption during network transfer, memory stomping bu... | 2020-09-25T14:46:40 | 2020-09-25T14:46:40 | {} | MEMBER |
ClickHouse | ClickHouse | 698,667,037 | 15,212 | alexey-milovidov | Not significant. | 2020-09-25T01:13:32 | 2020-09-25T01:13:32 | {} | MEMBER |
ClickHouse | ClickHouse | 698,697,917 | 15,257 | amosbird | Ok.. Then the operator would also need to be rewritten. | 2020-09-25T03:20:15 | 2020-09-25T03:20:15 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 698,703,646 | 7,948 | amosbird | It's fixed in https://github.com/ClickHouse/ClickHouse/pull/13497 . And `toDate(timestamp_ms / 1000) ` will also work because of https://github.com/ClickHouse/ClickHouse/pull/14513 . | 2020-09-25T03:45:59 | 2020-09-25T03:45:59 | {
"+1": 3,
"rocket": 1
} | COLLABORATOR |
ClickHouse | ClickHouse | 698,732,572 | 14,002 | alexey-milovidov | Fixed in this commit: https://github.com/ClickHouse/ClickHouse/pull/13528/commits/8d616444e4fa723e6cb5e70d2189cb005c6b5d9e | 2020-09-25T05:44:27 | 2020-09-25T05:44:27 | {} | MEMBER |
ClickHouse | ClickHouse | 698,786,969 | 12,741 | alex-zaitsev | Since short circuit evaluation https://github.com/ClickHouse/ClickHouse/issues/12587 is not available, let's finalize this one since it is a critical performance feature for some users with row-level security. | 2020-09-25T08:04:08 | 2020-09-25T08:04:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 699,032,091 | 14,414 | tavplubix | > 01249_bad_arguments_for_bloom_filter
> https://clickhouse-test-reports.s3.yandex.net/0/788b62b931c28b63ae47f1adce4ce5ca80cd3c68/functional_stateless_tests_(release,_databaseatomic).html
Already fixed | 2020-09-25T16:40:20 | 2020-09-25T16:40:20 | {} | MEMBER |
ClickHouse | ClickHouse | 699,028,006 | 15,314 | den-crane | You have no choice except to move un_id to the beginning of PRIMARYINDEX.
And try to clean un_id from 'us-' and save UUID as UUID (or as hex/FIXED String) not a String.
```
create table product (
country String,
un_id UUID,
organization_id LowCardinality(String),
request... | 2020-09-25T16:31:29 | 2020-09-25T16:56:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 698,682,339 | 14,977 | ghost | @alexey-milovidov
I added `indexOfCanUseBloomFilter()` to give hint whether `indexOf()` could use bloom_filter or not.
`indexOfCanUseBloomFilter()` returns false in the case `indexOf(arr, x) = 0` because test for non-existence is impossible at bloom_filter. | 2020-09-25T02:17:45 | 2020-09-25T02:17:45 | {} | NONE |
ClickHouse | ClickHouse | 698,790,654 | 15,283 | akuzm | https://github.com/ClickHouse/ClickHouse/issues/14695 | 2020-09-25T08:12:18 | 2020-09-25T08:12:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 698,847,342 | 15,144 | akuzm | https://clickhouse-test-reports.s3.yandex.net/15144/26abe8cb30819eea1c1f3383cc99620ab2e9da9a/performance_comparison/report.html#fail1
disabling percpu arenas doesn't really influence the performance | 2020-09-25T10:12:54 | 2020-09-25T10:12:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 698,964,319 | 15,305 | alexey-milovidov | > Docs check — Docs check passed | 2020-09-25T14:32:46 | 2020-09-25T14:32:46 | {} | MEMBER |
ClickHouse | ClickHouse | 699,042,749 | 15,311 | den-crane | https://stackoverflow.com/questions/64064438/getting-lot-of-dbnetexception-connection-reset-by-peer-while-reading-from-s | 2020-09-25T17:01:59 | 2020-09-25T17:01:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 699,059,664 | 15,317 | alexey-milovidov | Yes, it will be better to add the support for COLUMNS in ORDER BY.
The only detail is to support ASC/DESC, COLLATE and WITH FILL modifiers. | 2020-09-25T17:36:19 | 2020-09-25T17:36:19 | {} | MEMBER |
ClickHouse | ClickHouse | 699,083,334 | 15,319 | bharatnc | cc: @den-crane | 2020-09-25T18:25:03 | 2020-09-25T18:25:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 699,246,966 | 15,321 | alexey-milovidov | Usage:
```
clickhouse-format --seed Hello --obfuscate <<< "SELECT 888"
``` | 2020-09-26T00:42:03 | 2020-09-26T00:42:03 | {} | MEMBER |
ClickHouse | ClickHouse | 703,615,766 | 15,596 | den-crane | > This solves connecting to another node using the `remote` function.
> But the replication problem still exists, when I insert in a region, the data is not replicated in the other region.
**select * from system.replication_queue** on all replicas | 2020-10-05T13:00:28 | 2020-10-05T13:00:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 703,836,311 | 15,590 | azat | @abyss7 what is broken here? Oh I see, some changes after breaks it, and has been fixed shortly in 674d8d44b9e2542e142339c3002da3fb4f3eeb34 | 2020-10-05T19:19:29 | 2020-10-05T19:19:29 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.