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 | 397,915,340 | 2,522 | alexey-milovidov | :+1: | 2018-06-18T00:02:24 | 2018-06-18T00:02:24 | {} | MEMBER |
ClickHouse | ClickHouse | 399,018,736 | 1,867 | paruru48 | any update for this issue ??? | 2018-06-21T08:24:23 | 2018-06-21T08:24:23 | {} | NONE |
ClickHouse | ClickHouse | 399,037,422 | 2,525 | adubovikov | confirm, same issue, if the query contains FINAL and PREWHERE:
```
2018.06.21 09:20:28.015993 [ 12 ] <Debug> data.test_table (SelectExecutor): Key condition: unknown, unknown, and, (column 0 in [1529536154257000, +inf)), (column 0 in (-inf, 1529540354257000]), and, and, unknown, and
2018.06.21 09:20:28.016047 [ 12... | 2018-06-21T09:27:27 | 2018-06-21T09:27:27 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 399,246,978 | 2,540 | tigertv | @filimonov, Действительно так и есть, спасибо за пояснение.
Полагаю, что это работает для всех функций которые принимают константные параметры, и нормальные компиляторы это оптимизируют таким образом. | 2018-06-21T21:19:57 | 2018-06-21T22:00:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 400,376,986 | 2,557 | tigertv | @filimonov,
OK. Ich verstehe das. | 2018-06-26T16:27:59 | 2018-06-26T16:27:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 400,608,891 | 2,070 | ben-motiv | Same here on latest Docker build. How do you achieve write sharding from Kafka without a distributed table to handle it? | 2018-06-27T09:37:25 | 2018-06-27T09:37:25 | {} | NONE |
ClickHouse | ClickHouse | 401,116,627 | 2,490 | alexey-milovidov | Looks perfectly correct. Thank you! | 2018-06-28T17:42:27 | 2018-06-28T17:42:27 | {} | MEMBER |
ClickHouse | ClickHouse | 401,157,305 | 2,482 | alexey-milovidov | `server_logs_level` is non obvious name.
The user may think that it will affect the log level directly on server.
Will rename... | 2018-06-28T20:07:50 | 2018-06-28T20:07:50 | {} | MEMBER |
ClickHouse | ClickHouse | 397,535,743 | 2,190 | filimonov | #2447 | 2018-06-15T07:19:54 | 2018-06-15T07:19:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 397,622,812 | 2,517 | sleshJdev | I just tested with same values but as _tsv_, and it works. But unfortunately I cannot migrate to _tsv_ right now. I need to find out some workaround for _csv_ to make it work with signle quote in field value. | 2018-06-15T13:38:17 | 2018-06-15T13:38:17 | {} | NONE |
ClickHouse | ClickHouse | 399,221,733 | 2,540 | tigertv | Alexey, do you really think that there is all right?
https://github.com/yandex/ClickHouse/blob/ea947ce808fcdcf36390d3ad9067b8969c619113/utils/compressor/mutator.cpp#L68-L73
here we see 3 calls strlen, the faster way is to change it on a call:
```cpp
size_t len = strlen("https");
if (pos + len <= end && 0 == me... | 2018-06-21T19:49:18 | 2018-06-21T20:34:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 400,288,969 | 2,529 | filimonov | @Joozty for JDBC - use setMaxQuerySize method.
For HTTP - send that additional parameter **together** with your query. Like post query to `http://localhost:8123/?max_query_size=100000000` or do `http://localhost:8123/?max_query_size=100000000&query=SELECT%201`
About users.xml - it SHOULD work. If it don't - proba... | 2018-06-26T12:26:49 | 2018-06-26T12:27:23 | {
"+1": 1,
"-1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 397,622,315 | 2,517 | sleshJdev | Maybe is there some option to force using a only double quote as a string delimiter, and consider single quote as plain symbol without special meaning? Please, any solution is welcome. | 2018-06-15T13:36:32 | 2018-06-15T13:50:08 | {} | NONE |
ClickHouse | ClickHouse | 399,767,035 | 2,535 | filimonov | 1) I doubt tgat performance of numbers table can be a bottleneck in some real scenarios.
2) you can simplyfy such a selects using `WITH` clause like
```sql
with number + 10 as number2
select number2 as a, number2 as b ...
```
3) may be it would be better to introduce range/enumerate table function which will s... | 2018-06-24T16:00:55 | 2018-06-24T16:01:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 399,857,318 | 2,529 | dveselov | I'm using this docker image too and have similar question.
Limits on `max_query_size`, that I defined in `/etc/clickhouse-server/config.xml` doesn't work. | 2018-06-25T07:25:08 | 2018-06-25T07:25:08 | {
"+1": 4
} | NONE |
ClickHouse | ClickHouse | 400,924,869 | 2,568 | inolddays | no . I have two shards and each shard has two replicas.
Only one shard's replica's table is dropped. | 2018-06-28T06:20:21 | 2018-06-28T06:20:21 | {} | NONE |
ClickHouse | ClickHouse | 399,857,857 | 2,529 | dveselov | If it help, I'm fetching data from ClickHouse with large number of entries in `IN ( ... )` clause.
And ClickHouse throws this exception:
```
DB::Exception: Syntax error: failed at position 262142 (line 3, col 262083): 29900, 31794, 30743, 27330, 86017, 1237, 86395, 26248, 76827, 8051, 68091, 32989, 15449, 84636, 63... | 2018-06-25T07:27:20 | 2018-06-25T07:27:20 | {} | NONE |
ClickHouse | ClickHouse | 400,062,751 | 2,552 | tigertv | @alexey-milovidov, Yes, maybe the compiler does it this way. I didn´t check it too :) | 2018-06-25T19:09:25 | 2018-06-25T19:09:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 400,209,541 | 1,128 | jackpgao | Does this bug have been fixed? | 2018-06-26T07:33:16 | 2018-06-26T07:33:16 | {} | NONE |
ClickHouse | ClickHouse | 400,304,963 | 2,556 | alexey-milovidov | Also worth to note that we don't care about performance in debug builds. | 2018-06-26T13:20:05 | 2018-06-26T13:20:05 | {} | MEMBER |
ClickHouse | ClickHouse | 398,763,554 | 2,534 | proller | How frequent it is?
Maybe you have network problems? (packet loss. ...) | 2018-06-20T14:10:52 | 2018-06-20T14:10:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 399,402,776 | 2,532 | filimonov | yep. We don't :)
The first result when you google ya.make is [catboost issue](https://github.com/catboost/catboost/issues/34) asking to get rid of it and use some widely used 'normal' build system instead of closed-source yandex in-house build system. :) And some comments about plans to open-source ya make. | 2018-06-22T10:47:22 | 2018-06-22T10:47:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 400,279,630 | 1,631 | filimonov | https://stackoverflow.com/questions/14370489/what-can-cause-a-resource-temporarily-unavailable-on-sock-send-command
Can you check if send_timeout is set? Are you using standard MTU size on your interface? | 2018-06-26T11:52:25 | 2018-06-26T11:52:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 400,932,261 | 2,565 | filimonov | I also always have those warnings when server starts, but it does not shut down.
May be because i have no Dictionary database? | 2018-06-28T06:56:42 | 2018-06-28T06:56:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 401,115,455 | 2,490 | alexey-milovidov | Merged manually
(because branch was created from `stable`, not from `master`) | 2018-06-28T17:38:20 | 2018-06-28T17:38:20 | {} | MEMBER |
ClickHouse | ClickHouse | 398,456,120 | 2,447 | Tezd | @alexey-milovidov it doesn't trigger completion on paste, but will clear tabs from pasted text (at least this is behavior in my env) If it shows different behavior I will investigate it. Also not sure why builds are failing from time to time. | 2018-06-19T16:10:34 | 2018-06-19T16:10:34 | {} | NONE |
ClickHouse | ClickHouse | 399,245,115 | 2,540 | filimonov | Посмотрите: https://godbolt.org/g/TWj764
Строку 10 компилятор на этапе компиляции не может вычислить и поэтому в скомпилированном коде (на строке 33) присутствует вызов функции strlen.
Вызов встроенной функции на строке 11 компилятор может произвести самостоятельно, поэтому на строке 44 в скомпилированном файле... | 2018-06-21T21:12:25 | 2018-06-21T21:12:25 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 399,388,831 | 2,266 | alexey-milovidov | We've tested ClickHouse on Cavium ThunderX2 with two sockets, 224 threads.
It can work faster than some Intel servers (for example, E2650v2 with two sockets, 32 threads) on heavy queries (on queries that run for 500 ms or longer).
But you have to pay attention to hyper threading. For example, 224 threads on this se... | 2018-06-22T09:51:57 | 2018-06-22T09:51:57 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 399,396,564 | 2,532 | alexey-milovidov | People have no idea, what does it mean. | 2018-06-22T10:21:18 | 2018-06-22T10:21:18 | {} | MEMBER |
ClickHouse | ClickHouse | 400,317,340 | 2,416 | derekperkins | @simPod I am interested in the same, but I've seen it mentioned in other issues that if you have identical subqueries, ClickHouse only actually runs it once. | 2018-06-26T13:54:29 | 2018-06-26T13:54:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 400,320,520 | 2,416 | simPod | @derekperkins cool, good to know, thanks! | 2018-06-26T14:03:24 | 2018-06-26T14:03:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 400,528,552 | 2,535 | zhang2014 | > I doubt tgat performance of numbers table can be a bottleneck in some real scenarios.
Yes, if you use `LIMIT m, n` you will have unnecessary wait offset, and if you use `number + m ... LIMIT n` you will have unnecessary operations, like '+'
> you can simplyfy such a selects using WITH clause like
That's great
... | 2018-06-27T03:06:26 | 2018-06-27T03:06:39 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 400,980,825 | 2,298 | marklit | @LiuYangkuan any ETA on when your changes will be merged into master and released? | 2018-06-28T09:55:06 | 2018-06-28T09:55:06 | {
"+1": 2
} | NONE |
ClickHouse | ClickHouse | 401,059,289 | 2,553 | ssmike | dbms/src/Common/tests/radix_sort.cpp
| 2018-06-28T14:43:03 | 2018-06-28T14:43:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 398,226,057 | 2,299 | k0st1an | I changed default settings to:
```
<listen_host>127.0.0.1</listen_host>
<listen_host>192.168.56.102</listen_host>
```
When executing command `$ sudo reboot`. When OS boot tries to connect to server CH. But get the error:
```
$ clickhouse-client
ClickHouse client version 1.1.54385.
Connecting to l... | 2018-06-18T23:25:00 | 2018-06-18T23:25:00 | {} | NONE |
ClickHouse | ClickHouse | 400,208,507 | 1,631 | jackpgao | The same problem happened. | 2018-06-26T07:29:23 | 2018-06-26T07:29:23 | {} | NONE |
ClickHouse | ClickHouse | 400,284,562 | 2,416 | dveselov | Bump. | 2018-06-26T12:11:20 | 2018-06-26T12:11:20 | {} | NONE |
ClickHouse | ClickHouse | 397,928,337 | 2,520 | alexey-milovidov | Install packages:
```
git
cmake
ninja-build
gcc gcc-c++ libstdc++-static
openssl-devel
libicu-devel
readline-devel
unixODBC-devel libtool-ltdl-devel
```
```
cmake -D ENABLE_MYSQL=0 ..
``` | 2018-06-18T02:39:38 | 2018-06-18T02:39:38 | {} | MEMBER |
ClickHouse | ClickHouse | 398,384,294 | 2,187 | xiaoshuai | try to add `/etc/yum.repos.d/clickhouse.repo`
```
[clickhouse]
name=clickhouse repo
baseurl=http://repo.yandex.ru/clickhouse/rpm/stable/$basearch/
gpgcheck=0
enabled=1
``` | 2018-06-19T12:39:37 | 2018-06-19T12:39:37 | {} | NONE |
ClickHouse | ClickHouse | 399,924,713 | 300 | hereTac | 1.Create tablea(memory).
2.Create tableb(mergetree). Primary key is column with type Int.
3. Load csv into the memory.
4.Check clickhouse-server cached 4G memory.
5.Exe SQL: Insert into tableb select * from tablea;
6.Exe SQL: Drop table tablea;
7.Check clickhouse-server still cache 4G+ memory.
It still here... | 2018-06-25T11:50:33 | 2018-06-25T11:50:53 | {} | NONE |
ClickHouse | ClickHouse | 400,307,695 | 2,557 | alexey-milovidov | For my opinion, this modification doesn't make sense. Code readability is about the same.
Tell me if there is some motivation that I have missed. | 2018-06-26T13:27:10 | 2018-06-26T13:27:10 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 400,910,992 | 2,490 | stavrolia | In updated version of this branch it is able to send >16K (I tested on 20K). I can add it to test if necessary. | 2018-06-28T04:59:00 | 2018-06-28T04:59:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 399,241,441 | 2,540 | tigertv | @filimonov, Т.е. получается, что в итоговом откомпилированном файле в месте где должна вызываться эта функция по сути она не вызывается, а подставляется результат вызова функции?
А является ли это компиляторозависимым? | 2018-06-21T20:58:19 | 2018-06-21T21:09:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 399,739,167 | 2,550 | alex-zaitsev | Look at Merge() engine.
https://clickhouse.yandex/docs/en/table_engines/merge/ | 2018-06-24T08:30:21 | 2018-06-24T08:30:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 399,987,649 | 904 | jackpgao | It seems that the problem comes from the zookeeper which is very slow.
| 2018-06-25T15:14:16 | 2018-06-25T15:14:16 | {} | NONE |
ClickHouse | ClickHouse | 400,304,637 | 2,556 | alexey-milovidov | It's not better because it makes code less readable and the logic becomes hidden behind the code.
But C++ compiler always implement so called "peephole optimizations".
https://ridiculousfish.com/blog/posts/will-it-optimize.html
(this quiz is slightly outdated - about 8 years ago) | 2018-06-26T13:19:08 | 2018-06-26T13:19:08 | {
"+1": 2
} | MEMBER |
ClickHouse | ClickHouse | 400,516,873 | 1,631 | jackpgao | Update:
This is what I met: https://github.com/yandex/ClickHouse/issues/1128 | 2018-06-27T01:49:49 | 2018-06-27T01:49:49 | {} | NONE |
ClickHouse | ClickHouse | 401,037,161 | 2,553 | alexey-milovidov | :+1: | 2018-06-28T13:36:47 | 2018-06-28T13:36:47 | {} | MEMBER |
ClickHouse | ClickHouse | 398,691,107 | 2,001 | perez1987 | any news? | 2018-06-20T09:48:12 | 2018-06-20T09:48:12 | {} | NONE |
ClickHouse | ClickHouse | 399,929,901 | 904 | jackpgao | I've got the same problem. | 2018-06-25T12:12:38 | 2018-06-25T12:12:38 | {} | NONE |
ClickHouse | ClickHouse | 401,037,216 | 2,553 | alexey-milovidov | Missing test. | 2018-06-28T13:36:57 | 2018-06-28T13:36:57 | {} | MEMBER |
ClickHouse | ClickHouse | 397,779,398 | 2,506 | alexey-milovidov | Superseded by #2514 | 2018-06-16T02:11:57 | 2018-06-16T02:11:57 | {} | MEMBER |
ClickHouse | ClickHouse | 399,829,020 | 2,551 | filimonov | Clickhouse uses embedded statically linked version of librdkafka.
Cmake script should be adjusted for making librdkafka "see" openssl and zlib (see also #2531). @proller ? | 2018-06-25T04:43:31 | 2018-06-25T06:05:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 400,043,874 | 2,552 | alexey-milovidov | Thank you!
BTW, it should not make any difference, because compiler is smart enough to figure out that `res` variable is not modified in first branch (because `res` and `pos` have different types and C++ has "strict aliasing" rule, that is turned on in our build (we use `-O3` option)) and consequently, first and sec... | 2018-06-25T18:08:14 | 2018-06-25T18:09:32 | {} | MEMBER |
ClickHouse | ClickHouse | 400,280,273 | 2,529 | filimonov | @Joozty what client software do you use? JDBC? | 2018-06-26T11:55:11 | 2018-06-26T11:57:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 399,884,381 | 2,529 | filimonov | I know that it's not clear from documentation where to put some config parameters.
Generally clickhouse have 2 biggest group of settings:
1) "**user**" settings
2) "**server**" settings
You can check if some setting is "user"-setting by checking it in one of those places:
1) `system.settings` table
2) CH so... | 2018-06-25T09:09:08 | 2018-06-26T12:30:20 | {
"+1": 5
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 400,331,387 | 2,557 | tigertv | If it is assumed to be called not very frequently it doesn't mean you don't need the enchancement.
So If you don't need, then don't get it. | 2018-06-26T14:33:32 | 2018-06-26T15:20:41 | {
"-1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 400,367,157 | 2,563 | nickola | @filimonov Да, как раз переключился на Log в качестве workaround. | 2018-06-26T16:03:20 | 2018-06-26T16:03:20 | {} | NONE |
ClickHouse | ClickHouse | 398,445,461 | 2,530 | blinkov | Join syntax is pretty limited at the moment and it's on the roadmap for this year to fix it.
Currently the most common option is joining subqueries connected with USING instead of ON, see the docs: https://clickhouse.yandex/docs/en/query_language/queries/#join-clause | 2018-06-19T15:40:37 | 2018-06-19T15:40:37 | {
"heart": 6
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 398,911,405 | 2,525 | proller | ready test:
```
DROP TABLE IF EXISTS test.final_test;
CREATE TABLE test.final_test (id String, version Date) ENGINE = ReplacingMergeTree(version, id, 8192);
INSERT INTO test.final_test (id, version) VALUES ('2018-01-01', '2018-01-01'); ... | 2018-06-20T21:59:39 | 2018-06-20T21:59:39 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 399,087,682 | 2,539 | filimonov | https://github.com/yandex/ClickHouse/issues/1661#issuecomment-352739726 | 2018-06-21T12:33:22 | 2018-06-21T12:33:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 399,239,906 | 2,540 | filimonov | Значение этой функции вычисляется на этапе компиляции, так как параметр является константой. Во время выполнения программы эти функции не вызываются, а сразу подставляются вычисленные на этапе компиляции значения. | 2018-06-21T20:52:39 | 2018-06-21T20:52:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 399,330,534 | 2,539 | sdlcwangsong | @filimonov thank you | 2018-06-22T06:07:14 | 2018-06-22T06:07:14 | {} | NONE |
ClickHouse | ClickHouse | 400,324,846 | 2,557 | tigertv | Let's say, old_size is checked only once and not 3 times in the worst case.
And I don't think the compiler will optimise it right. | 2018-06-26T14:15:40 | 2018-06-26T14:27:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 400,360,541 | 2,563 | filimonov | Похоже так происходит только если таблица была изначально пустая.
В качестве workaround для быстрого решения - используйте движок Log. С ним проблема не воспроизводится.
```sql
CREATE TABLE data ( value Int8 ) ENGINE = Log;
INSERT INTO data SELECT 1 WHERE 1 = 0;
SELECT count() FROM data
┌─count()─┐
│ ... | 2018-06-26T15:48:44 | 2018-06-26T15:48:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 400,934,681 | 2,568 | inolddays | thanks I'll have a try . | 2018-06-28T07:08:07 | 2018-06-28T07:08:07 | {} | NONE |
ClickHouse | ClickHouse | 397,624,349 | 2,517 | sleshJdev | I asked the [question](https://stackoverflow.com/questions/50876839/couldnt-parse-csv-file-if-value-start-from-or) on the SO also. | 2018-06-15T13:44:04 | 2018-06-15T13:44:04 | {} | NONE |
ClickHouse | ClickHouse | 397,928,266 | 2,520 | alexey-milovidov | Successfully built in Docker. | 2018-06-18T02:39:06 | 2018-06-18T02:39:06 | {} | MEMBER |
ClickHouse | ClickHouse | 398,002,527 | 2,517 | KochetovNicolai | Now there is no option to use only double quote. It seems that the best way is to prepare data before insertion. For example, wrap `'Ewa Beach` into double quotes. | 2018-06-18T09:54:01 | 2018-06-18T09:54:01 | {} | MEMBER |
ClickHouse | ClickHouse | 398,098,686 | 2,524 | proller | please show how you run cmake and cmake console output
you can try redefine rdkafka libs:
```cmake .. -DRDKAFKA_LIB=/path/to/rdkafka;/path/to/libgnutls.a``` | 2018-06-18T15:40:51 | 2018-06-18T15:40:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 398,995,558 | 2,507 | dengwanc | ```
select uid, if(toUInt8(uid), 's', 'v') as t from nxlog.table limit 10
union all
select uid, if(toUInt8(uid), 's', 'v') as s from nxlog.table limit 10
```
worked but stupid | 2018-06-21T06:52:31 | 2018-06-21T06:52:44 | {
"laugh": 1
} | NONE |
ClickHouse | ClickHouse | 399,182,396 | 2,540 | alexey-milovidov | strlen is a builtin, and strlen of string literal is a compile time constant.
Your changes actually made code slower, because std::string::length is not a compile time constant in this case. | 2018-06-21T17:28:22 | 2018-06-21T17:28:31 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 399,840,875 | 2,226 | chenxing-xc | is there any progress on this fix? we hit the same problem multiple times and has to disable optimize_move_to_prewhere to workaround the problem. | 2018-06-25T06:02:14 | 2018-06-25T06:02:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 400,253,381 | 2,490 | stavrolia | There is a basic test, and what do you mean by external tables? | 2018-06-26T10:04:03 | 2018-06-26T10:04:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 400,283,611 | 2,529 | Joozty | @filimonov Both JDBC and also HTTP. | 2018-06-26T12:07:31 | 2018-06-26T12:07:31 | {} | NONE |
ClickHouse | ClickHouse | 400,553,926 | 2,045 | filimonov | Update: as Clickhouse stores column data in blocks and each block is compressed and can be processed separately, encryprion can also be done on block level. In that case initialization vector could be unique per block, and it will give really tiny overhead. | 2018-06-27T06:09:43 | 2018-06-27T06:10:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 401,090,068 | 2,571 | den-crane | This is because of the features:
>ClickHouse release 1.1.54343
>Now the table index is used for conditions like expr IN (subquery).
> (ClickHouse release 1.1.54356)
> An index of MergeTree tables is used when IN is applied to a tuple of expressions from the
>columns of the primary key. Example: WHERE (UserID,... | 2018-06-28T16:14:24 | 2018-06-28T16:25:21 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 397,842,853 | 2,447 | alexey-milovidov | TODO: It shouldn't work when you paste a text with tabs in terminal. | 2018-06-16T22:43:01 | 2018-06-16T22:43:01 | {} | MEMBER |
ClickHouse | ClickHouse | 397,942,540 | 2,520 | chipitsine | it died with OOM on linking :) | 2018-06-18T05:01:56 | 2018-06-18T05:01:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 398,546,643 | 2,299 | proller | Code: 210. DB::NetException: Connection refused: (localhost:9000, ::1)
- your client tries to use ipv6.
change
`<listen_host>127.0.0.1</listen_host>`
to
`<listen_host>::1</listen_host>`
or you can configure default host for client:
/etc/clickhouse-client/config.xml:
`<host>127.0.0.1</host>`
| 2018-06-19T21:08:58 | 2018-06-19T21:08:58 | {
"+1": 6,
"laugh": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 398,618,516 | 2,517 | filimonov | Duplicate of #2192
Important notice - because of that clickhouse csv parser is not RFC compatible. | 2018-06-20T04:11:05 | 2018-06-20T04:11:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 398,683,219 | 2,536 | valo | My bad. Wrong place for this PR | 2018-06-20T09:20:50 | 2018-06-20T09:20:50 | {} | NONE |
ClickHouse | ClickHouse | 399,392,330 | 2,535 | alexey-milovidov | You didn't provide any motivation for this change.
So I have to guess, is it any useful.
At first glance, it seems completely useless,
because the user can always write
```
SELECT number + 10 FROM numbers(123)
```
instead of
```
SELECT number FROM numbers(10, 123)
```
and proposed change will only make que... | 2018-06-22T10:04:35 | 2018-06-22T10:04:35 | {} | MEMBER |
ClickHouse | ClickHouse | 400,320,334 | 2,416 | filimonov | Clickhouse supports `WITH` clause since mid 2017, but only for scalar values.
Sample syntax:
```
WITH
[1,2,3,4,5,6,7,8,9] as numbers,
arrayMap(x -> x*2, numbers) as doubled_numbers,
arrayMap(x -> x*x, numbers) as squared_numbers,
arrayMap( (x,y) -> x+y, doubled_numbers, squared_numbers) as summed_num... | 2018-06-26T14:02:54 | 2018-06-26T14:02:54 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 400,327,987 | 2,556 | tigertv | @alexey-milovidov, Thanks for the link. Debug builds becomes releases.) | 2018-06-26T14:24:21 | 2018-06-26T14:54:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 400,702,592 | 2,566 | kvap | cc @vavrusa, @alexey-milovidov, @twalwyn, @lwolf, @Dorokhov | 2018-06-27T14:55:02 | 2018-06-27T14:55:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 400,910,571 | 2,568 | filimonov | Did you drop it on all the replicas? | 2018-06-28T04:55:35 | 2018-06-28T04:55:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 401,122,291 | 2,571 | alexey-milovidov | It's not normal. Will fix. | 2018-06-28T18:01:57 | 2018-06-28T18:01:57 | {
"+1": 3,
"hooray": 1
} | MEMBER |
ClickHouse | ClickHouse | 399,296,331 | 2,263 | hereTac | does `--format_csv_delimiter=` support in v1.1.54385-stable?
Errors:
Bad arguments: unrecognised option '--format_csv_delimiter= ' | 2018-06-22T01:54:07 | 2018-06-22T01:54:07 | {
"heart": 2
} | NONE |
ClickHouse | ClickHouse | 399,495,388 | 2,535 | zhang2014 | Sorry, I didn't introduce motivation.
> At first glance, it seems completely useless,
> because the user can always write
> SELECT number + 10 FROM numbers(123)
> instead of
> SELECT number FROM numbers(10, 123)
> and proposed change will only make query language more heavy.
You are right, but
- it's alw... | 2018-06-22T16:09:52 | 2018-06-22T16:24:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 399,765,640 | 2,550 | Slind14 | Thank you, is it possible to supply an array of table names? Unfortunately we use uuids as table names.. | 2018-06-24T15:39:40 | 2018-06-24T15:39:40 | {} | NONE |
ClickHouse | ClickHouse | 400,267,491 | 2,529 | Joozty | Thanks, @filimonov I tried to put it into `user.xml` and `config.xml` files without any effect.
I also tried to send http request:
`http://localhost:8123/?max_query_size=100000000`
But this seems it doesn't work at all.
This is the response `Code: 62, e.displayText() = DB::Exception: Empty query, e.what() = ... | 2018-06-26T11:00:01 | 2018-06-26T11:00:01 | {} | NONE |
ClickHouse | ClickHouse | 400,800,397 | 1,179 | filimonov | See #2490 | 2018-06-27T19:25:20 | 2018-06-27T19:25:20 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 397,571,773 | 2,480 | laoguanjie | @KochetovNicolai thx your reply:)
i use version 1.1.54381, and it's easy to reproduce.
i test ch with two points:
> 1、100 threads add columns concurrently.
2、column which these threads try to add may by the same, or not.
i use clickhouse-jdbc to try it, and my code like
```
ClickhouseClient client = n... | 2018-06-15T09:49:49 | 2018-06-15T09:50:11 | {} | NONE |
ClickHouse | ClickHouse | 397,918,414 | 2,520 | alexey-milovidov | Relevant details: https://stackoverflow.com/questions/20759750/resolving-redefinition-of-timespec-in-time-h | 2018-06-18T00:51:41 | 2018-06-18T00:51:41 | {} | MEMBER |
ClickHouse | ClickHouse | 400,329,210 | 2,557 | alexey-milovidov | `Allocator::realloc` is assumed to be called not very frequently (for example, less than a million times per second and this is huge overestimation). Excessive comparison doesn't make any difference in this case. | 2018-06-26T14:27:46 | 2018-06-26T14:27:46 | {} | MEMBER |
ClickHouse | ClickHouse | 400,348,910 | 2,557 | filimonov | @tigertv it's nice that you like to help open source project, but please stop fixing the things that are not broken :) "if it ain't broke don't fix it".
About performance improvements: the main concept behind each performance improvement is measurement, as a prove of speed up (otherwise you have a risk to slow down ... | 2018-06-26T15:19:47 | 2018-06-26T15:21:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 400,860,287 | 2,567 | den-crane | your columns have different types that is a bad idea (UInt64 vs UInt8)
Check clickhouse.err.log it should have more information.
though could not reproduce
```
create table testx(errorCount UInt64, erroneous UInt8 MATERIALIZED errorCount) Engine=MergeTree partition by tuple() order by tuple();
insert int... | 2018-06-27T23:21:10 | 2018-06-27T23:28:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 400,930,868 | 2,568 | filimonov | And that is the reason. If you will drop table on all the replicas then table metadata will be removed from zookeeper. Till that moment any attempts to create table on the same zookeeper path will be considered as attempts to make another replica of that table (and all table replicas should have the same structure). | 2018-06-28T06:49:47 | 2018-06-28T06:49:47 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 401,084,890 | 2,569 | alex-zaitsev | Fixed in 1.1.54388 | 2018-06-28T15:58:15 | 2018-06-28T15:58:15 | {} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.