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 | 376,008,502 | 2,107 | alexey-milovidov | Each table use single shared ZooKeeper instance from the global Context (at least it should work this way).
When we have more than one active ZK sessions in a program, various bad issues may happen: loss of sequential consistency, dangling locks, etc. | 2018-03-25T22:22:21 | 2018-03-25T22:22:52 | {} | MEMBER |
ClickHouse | ClickHouse | 376,062,380 | 2,106 | jonatasfreitasv | Nice! Thx a lot. | 2018-03-26T06:44:07 | 2018-03-26T06:44:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 376,074,331 | 1,771 | dvahram | @tobad357 that's correct. | 2018-03-26T07:41:26 | 2018-03-26T07:41:26 | {} | NONE |
ClickHouse | ClickHouse | 377,961,395 | 2,156 | KochetovNicolai | Can you try to create an example how to reproduce this problem?
I've created Distributed over MergeTree with the same structure and inserted some data, but both queries on 1.1.54370 works fine for me. | 2018-04-02T15:52:41 | 2018-04-02T15:52:41 | {} | MEMBER |
ClickHouse | ClickHouse | 378,172,760 | 2,159 | santaux | Done | 2018-04-03T08:36:26 | 2018-04-03T08:36:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,243,701 | 1,770 | alexm93 | Hello!
Could you say when this bug will be fixed?
It's very critical for our development team now.
| 2018-04-03T13:08:49 | 2018-04-03T13:08:49 | {} | NONE |
ClickHouse | ClickHouse | 378,427,318 | 2,161 | jonatasfreitasv | @alexey-milovidov | 2018-04-03T23:06:38 | 2018-04-03T23:06:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 376,056,593 | 2,110 | filimonov | P.s. I didn't review all the changes. Probably there are more problems. | 2018-03-26T06:09:38 | 2018-03-26T06:09:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 377,363,222 | 867 | filimonov | Currenty NULLs looks quite mature. Probably documentation should be added. | 2018-03-29T20:32:00 | 2018-03-29T20:32:00 | {
"+1": 3
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,272,227 | 1,770 | deonwise | Всё по документации:
**UUIDStringToNum(str)**
Принимает строку, содержащую 36 символов в формате 123e4567-e89b-12d3-a456-426655440000, и возвращает в виде набора байт в FixedString(16).
**UUIDNumToString(str)**
Принимает значение типа FixedString(16). Возвращает строку из 36 символов в текстовом виде.
То ест... | 2018-04-03T14:35:22 | 2018-04-03T14:35:22 | {} | NONE |
ClickHouse | ClickHouse | 378,278,692 | 2,124 | filimonov | Это поведение задокументировано: https://clickhouse.yandex/docs/ru/query_language/queries/#_8
> При использовании обычного IN-а, запрос отправляется на удалённые серверы, и на каждом из них выполняются подзапросы в секциях IN / JOIN.
Попробуйте заменить IN на GLOBAL IN.
| 2018-04-03T14:53:23 | 2018-04-03T14:53:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,426,196 | 2,161 | jonatasfreitasv | In this image is the Best MVP Replicated/Distributed ClickHouse Architecture?
https://ibb.co/iHDygc | 2018-04-03T23:01:16 | 2018-04-03T23:01:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,502,166 | 2,161 | filimonov | > But in the case of using shard and replica at the same time, will not conflicts happen?
It is 100% safe, and in it is a standard way of ClickHouse usage.
Use replication always.
Use replication + sharding if all your data does not fit to one server.
> In this image is the Best MVP Replicated/Distributed Clic... | 2018-04-04T07:07:22 | 2018-04-04T07:08:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,559,017 | 2,146 | KochetovNicolai | https://github.com/yandex/ClickHouse/pull/2166 | 2018-04-04T10:48:32 | 2018-04-04T10:48:32 | {} | MEMBER |
ClickHouse | ClickHouse | 376,165,901 | 2,112 | silviucpp | @filimonov yes it's a bug in clickhouse that when you do this filtering : `f32 = 10.2`, 10.2 is casted to `Float64` and not to the type of `f32`. At least this is my opinion that as you also spotted the proper cast is not triggering the bug.
| 2018-03-26T13:29:39 | 2018-03-26T13:30:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,306,865 | 2,165 | ztlpn | @filimonov is right, but the particular error that you get is a parsing error. FORMAT clause should be after the SETTINGS clause. | 2018-04-03T16:14:17 | 2018-04-03T16:14:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,328,355 | 2,162 | alexey-milovidov | No. You can create Replicated tables even if you have single ClickHouse server. | 2018-04-03T17:18:49 | 2018-04-03T17:18:49 | {} | MEMBER |
ClickHouse | ClickHouse | 378,566,552 | 1,770 | deonwise | Я бинарном виде вставляю, проблем в 1000 строк не было... | 2018-04-04T11:21:46 | 2018-04-04T11:21:46 | {} | NONE |
ClickHouse | ClickHouse | 376,164,790 | 2,112 | filimonov | >I know that comparing floats is not accurate but I'm expecting to work at least on the same limitations c++ has
@silviucpp but in pure c++ there is the same limitation.
Please try: http://cpp.sh/2yjdn | 2018-03-26T13:26:03 | 2018-03-26T13:27:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 376,332,227 | 2,084 | rajadayalan | Thank You Alex. that is useful information to know. | 2018-03-26T22:27:25 | 2018-03-26T22:27:25 | {} | NONE |
ClickHouse | ClickHouse | 377,777,911 | 2,146 | alex-zaitsev | @derekperkins , 54370 has a number of problems with filters. These are regressions of some re-work performed by Yandex team in order to optimize in () queries that did not use indexes in some cases (e.g. in views). Some of those problems have been already addressed, so next release should be good. So far, the last rele... | 2018-04-01T10:39:48 | 2018-04-01T10:39:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 377,890,845 | 2,155 | jpiper | Gah, I have no idea what I'm doing here (I'm not a C++ dev)... | 2018-04-02T09:10:48 | 2018-04-02T09:10:48 | {} | NONE |
ClickHouse | ClickHouse | 378,302,908 | 2,074 | erdebee | Depending on the way you store your data you can remove data with the following strategies as well:
Use primary key that cannot be used for singling out any personal data after deletion, use ReplacingMergeTree, replace the existing row with one that contains only blanked or default values. Execute (or wait for) opti... | 2018-04-03T16:01:57 | 2018-04-03T16:03:38 | {} | NONE |
ClickHouse | ClickHouse | 378,495,723 | 2,115 | silviucpp | Hello @alexey-milovidov ,
I'm trying to investigate this issue and I have a question: For the example I provided above what branch in FunctionArray it's expected to be executed starting from line 855:
Currently the second one is executed, but I have a feeling that first one is the proper one and `item_arg->onlyNu... | 2018-04-04T06:35:33 | 2018-04-04T06:35:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 376,087,282 | 2,107 | silviucpp | Thanks for answer ! so there is no problem :) | 2018-03-26T08:31:55 | 2018-03-26T08:31:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 377,374,600 | 2,138 | alexey-milovidov | I've installed the latest gcc 7.3.0 that I'm able to download in source tarball from https://gcc.gnu.org/ and I see no issues like that. Also I have no ideas why the compiler may get fool in this case.
Nevertheless, this modification is good for me. | 2018-03-29T21:14:38 | 2018-03-29T21:16:25 | {} | MEMBER |
ClickHouse | ClickHouse | 377,658,833 | 2,148 | zhang2014 | Reference http://www.man7.org/linux/man-pages/man2/sigaction.2.html
```
Some suberrors of SIGBUS, in particular BUS_MCEERR_AO and
BUS_MCEERR_AR, also fill in si_addr_lsb. This field indicates the
least significant bit of the reported address and therefore the
extent of the corruption. ... | 2018-03-31T02:16:41 | 2018-03-31T02:16:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,245,315 | 1,770 | deonwise | Разработчики ничего не ответили, так что я перешел на использование FixedString(16) и UUIDStringToNum(). Это работает, как я понимаю с той же скоростью... | 2018-04-03T13:13:51 | 2018-04-03T13:14:18 | {} | NONE |
ClickHouse | ClickHouse | 378,283,244 | 2,074 | filimonov | Just to clarify: that issue is not connected with GDPR.
About GDPR: as far as i know some special syntax of real data deletion is planned. Hope that it will be delivered till may of 2018.
For now the only way to delete some rows from clickhouse is full rewrite of one partition. So create the table with the same ... | 2018-04-03T15:05:56 | 2018-04-03T15:06:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,300,201 | 2,165 | filimonov | Unfortunately currently CapnProto is supported only as input format:
https://github.com/yandex/ClickHouse/blob/master/dbms/src/DataStreams/FormatFactory.cpp | 2018-04-03T15:53:52 | 2018-04-03T15:53:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 377,958,518 | 217 | alexey-milovidov | How to reproduce:
```
SET max_execution_time = 3;
SELECT * FROM remote('127.0.0.{2,3}', system.numbers) WHERE number < 10;
``` | 2018-04-02T15:40:49 | 2018-04-02T15:40:49 | {} | MEMBER |
ClickHouse | ClickHouse | 378,002,364 | 217 | alexey-milovidov | Fixed in master. | 2018-04-02T18:27:09 | 2018-04-02T18:27:09 | {} | MEMBER |
ClickHouse | ClickHouse | 378,016,239 | 2,159 | alexey-milovidov | Could you please add also to `docs/en`? | 2018-04-02T19:18:51 | 2018-04-02T19:18:51 | {} | MEMBER |
ClickHouse | ClickHouse | 376,072,345 | 1,771 | dvahram | @tobad357 I believe not really, except the commit message which is not very detailed. If you want to try it out I can help you.
@alexey-milovidov also I can provide some docs on the functionality. Where do I commit that? To changelog? | 2018-03-26T07:32:41 | 2018-03-26T07:32:41 | {} | NONE |
ClickHouse | ClickHouse | 376,085,228 | 2,110 | silviucpp | @filimonov great :) I thought the features were removed :) | 2018-03-26T08:24:22 | 2018-03-26T08:24:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 376,168,585 | 2,112 | filimonov | But if you really need to do binary comparison, you can always force typecast, so if you want to look for exact Float32 value you can use something like ```WHERE f32 = toFloat32(10.2)```.
But still in most cases it's a very bad idea, as in some cases even same float can have multiple machine representation, which wi... | 2018-03-26T13:38:06 | 2018-03-26T13:44:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 377,639,570 | 2,146 | derekperkins | I am running version 1.1.54370 and I am seeing a similar issue, even without a tuple.
`SELECT field1 FROM foo WHERE (dateTimeField IN '2018-03-02 00:00:00') AND (SomeID=1234);`
returns no rows
`SELECT field1 FROM foo WHERE (dateTimeField = '2018-03-02 00:00:00') AND (SomeID=1234);`
returns correct rows
This ... | 2018-03-30T22:42:13 | 2018-03-30T22:42:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 377,809,866 | 1,633 | alexey-milovidov | @KochetovNicolai is working on this PR. | 2018-04-01T19:17:22 | 2018-04-01T19:17:22 | {} | MEMBER |
ClickHouse | ClickHouse | 377,964,622 | 2,156 | lamberken | @KochetovNicolai, here are demo, you can try.
if the query with *`limit 1`*, it'll throw error.
```
use default;
CREATE TABLE centerlog_shard ON CLUSTER datacenter
( date Date, time DateTime, appname String)
ENGINE = MergeTree(date, (time, appname), 8192);
CREATE TABLE centerlog ON CLUSTER datacent... | 2018-04-02T16:05:31 | 2018-04-02T16:18:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,421,567 | 2,161 | jonatasfreitasv | @filimonov @KochetovNicolai | 2018-04-03T22:37:44 | 2018-04-03T22:37:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 376,055,824 | 2,110 | filimonov | Very innacurate master merge before push. Lot of regressions | 2018-03-26T06:04:30 | 2018-03-26T06:04:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 376,070,823 | 1,771 | tobad357 | @alexey-milovidov @dvahram did this functionality get documented somewhere in the docs? | 2018-03-26T07:26:00 | 2018-03-26T07:26:00 | {} | NONE |
ClickHouse | ClickHouse | 376,728,386 | 2,096 | lamberken | @sundy-li，:+1: :+1:，太给力了，尝试使用下，期待发布到正式版。
但是这里有一个矛盾点，若分布式表指定了shard key，但实际是推荐往本地表写入的，那查询分布式表时，shard key还有作用吗？
| 2018-03-28T01:22:32 | 2018-03-28T01:27:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 376,743,129 | 1,925 | zhang2014 | Maybe we should continue : ) | 2018-03-28T02:49:54 | 2018-03-28T02:49:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 377,079,038 | 2,131 | alexey-milovidov | Излишнее логгирование:
```
LOG_DEBUG(log, "Will use existing table " << description << " for " + LogElement::name());
``` | 2018-03-29T00:10:41 | 2018-03-29T00:10:41 | {} | MEMBER |
ClickHouse | ClickHouse | 377,802,820 | 2,149 | alexey-milovidov | Reasonable. Although we assume that arrays in ClickHouse are generally small, you can have few large arrays as a result of calculations - just like in your example. | 2018-04-01T17:32:11 | 2018-04-01T17:32:11 | {} | MEMBER |
ClickHouse | ClickHouse | 378,269,756 | 1,770 | alexm93 | То есть uuid храните в CH как FixedString(16)?
А insert как делаете? Есть ли способ инсертить uuid как 36 символьную строку, чтобы clickhouse сам ее переводил в FixedString(16)? | 2018-04-03T14:28:36 | 2018-04-03T14:28:36 | {} | NONE |
ClickHouse | ClickHouse | 378,284,888 | 198 | filimonov | Clickhouse version 1.1.54356 (2018-03-06) introduce the `clickhouse-copier` tool for copying between clusters and resharding data.
Please, check documentation: https://clickhouse.yandex/docs/en/utils/clickhouse-copier/ | 2018-04-03T15:10:34 | 2018-04-03T15:10:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 376,163,091 | 2,112 | filimonov | But may be ClickHouse can implement `almost_equal` function (with machine epsilon) for floats.
For general usage it's not always nice to use such approximations, as they are not so cpu-efficient, and also can have some undesired side effects, one of them - loose transitive property. So if
almost_equal(a,b) and alm... | 2018-03-26T13:20:23 | 2018-03-26T13:20:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 377,309,792 | 2,139 | proller | https://github.com/pocoproject/poco/pull/2257
| 2018-03-29T17:24:31 | 2018-03-29T17:24:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 377,791,871 | 2,141 | jpiper | Ah, didn't even know that ClickHouse had DNS caching. Would be useful to disable this in situations where the active endpoint for a service (e.g. an external dictionaries endpoint) is managed via DNS with a short TTL. | 2018-04-01T14:48:43 | 2018-04-01T14:48:43 | {} | NONE |
ClickHouse | ClickHouse | 378,265,749 | 2,146 | derekperkins | @alex-zaitsev Sounds good, I just wanted to make sure this was on the list to get fixed in the next release. Thanks! | 2018-04-03T14:17:15 | 2018-04-03T14:17:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,300,678 | 2,165 | tobad357 | That would explain it
It seems like the documentation is a bit eager
https://clickhouse.yandex/docs/en/formats/capnproto/ | 2018-04-03T15:55:11 | 2018-04-03T15:55:11 | {} | NONE |
ClickHouse | ClickHouse | 376,569,972 | 2,125 | lamberken | set `null_value` 0
```
<null_value>0</null_value>
``` | 2018-03-27T15:33:46 | 2018-03-27T15:33:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 377,227,141 | 2,136 | proller | also not required now: https://github.com/yandex/ClickHouse/commit/62b27b539af552c41457cd9c16a18a0833476759 | 2018-03-29T12:59:19 | 2018-03-29T12:59:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 377,237,527 | 2,131 | alexey-milovidov | А теперь недостаточное логгирование (потеряно логгирование при старте). | 2018-03-29T13:37:40 | 2018-03-29T13:37:40 | {} | MEMBER |
ClickHouse | ClickHouse | 377,639,447 | 2,135 | derekperkins | I don't have anything to compare it to, but it does seem to be working quickly enough for a materialized view definition. Thanks @alex-zaitsev! | 2018-03-30T22:41:23 | 2018-03-30T22:41:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 377,822,406 | 2,153 | den-crane | Seems it's my environment issue (locale LANG=C... ), but clickhouse-client uses UTF8 characters.
anyway -U fixes the issue (screen -U).
Although clickhouse-client should not use UTF8 if a current locale does not support UTF (LANG=C) and raise an error or use ASCII-128 characters such as :
`+-----+--+`
`|col1 |aa|... | 2018-04-01T22:40:15 | 2018-04-01T23:19:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 377,894,845 | 2,154 | jpiper | I dug further, and the IP address in the container is different than the address that's exposed outside of the container, which is what's causing the issue.
I think a solution would be to implement #2155 properly (I don't know how to do it, unfortunately). | 2018-04-02T09:35:49 | 2018-04-02T09:35:49 | {} | NONE |
ClickHouse | ClickHouse | 378,131,367 | 2,163 | tobad357 | @vavrusa FYI, in case you are using this in production | 2018-04-03T05:19:51 | 2018-04-03T05:19:51 | {} | NONE |
ClickHouse | ClickHouse | 378,281,666 | 2,124 | antonio-antuan | Спасибо, разобрался | 2018-04-03T15:01:29 | 2018-04-03T15:01:29 | {} | NONE |
ClickHouse | ClickHouse | 378,282,065 | 1,770 | alexm93 | Спасибо, стало понятно. Я просто использовал JSONEachRow при вставке. При его использовании, как я понял, конструкцию UUIDStringToNum(текстовое представление UUID) не удастся вставить.
Придется переходить на другой формат. | 2018-04-03T15:02:35 | 2018-04-03T15:04:17 | {} | NONE |
ClickHouse | ClickHouse | 378,306,878 | 2,074 | filimonov | About GDPR: actually as far as i know at that point 'right to be forgotten' doesn't enforce to remove that data instantly, you need to have protocol describing that procedure, and up to 1 month to really remove the data. So i don't think that running OPTIMIZE FINAL will be a big issue during that month. | 2018-04-03T16:14:19 | 2018-04-03T16:14:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 376,009,954 | 2,100 | alexey-milovidov | Thanks!
BTW, Is it true that topic name cannot contain `,`? | 2018-03-25T22:43:43 | 2018-03-25T22:43:43 | {} | MEMBER |
ClickHouse | ClickHouse | 376,441,492 | 2,096 | sundy-li | @Jaxvk Sorry for the delay, If you are still looking for the windowFunnel problem solution. You may try the release with embedded funnelWindow function.
Usage as follows:
```
SELECT
sumIf(c, level >= 1) AS _1,
sumIf(c, level >= 2) AS _2,
sumIf(c, level >= 3) AS _3
FROM
(
SELECT
... | 2018-03-27T08:32:17 | 2018-03-27T08:32:17 | {
"+1": 3
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 376,602,844 | 2,084 | ztlpn | Fixed in master. | 2018-03-27T17:11:17 | 2018-03-27T17:11:17 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,272,381 | 1,770 | deonwise | тут много было про данный формат. | 2018-04-03T14:35:49 | 2018-04-03T14:35:49 | {} | NONE |
ClickHouse | ClickHouse | 375,951,717 | 2,107 | silviucpp | Hello @alexey-milovidov , I was looking via your pull request and was wonder if you can have multiple ZooKeeperImpl instances running in the same time ? And if yes if they grow with each replicated table you create . I'm asking because ZooKeeperImpl it's creating 2 threads per instance . In case it's one instance or a ... | 2018-03-25T07:41:03 | 2018-03-25T07:41:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 376,009,651 | 2,110 | alexey-milovidov | Thanks!
BTW, the files became executable. This is bad but tolerable. | 2018-03-25T22:39:53 | 2018-03-25T22:39:53 | {} | MEMBER |
ClickHouse | ClickHouse | 376,009,701 | 2,106 | alexey-milovidov | :+1: | 2018-03-25T22:40:33 | 2018-03-25T22:40:33 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 376,161,424 | 2,112 | silviucpp | @filimonov I know that comparing floats is not accurate but I'm expecting to work at least on the same limitations c++ has. Here problem I'm sure is not because of the way real numbers are represented and most probably is a bug in clickhouse somewhere. | 2018-03-26T13:15:10 | 2018-03-26T13:15:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 376,384,845 | 783 | linuxgood1230 | if data is None:
data = "NULL" | 2018-03-27T03:26:40 | 2018-03-27T03:26:40 | {} | NONE |
ClickHouse | ClickHouse | 377,191,864 | 2,136 | proller | ninja-build:
Conflicts: ninja
only ninja-build required | 2018-03-29T10:22:07 | 2018-03-29T10:22:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 377,795,744 | 2,149 | lamberken | ClickHouse is designed for analytics, may some table store trillion rows,
so the return type of countEqual is defined as UInt64 is better.
for example, you'll get zero.
SELECT countEqual(datas, 1)
FROM
(
SELECT groupArray(1) AS datas
FROM
(
SELECT 1
FROM system.numbers
LIMIT 4294967296
)
)
┌─countEqual(datas, 1)─┐
│ ... | 2018-04-01T15:47:05 | 2018-04-01T16:03:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 377,927,020 | 2,151 | KochetovNicolai | Yes, it's ok. | 2018-04-02T13:16:53 | 2018-04-02T13:16:53 | {} | MEMBER |
ClickHouse | ClickHouse | 377,994,765 | 2,156 | KochetovNicolai | @lamber-ken, thank you for demo. I've reproduced on cluster with 2 shards. | 2018-04-02T17:59:29 | 2018-04-02T17:59:29 | {} | MEMBER |
ClickHouse | ClickHouse | 378,197,593 | 2,161 | filimonov | Yep. I remember that for me all that engines was a bit confusing at the beginning too.
Generally: the **main** engine in Clickhouse is called [MergeTree](https://clickhouse.yandex/docs/en/table_engines/mergetree/). It allows to store and process data on one server and feel all the advantages of Clickhouse. Basic us... | 2018-04-03T10:04:17 | 2018-04-03T10:39:32 | {
"+1": 234,
"heart": 93,
"hooray": 26,
"laugh": 6,
"rocket": 19
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,323,029 | 2,161 | jonatasfreitasv | Thx a lot guys!!!
But in the case of using shard and replica at the same time, will not conflicts happen?
For the records will be splited between the servers but at the same time all parts will be replicated.
Or the correct would be to have, for example, 2 servers in replica, being one of these part of the sha... | 2018-04-03T17:02:20 | 2018-04-03T17:02:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,249,592 | 2,099 | VladRassokhin | Found another peculiar behavior which could be potentially optimized:
```text
:) SELECT count() from (SELECT DISTINCT id FROM test.data) FORMAT TSV;
1024
1 rows in set. Elapsed: 0.852 sec. Processed 1.07 billion rows, 4.29 GB (1.26 billion rows/s., 5.04 GB/s.)
:) SELECT uniqExact(id) FROM test.data FORMAT TSV;
10... | 2018-04-03T13:27:32 | 2018-04-03T17:15:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,373,405 | 2,093 | k4otix | This would be great. My use case is doing range queries for IP addresses but I cannot use `ip_trie` since many of the ranges cannot be represented as CIDR blocks, so I would want to use the FixedString(16) representation resulting from `IPv6StringToNum()`.
However, it would be nice for any ordered type to be able t... | 2018-04-03T19:47:44 | 2018-04-03T19:47:44 | {} | NONE |
ClickHouse | ClickHouse | 376,067,973 | 2,110 | silviucpp | Hello guys,
Are you sure about : "In contrast to the native interface, the HTTP interface does not support the concept of sessions or session settings, does not allow aborting a query (to be exact, it allows this in only a few cases)". I mean was this functionality removed ? because I'm using session_id, session_tim... | 2018-03-26T07:13:02 | 2018-03-26T07:13:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 377,063,504 | 2,135 | alex-zaitsev | Storing tags in comma separated strings is not that slow comparing to array. You can splitByChar or use match() function if you search by tags. Since you can filter dictionary first it is not slow at all. | 2018-03-28T22:49:31 | 2018-03-28T22:49:31 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 377,201,545 | 2,136 | kshvakov | fixed | 2018-03-29T11:06:58 | 2018-03-29T11:06:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 377,543,126 | 2,146 | lamberken | According to your steps，I test `ClickHouse server version 1.1.54343` and `ClickHouse server version 1.1.54343`
| version | result |
| - | -: |
| 1.1.54370 | wrong |
| 1.1.54343 | right |
| 2018-03-30T15:01:14 | 2018-03-30T15:01:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,028,507 | 2,115 | silviucpp | `SELECT has([null, 'str1', 'str2'], 'str1')` works as well. | 2018-04-02T20:05:42 | 2018-04-02T20:05:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,204,594 | 2,074 | erdebee | A problem you are not regarding here is that merging only happens when OPTIMIZE TABLE is being executed... until that point you still have 2 rows, 1 marked with deleted and the other not.
To comply with GDPR your data doesnt just have to be marked as deleted, it really has to be gone, you can never "delete" your pri... | 2018-04-03T10:32:19 | 2018-04-03T10:32:19 | {} | NONE |
ClickHouse | ClickHouse | 376,155,747 | 1,882 | KochetovNicolai | Improves for long strings. Before:
```SQL
SELECT count()
FROM system.numbers
WHERE NOT ignore(lengthUTF8(materialize('https://zh.wikipedia.org/wiki/%E4%B8%AD%E6%97%A5%E9%9F%93%E7%B5%B1%E4%B8%80%E8%A1%A8%E6%84%8F%E6%96%87%E5%AD%97%E6%93%B4%E5%B1%95%E5%8D%80F')))
↑ Progress: 223.87 million rows, 1.79 GB (10.90 mi... | 2018-03-26T12:55:57 | 2018-03-26T12:55:57 | {} | MEMBER |
ClickHouse | ClickHouse | 377,476,522 | 2,143 | lamberken | It's friendly for readers | 2018-03-30T08:25:37 | 2018-03-30T08:25:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,561,965 | 1,770 | alexm93 | А вы в каком формате делаете инсерт?
Я пытаюсь с помощью обычного VALUES вставлять
Если вставлять по 100 строк, то все ок
Если по 1000 и больше начинают валиться подобные ошибки:
DB::Exception: Unknown identifier: UUIDStringToNum, e.what() = DB::Exception
Или
DB::Exception: Unknown identifier: UUIDStrin, e.what()... | 2018-04-04T11:01:16 | 2018-04-04T11:01:16 | {} | NONE |
ClickHouse | ClickHouse | 376,080,262 | 2,110 | filimonov | @silviucpp it looks like regression of #2037 (and few other doc commits) caused by inaccurate conflict resolution during merge. | 2018-03-26T08:05:37 | 2018-03-26T08:05:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 376,156,532 | 2,112 | filimonov | The reason is that for floats which appear in query ClickHouse uses Float64 by default.
```
:) select toTypeName(10.2);
┌─toTypeName(10.2)─┐
│ Float64 │
└──────────────────┘
```
So 10.2 in Float64 from query exactly (binary) equal to 10.2 in Float64 in table.
But for 10.2 (in Float64) != 10.2 (in... | 2018-03-26T12:58:41 | 2018-03-26T12:58:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 377,079,721 | 2,127 | alexey-milovidov | You are relying on an assumption that `is_cancelled` flag in processlist is set only when the query was cancelled by `replace_running_query`. This is non obvious for me. | 2018-03-29T00:14:51 | 2018-03-29T00:14:51 | {} | MEMBER |
ClickHouse | ClickHouse | 377,775,330 | 2,149 | zhang2014 | The maximum length of an array cannot greater then `UIn32::max_value` ? : ) | 2018-04-01T09:44:59 | 2018-04-01T09:44:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,201,967 | 2,161 | KochetovNicolai | Hi,
In order to enable replication you can use tables from [Replicated](https://clickhouse.yandex/docs/en/table_engines/replication/#replicatedmergetree) family. You can add `Replicated` to any `*MergeTree` engine and specify znode and replica name in two extra parameters.
You should have access to external zookeep... | 2018-04-03T10:21:27 | 2018-04-03T10:21:27 | {
"+1": 11,
"heart": 6
} | MEMBER |
ClickHouse | ClickHouse | 378,228,864 | 2,162 | jonatasfreitasv | To work, it is mandatory that there are two or more ClickHouse servers. | 2018-04-03T12:14:42 | 2018-04-03T12:14:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,285,109 | 1,169 | filimonov | Clickhouse version 1.1.54356 (2018-03-06) introduce the `clickhouse-copier` tool for copying between clusters and resharding data.
Please, check documentation: https://clickhouse.yandex/docs/en/utils/clickhouse-copier/ | 2018-04-03T15:11:12 | 2018-04-03T15:11:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 376,022,316 | 2,100 | sundy-li | @alexey-milovidov the rule is as follow:
https://github.com/apache/kafka/blob/trunk/clients/src/main/java/org/apache/kafka/common/internals/Topic.java#L77 | 2018-03-26T01:34:47 | 2018-03-26T01:34:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 376,073,105 | 1,771 | tobad357 | @dvahram so if I understand correctly for a mysql dictionary we would add the update field and
then each time CK checks for updates it will use the update_field to only select rows from the last update?
```
<update_field>updated</update_field>
```
becomes
```
select col1, col2, col3 from dict_table where upda... | 2018-03-26T07:36:07 | 2018-03-26T07:36:07 | {} | NONE |
ClickHouse | ClickHouse | 377,907,340 | 2,154 | jpiper | So to summarise:
Internal IP (in the container) looks like `172.16.1.2`
However, with docker the ports 9000, 9009, and 8123 on the host machine (`10.1.2.3`) are mapped inside.
The remote server uses the DNS name of the host machine, which resolves to `10.1.2.3`, and then `isLocalAddress` fails as it is comparing... | 2018-04-02T11:00:50 | 2018-04-02T11:00:50 | {} | NONE |
ClickHouse | ClickHouse | 378,307,770 | 2,165 | filimonov | > It seems like the documentation is a bit eager
Yep. May be @vavrusa was planning to add that as an output format too? It doesn't sound like a big issue. | 2018-04-03T16:17:01 | 2018-04-03T16:17:01 | {} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.