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 | 432,184,002 | 3,447 | alexey-milovidov | As far as I know, `pointInEllipses` works on plane, so the order of arguments and their units are insignificant. This is documentation issue. | 2018-10-23T10:05:37 | 2018-10-23T10:05:37 | {} | MEMBER |
ClickHouse | ClickHouse | 432,208,937 | 3,436 | amosbird | @alexey-milovidov I've rewrite almost all call sites of Field constructions, leaving some non trivial conversions intact.
reinterpret_cast generates the same instructions as static_cast for signed and unsigned conversion
https://godbolt.org/z/wurrxV
The main non-obvious part is accidentally place floats into i... | 2018-10-23T11:29:50 | 2018-10-23T11:29:50 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 432,312,894 | 2,014 | alexey-milovidov | No. | 2018-10-23T16:10:14 | 2018-10-23T16:10:14 | {} | MEMBER |
ClickHouse | ClickHouse | 432,807,125 | 1,567 | valyala | @KochetovNicolai , @alexey-milovidov , as I understand this feature has been already implemented | 2018-10-24T20:00:30 | 2018-10-24T20:00:30 | {} | NONE |
ClickHouse | ClickHouse | 432,937,071 | 3,372 | BayoNet | Исправил. | 2018-10-25T06:59:47 | 2018-10-25T06:59:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 433,404,171 | 3,333 | filimonov | avg always return Float64 (the only exception is Decimal):
```select toTypeName( arrayReduce('avg',[0]) );```
It's here:
https://github.com/yandex/ClickHouse/blob/47dd1730072d98ff2d598edbbabb3d108a37373d/dbms/src/AggregateFunctions/AggregateFunctionAvg.h#L47
| 2018-10-26T13:17:52 | 2018-10-26T13:17:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 433,453,151 | 3,411 | ztlpn | The bug appeared when data was inserted into a Distributed table in the Native format. It is fixed in master and in the upcoming v18.14.11 fix-release. | 2018-10-26T15:46:14 | 2018-10-26T15:46:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 433,461,775 | 3,472 | filimonov | > А вот ALTER TABLE UPDATE работает на удивление быстро.
Если условие для UPDATE позволяет выбрать маленькое количество частей то будет быстро. В общем случае это может быть эквивалентно перезаписыванию всех затронутых колонок во всей таблице. | 2018-10-26T16:13:34 | 2018-10-26T16:13:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 433,877,545 | 3,470 | ghuname | How come that 1970-01-01 is not supported? It is zero epoch. | 2018-10-29T11:31:25 | 2018-10-29T11:31:25 | {} | NONE |
ClickHouse | ClickHouse | 433,971,034 | 3,470 | filimonov | Caused by the fact that when it was 1970-01-01 in UTC in some other timezones it still was 1969 which can'be properly processed. Accurate processing of that date is possible but AFIAK it was just simpler to assume that epoch starts a little bit later, and no extra costs on processing border effects of first day of epo... | 2018-10-29T16:08:45 | 2018-10-29T16:09:29 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 433,167,041 | 3,411 | denisov-vlad | 1. Yes.
2. No
3. Same; no; they were recently created.
4. Yes.
5. I can show you config-preprocessed file via telegram.
6. No.
7. I'll send you a message through telegram. | 2018-10-25T18:58:02 | 2018-10-25T18:58:21 | {} | NONE |
ClickHouse | ClickHouse | 433,391,406 | 3,456 | BenoitSa | Ok. I get it.
While `datatime `is just a timestamp and does not contain time zone dependent information, the type (of the result-set column) contains the time zone. Thus `now()` and `toDateTime(now(),'America/New_York')` have the same value, are considered as equal, but not the same type, thus not displayed the same... | 2018-10-26T12:30:07 | 2018-10-26T15:26:39 | {} | NONE |
ClickHouse | ClickHouse | 433,650,679 | 3,480 | dorantor | Forgot to mention. If this errors happens for some time(as we had) then you'll get segfaults in threads and after some more time server will silently close. This makes me think that memory leaks are very possible. | 2018-10-27T19:57:46 | 2018-10-27T19:57:46 | {} | NONE |
ClickHouse | ClickHouse | 433,861,905 | 3,162 | mlex | Hi,
we were able to create a small example to reproduce the bug. You need a cluster of at least 2 nodes.
```
CREATE TABLE local_test ON CLUSTER testcluster (
t UInt64,
date Date MATERIALIZED toDate(t/1000),
shard UInt64,
col1 String,
col2 String
) ENGINE = MergeTree
PARTITION BY toRelativeDayNum... | 2018-10-29T10:34:23 | 2018-10-29T13:37:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 433,946,542 | 3,487 | alesapin | Failed test is `stateless.00747_contributors`. | 2018-10-29T15:07:56 | 2018-10-29T15:07:56 | {} | MEMBER |
ClickHouse | ClickHouse | 433,968,003 | 3,480 | alesapin | Fixed in https://github.com/yandex/ClickHouse/pull/3487. | 2018-10-29T16:00:51 | 2018-10-29T16:00:51 | {} | MEMBER |
ClickHouse | ClickHouse | 432,444,409 | 3,445 | chripede | Syslog is being spammed with messages like this while ClickHouse is at 100%
```
[46472.561964] netlink: 'BackgrProcPool': attribute type 1 has an invalid length.
[46487.811097] netlink: 'HTTPHandler': attribute type 1 has an invalid length.
[46487.811668] netlink: 'HTTPHandler': attribute type 1 has an invalid le... | 2018-10-23T22:26:38 | 2018-10-23T22:26:38 | {} | NONE |
ClickHouse | ClickHouse | 433,174,816 | 3,468 | den-crane | try
set max_bytes_before_external_group_by = 10000000000;
max_bytes_before_external_sort, max_bytes_before_external_group_by
https://clickhouse.yandex/docs/en/query_language/select/
| 2018-10-25T19:24:01 | 2018-10-25T19:24:01 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 433,711,561 | 3,424 | bgranvea | Very strange: if the table contains the column `foo` from the beginning, I can drop and re-add it:
```
CREATE TABLE test (ts DATETIME,`foo.bar` DOUBLE, `foo` DOUBLE) ENGINE=MergeTree PARTITION BY toStartOfDay(ts) ORDER BY tuple()
Ok
alter table test drop column `foo`
Ok
alter table test add column `foo` DOU... | 2018-10-28T14:41:19 | 2018-10-28T14:41:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 433,929,210 | 3,477 | ztlpn | Thanks! A note for your future pull requests: local variable should be named in snake_case: https://clickhouse.yandex/docs/en/development/style/#names | 2018-10-29T14:24:20 | 2018-10-29T14:24:20 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 432,172,899 | 3,445 | chripede | @blinkov Of course
```
Thread 33 (Thread 0x7f9f491fc700 (LWP 2038)):
#0 0x00007fa051aa5f85 in futex_abstimed_wait_cancelable (private=<optimized out>, abstime=0x7f9f491f8c40, expected=0, futex_word=0x7f... | 2018-10-23T09:34:03 | 2018-10-23T09:34:03 | {} | NONE |
ClickHouse | ClickHouse | 432,571,183 | 3,410 | alesapin | We disable compile expressions by default in 14.10. | 2018-10-24T08:52:24 | 2018-10-24T08:52:24 | {} | MEMBER |
ClickHouse | ClickHouse | 433,107,640 | 3,415 | 4ertus2 | Возможно исправлено тут https://github.com/yandex/ClickHouse/pull/3460 | 2018-10-25T15:57:37 | 2018-10-25T15:57:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 433,400,708 | 3,414 | filimonov | Workaround is simple:
```sql
SELECT
if ( countIf(condition)>0, <aggregatefunction>If(..., condition), default_value)
``` | 2018-10-26T13:05:49 | 2018-10-26T13:05:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 433,481,456 | 3,198 | fokkoru | ```
2018.10.26 17:11:45.063142 [ 616383 ] {} <Error> void DB::ThreadStatus::initPerformanceCounters(): Code: 412, e.displayText() = DB::Exception: Can't create PF_NETLINK socket, errno: 24, strerror: Too many open files, e.what() = DB::Exception, Stack trace:
0. /usr/bin/clickhouse-server(StackTrace::StackTrace()+0... | 2018-10-26T17:20:32 | 2018-10-26T17:20:32 | {} | NONE |
ClickHouse | ClickHouse | 433,592,914 | 2,609 | silviucpp | @alexey-milovidov do you think it is possible to resume work and complete this PR ? | 2018-10-27T05:35:39 | 2018-10-27T05:35:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 433,917,635 | 3,479 | alesapin | It's strange case for using ClickHouse, when all your data can fit into memory.
ClickHouse doesn't cache compressed data from MergeTree, it relies on OS page cache. But if you really want to load all your data into memory you can place your clickhouse data directory into [tmpfs](https://en.wikipedia.org/wiki/Tmpfs)... | 2018-10-29T13:52:51 | 2018-10-29T13:52:51 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 434,260,173 | 3,162 | mikeeremin | Confirm for 18.14.11 - not fixed. | 2018-10-30T11:11:06 | 2018-10-30T11:11:06 | {} | NONE |
ClickHouse | ClickHouse | 432,365,722 | 3,450 | TH-HA | Fyi , i ve tried another view definition to avoid the join ( using the operator arrayJoin instead ) with the same result
```
CREATE MATERIALIZED VIEW test.mv (
date_time DateTime('UTC'),
site String,
lan_byte AggregateFunction(sum, UInt64),
cnt AggregateFunction(count, UInt8)
) ENGINE = AggregatingMe... | 2018-10-23T18:28:46 | 2018-10-23T18:29:25 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 433,014,100 | 3,461 | ztlpn | 👍 | 2018-10-25T11:25:02 | 2018-10-25T11:25:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 433,163,630 | 3,411 | ztlpn | @denisov-vlad Thanks, but still no luck.
Some additional questions:
1. Are the local tables replicated (I assume they are, because internal_replication is true)
2. Are there any materialized views attached to local/distributed tables?
3. Is the set of columns the same for local and distributed tables? Do the tabl... | 2018-10-25T18:47:33 | 2018-10-25T18:57:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 433,474,175 | 3,292 | ztlpn | FinishSortingBlockInputStream is intended as a building block for implementing ORDER BY clause execution that uses the ordering in MergeTree parts. Your example (at least for now) will just be sorted twice. | 2018-10-26T16:54:48 | 2018-10-26T16:54:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 434,203,801 | 3,488 | AlexanderJLiu | @alesapin This is awsome, but the clickhouse version doesn't support `update`. | 2018-10-30T07:52:56 | 2018-10-30T07:52:56 | {} | NONE |
ClickHouse | ClickHouse | 434,214,198 | 3,490 | lapkofear | It seems, that this bug is reproducing on all 16 bit types, because next queries executes the same:
```
SELECT 1 FROM test.perf_test GROUP BY toUInt16(1)
SELECT 1 FROM test.perf_test GROUP BY toUInt32(1)
```
First one up to 20s, second one almost immediately. | 2018-10-30T08:34:39 | 2018-10-30T08:34:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 434,228,326 | 3,488 | alesapin | It was introduced in 18.12.14. You have to update, current stable is 18.14.11. | 2018-10-30T09:24:32 | 2018-10-30T09:29:54 | {} | MEMBER |
ClickHouse | ClickHouse | 434,274,742 | 3,478 | 4ertus2 | There was analogous crash in another part of Aggregator. The reason was aligned instructions in gcc release build with unaligned data in Aggregator. | 2018-10-30T12:05:27 | 2018-10-30T12:05:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 432,576,564 | 2,684 | CompadreP | @Deninc We decided to use postgres for this case because it handles this type of workload much better | 2018-10-24T09:06:50 | 2018-10-24T09:06:50 | {} | NONE |
ClickHouse | ClickHouse | 433,890,877 | 3,162 | alesapin | Can you provide your cluster config or test it yourself on 18.14.11? This error seems to be fixed. Currently I can't reproduce it. | 2018-10-29T12:25:30 | 2018-10-29T12:25:30 | {} | MEMBER |
ClickHouse | ClickHouse | 434,027,727 | 3,195 | ztlpn | Also broke the example clickhouse-client usage from the README:
```
$ docker run -it --rm --link some-clickhouse-server:clickhouse-server yandex/clickhouse-client:18.14.11 --host clickhouse-server
docker: Error response from daemon: OCI runtime create failed: container_linux.go:348: starting container process caused... | 2018-10-29T18:30:37 | 2018-10-29T18:35:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 432,167,659 | 3,445 | blinkov | @chripede could you run and post the results of:
* `sudo perf top -p $(pidof clickhouse-server)`
* `sudo gdb -p $(pidof clickhouse-server)` + `thread apply all bt` | 2018-10-23T09:20:08 | 2018-10-23T09:20:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 432,587,845 | 3,450 | TH-HA | Seems to be related to #127 and #448 except it's on one single input table.
Can we help to make progress on this issue ? Materialized View combine with state are a killer feature , by solving this (especially since you can process MV concurrently ) the possibility are just enormous in term of automated and efficien... | 2018-10-24T09:40:43 | 2018-10-24T09:40:43 | {} | NONE |
ClickHouse | ClickHouse | 432,638,626 | 3,441 | proller | this brokes macos build:
```
FAILED: dbms/CMakeFiles/dbms.dir/src/Client/Connection.cpp.o
/usr/local/bin/ccache /usr/local/bin/g++-8 -DBOOST_SYSTEM_NO_DEPRECATED -DLZ4_DISABLE_DEPRECATE_WARNINGS=1 -DPOCO_STATIC -DPOCO_UNBUNDLED_ZLIB -DUNALIGNED_OK -DWITH_GZFILEOP -DX86_64 -DZLIB_COMPAT -isystem ../contrib/libspars... | 2018-10-24T12:40:50 | 2018-10-24T12:40:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 433,326,798 | 3,468 | PetarLotrean | Great, thanks! This solves the failing issue, the query isn't failing anymore.
But still, it uses a huge amount of disc space, and the query is much slower because of writing/reading, and it doesn't have to be. | 2018-10-26T08:16:31 | 2018-10-26T08:16:31 | {} | NONE |
ClickHouse | ClickHouse | 433,371,865 | 3,456 | filimonov | It's not a bug, it's a feature :)
Important information here is that DateTime can have the optional parameter - timezone used to output the time, instead of the system default.
So you can write something like
```CREATE TABLE test (id UInt8, timestamp DateTime('UTC')) Engine = Log;```
and timestamp will be ... | 2018-10-26T11:03:25 | 2018-10-26T11:04:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 432,299,739 | 3,450 | TH-HA | I ve just realized that a 18.14.9 is available. I'll try it to see if some fixes are relevant for me.
| 2018-10-23T15:39:17 | 2018-10-23T15:39:17 | {} | NONE |
ClickHouse | ClickHouse | 433,379,076 | 3,472 | BooBSD | @filimonov Спасибо.
Правильно ли я понимаю, что, если мне нужны точечные апдейты по первичному ключу, то после втавки исправленной строки в `ReplacingMergeTree`, можно сразу же сделать `SELECT FINAL` по этому ключу и все остальные `SELECT` уже можно использовать без `FINAL`, и данные уже будут актуальными (без дубле... | 2018-10-26T11:36:55 | 2018-10-26T11:44:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 433,405,890 | 3,468 | PetarLotrean | Ok, great, that's what I needed to hear, thanks! | 2018-10-26T13:23:42 | 2018-10-26T13:23:42 | {} | NONE |
ClickHouse | ClickHouse | 434,200,964 | 2,198 | AlexMAS | Is there any progress for this issue? :) | 2018-10-30T07:41:22 | 2018-10-30T07:41:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 434,202,219 | 3,478 | grigory51 | Sorry, i try again and everything okey.
Found this logs in clickhouse-server.err.log
```
2018.10.27 09:38:08.384072 [ 2 ] {} <Error> BaseDaemon: ########################################
2018.10.27 09:38:08.384147 [ 2 ] {} <Error> BaseDaemon: (from thread 23) Received signal Segmentation fault (11).
2018.10.27 09... | 2018-10-30T07:46:36 | 2018-10-30T07:46:36 | {} | NONE |
ClickHouse | ClickHouse | 434,202,764 | 3,488 | AlexanderJLiu | > Dropping the column and adding the new one with the same name should work.
@filimonov Thanks，but the data already existed would be lost. This would be the last option I choose. 😭 | 2018-10-30T07:48:42 | 2018-10-30T07:49:54 | {} | NONE |
ClickHouse | ClickHouse | 432,176,193 | 2,014 | leif-larsen | Has this been implemented yet? | 2018-10-23T09:43:20 | 2018-10-23T09:43:20 | {} | NONE |
ClickHouse | ClickHouse | 433,919,965 | 3,162 | mlex | The bug is still there with the clickhouse docker image 18.14.11.
My cluster.xml (for node-1):
```
<yandex>
<macros>
<cluster>testcluster</cluster>
<shard>1</shard>
</macros>
<remote_servers>
<testcluster>
<shard>
<replica>
... | 2018-10-29T13:59:23 | 2018-10-29T13:59:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 434,163,583 | 3,494 | blinkov | In ClickHouse only single quotes can be used for string literals. Double quotes work the same as backticks, for identifiers. | 2018-10-30T03:58:03 | 2018-10-30T03:58:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 432,302,930 | 3,450 | TH-HA | switching to 18.14.9 don't change the behavior. | 2018-10-23T15:46:20 | 2018-10-23T15:46:20 | {} | NONE |
ClickHouse | ClickHouse | 432,448,711 | 2,156 | alexey-milovidov | Fixed in master. | 2018-10-23T22:46:08 | 2018-10-23T22:46:08 | {} | MEMBER |
ClickHouse | ClickHouse | 433,043,162 | 3,292 | nvartolomei | Keeping in mind how columns are resolved during query parsing, the following query might change it's behaviour if FinishSortingBlockInputStream is plumbed as is.
```sql
select -a as a, b
from (
select a, b
from (
select number as a, rand() as b from system.numbers limit 10
... | 2018-10-25T13:07:17 | 2018-10-25T13:07:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 433,148,589 | 3,411 | denisov-vlad | @ztlpn
engine = ReplicatedMergeTree
cluster:
```
<CLUSTER_NAME>
<shard>
<internal_replication>true</internal_replication>
<weight>0</weight>
<replica>
<host>hdd1</host>
<port>9000</port>
</replica>
<replica>
<host>hdd2</host>
... | 2018-10-25T18:00:52 | 2018-10-25T18:00:52 | {} | NONE |
ClickHouse | ClickHouse | 433,375,997 | 3,472 | filimonov | Лучше не использовать, ни то ни другое.
ReplacingMergeTree обновит данные "когда-нибудь" в процессе фоновых слияний, а может и совсем не обновит. Нужно будет всегда использовать FINAL или постоянно запускать OPTIMIZE.
ALTER TABLE UPDATE обновит точно но будет работать долго, мучительно перемалывая все данные.
... | 2018-10-26T11:22:49 | 2018-10-26T11:22:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 433,458,392 | 3,410 | alesapin | Seems to be fixed in master https://github.com/yandex/ClickHouse/pull/3457. But we will test this feature more carefully. | 2018-10-26T16:02:53 | 2018-10-26T16:02:53 | {} | MEMBER |
ClickHouse | ClickHouse | 433,858,293 | 3,481 | alesapin | Thanks! | 2018-10-29T10:21:46 | 2018-10-29T10:21:46 | {} | MEMBER |
ClickHouse | ClickHouse | 434,049,821 | 3,292 | nvartolomei | @ztlpn I was hoping this is a building block for #1344 :)
| 2018-10-29T19:37:15 | 2018-10-29T19:37:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 434,244,075 | 3,463 | filimonov | > Isn't it more correct ?
>
> ```dockerfile
> ADD https://github.com/tianon/gosu/releases/download/${gosu_ver}/gosu-amd64 /bin/gosu
> ```
Can you test it and create pull request? | 2018-10-30T10:15:42 | 2018-10-30T10:15:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 432,588,394 | 1,460 | Tasselmi | > Just use [symbolic links](https://en.wikipedia.org/wiki/Symbolic_link).
>
> Suppose you have `<path>/opt/clickhouse/</path>`. Then create symbolic links in `/opt/clickhouse/data/` for every db you need:
>
> ```
> ln -s /mnt/ssd/fastdb /opt/clickhouse/data/fastdb
> ln -s /mnt/hdd/slowdb /opt/clickhouse/data/sl... | 2018-10-24T09:42:25 | 2018-10-24T09:42:25 | {} | NONE |
ClickHouse | ClickHouse | 432,629,223 | 3,458 | blinkov | @sundy-li @zhang2014 | 2018-10-24T12:08:51 | 2018-10-24T12:08:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 432,938,221 | 3,113 | BayoNet | I rewrote the description of 'on_duplicate_clause', it should become more clear.
https://clickhouse.yandex/docs/en/operations/table_engines/mysql/ | 2018-10-25T07:04:59 | 2018-10-25T07:04:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 433,879,767 | 3,472 | ztlpn | @BooBSD В процессе апдейта создаётся новый кусок, содержащий изменённые данные старого. Если в нём изменяется хотя бы одна строка, то изменяемые столбцы будут полностью перезаписаны. Остальные столбцы перезаписаны не будут (файлы столбцов в новом куске будут хардлинками к файлам столбцов в старом куске). Также не будет... | 2018-10-29T11:40:56 | 2018-10-29T12:17:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 433,915,516 | 3,482 | alesapin | Can you provide some self-contained example to reproduce the problem? I think you have zeros in your `GROUP BY` columns, but even with them I cannot reproduce this behaviour. | 2018-10-29T13:46:47 | 2018-10-29T13:46:47 | {} | MEMBER |
ClickHouse | ClickHouse | 432,537,051 | 2,788 | millin | @vavrusa v.18.14.10. Issue still not fixed...
```log
2018.10.24 09:56:08.324216 [ 34 ] {} <Error> void DB::StorageKafka::streamThread(): Code: 60, e.displayText() = DB::Exception: Table stat.kafka_log doesn't exist., e.what() = DB::Exception, Stack trace:
0. clickhouse-server(StackTrace::StackTrace()+0x16) [0x56... | 2018-10-24T07:01:36 | 2018-10-24T07:02:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 432,537,220 | 2,684 | Deninc | Hi @CompadreP, have you done any further test in this regard? | 2018-10-24T07:02:16 | 2018-10-24T07:02:16 | {} | NONE |
ClickHouse | ClickHouse | 433,216,040 | 3,450 | alex-zaitsev | Could be the same as #2878 | 2018-10-25T21:43:56 | 2018-10-25T21:43:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 433,399,866 | 3,468 | filimonov | >The thing is, I can imagine a workflow where this could work
Right now ClickHouse has very primitive query optimizer, and there are a lot of optimizations possible. You can contribute the project to make it better. Or if you really need to get it done that way right now - you can do that optimization 'by hand': yo... | 2018-10-26T13:02:47 | 2018-10-26T13:02:47 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 433,895,893 | 3,469 | ztlpn | Confirmed. Internal detail: `CSVRowInputStream::readPrefix()` is called twice. | 2018-10-29T12:43:44 | 2018-10-29T12:43:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 433,966,136 | 3,488 | alesapin | You can use new `ALTER UPDATE` feature. Example:
```
:) CREATE TABLE default.some_table ( client_id Int32, some_column String) ENGINE = MergeTree ORDER BY tuple() SETTINGS index_granularity = 8192
:) INSERT INTO some_table VALUES (1, '1.0'), (2, '3.6'), (3, ''), (5, '172');
:) SELECT * FROM some_table;
... | 2018-10-29T15:56:11 | 2018-10-29T15:56:11 | {
"+1": 6
} | MEMBER |
ClickHouse | ClickHouse | 434,214,115 | 3,463 | spyzhov | Isn't it more correct ?
```Dockerfile
ADD https://github.com/tianon/gosu/releases/download/${gosu_ver}/gosu-amd64 /bin/gosu
``` | 2018-10-30T08:34:22 | 2018-10-30T08:34:22 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 432,938,989 | 1,744 | BayoNet | I added the warning about such of behaviour in docs.
https://clickhouse.yandex/docs/en/query_language/functions/date_time_functions/#tostartofmonth | 2018-10-25T07:08:42 | 2018-10-25T07:08:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 432,955,951 | 3,411 | MykhailoSobol | We are having the same issue.
ClickHouse server version 18.14.10 revision 54409
ubuntu 18.04. | 2018-10-25T08:11:55 | 2018-10-25T08:11:55 | {} | NONE |
ClickHouse | ClickHouse | 433,921,550 | 3,467 | alesapin | Unfortunately this information available for MergeTree family https://github.com/yandex/ClickHouse/blob/master/dbms/src/Storages/System/StorageSystemColumns.cpp#L109-L120 only. | 2018-10-29T14:03:42 | 2018-10-29T14:03:42 | {} | MEMBER |
ClickHouse | ClickHouse | 433,922,505 | 3,463 | alesapin | It's used inside [entrypoint.sh](https://github.com/yandex/ClickHouse/blob/master/docker/server/entrypoint.sh#L34) instead of `sudo`. | 2018-10-29T14:06:32 | 2018-10-29T14:06:32 | {} | MEMBER |
ClickHouse | ClickHouse | 432,403,679 | 3,453 | alexey-milovidov | This is 2000th merged pull request, congrats! | 2018-10-23T20:19:08 | 2018-10-23T20:19:08 | {} | MEMBER |
ClickHouse | ClickHouse | 433,120,962 | 3,411 | ztlpn | Somehow I can't reproduce the issue, but judging from the number of upvotes it should be easy. Can anyone affected share their cluster configuration (`<remote_servers>`, what kind of ENGINE local tables use)? | 2018-10-25T16:36:16 | 2018-10-25T16:36:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 433,327,845 | 3,462 | BayoNet | It looks that there is no something to do for me :-). I'll add this code block into the Russian part of the docs. | 2018-10-26T08:20:30 | 2018-10-26T08:20:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 433,465,461 | 3,472 | ztlpn | Да, всё верно - FINAL выкидывает старые версии на ходу, на диске он ничего не меняет. Основной недостаток в том, что он выполняет запрос в 1 поток, отсюда и замедление. OPTIMIZE эквивалентен перезаписи вообще всех данных в партиции. ALTER UPDATE перезаписывает только изменяющиеся столбцы в изменяющихся кусках, что тоже... | 2018-10-26T16:25:32 | 2018-10-26T16:25:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 433,834,415 | 3,476 | zhang2014 | 👍 | 2018-10-29T09:01:49 | 2018-10-29T09:01:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 433,930,290 | 3,488 | filimonov | Dropping the column and adding the new one with the same name should work. | 2018-10-29T14:27:09 | 2018-10-29T14:27:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 432,186,979 | 1,128 | CurtizJ | Fixed in master #3404. | 2018-10-23T10:15:11 | 2018-10-23T10:15:11 | {} | MEMBER |
ClickHouse | ClickHouse | 432,392,239 | 3,454 | blinkov | I think it's about front page, not documentation itself.
There's a second link to documentation at "Quick start" section near the bottom, so it's just two taps if you know what you are doing. Maybe we could add a third link as some sort of icon on first screen in narrow mode, but there's not much room and it'll dis... | 2018-10-23T19:46:12 | 2018-10-23T19:46:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 434,240,760 | 3,480 | dorantor | Awesome! Thanks! | 2018-10-30T10:04:42 | 2018-10-30T10:04:42 | {} | NONE |
ClickHouse | ClickHouse | 432,156,548 | 3,444 | alexey-milovidov | Yes, just write `FROM system.one`
https://clickhouse.yandex/docs/en/operations/system_tables/#systemone | 2018-10-23T08:50:13 | 2018-10-23T08:50:13 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 432,744,791 | 3,198 | fokkoru | I have kind of the same issue. But my logs have huge numbers of error.
I have Ubuntu 18.04.1, Clickhouse 18.14.9 revision 54409, instaled without docker from packages, 3
Kafka nodes.
```
2018.10.24 12:38:08.038702 [ 22 ] {} <Error> r0.events (ReplicatedMergeTreeRestartingThread): void DB::ReplicatedMergeTreeRe... | 2018-10-24T17:06:59 | 2018-10-24T17:06:59 | {} | NONE |
ClickHouse | ClickHouse | 433,373,803 | 3,470 | filimonov | the same:
```
select toDateTime('1969-12-29 00:00:01') ; -- 1970-01-01 06:28:17
select toDateTime('1008-02-29 00:00:01') ; -- 1970-01-01 06:28:17
```
Generally those dates are out of clickhouse supported range, but probably should be printed as '0000-00-00 00:00:00' | 2018-10-26T11:12:43 | 2018-10-26T11:12:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 433,408,959 | 3,472 | BooBSD | Проверил. Я неправильно понимал. `FINAL` считает в памяти и ничего не меняет на диске. То есть надо делать `OPTIMIZE`, что ооочень долго. Каждый раз использовать `SELECT FINAL` тоже не вариант, так как в 8 раз медленнее, чем просто `SELECT`.
А вот `ALTER TABLE UPDATE` работает на удивление быстро. | 2018-10-26T13:33:16 | 2018-10-26T13:33:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 433,471,266 | 3,448 | ztlpn | Функция `pointInEllipses` работает со своими аргументами, как с декартовыми координатами. Соответственно, расстояние считается на плоскости, а не на поверхности Земли. | 2018-10-26T16:44:41 | 2018-10-26T16:44:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 433,529,908 | 3,472 | BooBSD | @ztlpn
>ALTER UPDATE перезаписывает только изменяющиеся столбцы в изменяющихся кусках
Не совсем понятно. Если я делаю точечный апдейт по конкретному первичному ключу, то перезаписывается весь кусок (из `system.parts`) или только небольшая часть куска, которая относится к конкретному первичному ключу? Иными словами... | 2018-10-26T20:13:37 | 2018-10-26T20:13:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 433,714,419 | 3,421 | bgranvea | I found a way to reproduce the issue.
I'm on a single server but with replication enabled:
```
<vscluster>
<shard>
<internal_replication>true</internal_replication>
<replica>
<host>myhostname</host>
<port>9000</port>
</replica>
</shard>
</vscluster>
```
Steps to reprod... | 2018-10-28T15:15:22 | 2018-10-28T15:15:22 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 433,868,943 | 2,601 | fillest | Another example.
`CREATE TABLE IF NOT EXISTS testdb.test (time DateTime, text String) ENGINE = MergeTree() PARTITION BY toDate(time) ORDER BY (time)`
I've inserted 17280000000 rows for total 99G uncompressed (resulted in ~2G compressed).
`SELECT time, text FROM testdb.test WHERE time > 1540581226 ORDER BY time ASC ... | 2018-10-29T10:58:36 | 2018-10-29T11:00:54 | {} | NONE |
ClickHouse | ClickHouse | 433,872,369 | 3,483 | alesapin | Can't reproduce:
config.xml:
```
...
<!-- Substitutions for parameters of replicated tables.
Optional. If you don't use replicated tables, you could omit that.
See https://clickhouse.yandex/docs/en/table_engines/replication/#creating-replicated-tables
-->
<macros>
... | 2018-10-29T11:11:12 | 2018-10-29T11:11:29 | {} | MEMBER |
ClickHouse | ClickHouse | 433,892,494 | 3,478 | alesapin | Maybe you can provide server logs and/or some data to reproduce the problem? On empty tables it doesn't reproduce. | 2018-10-29T12:31:41 | 2018-10-29T12:31:52 | {} | MEMBER |
ClickHouse | ClickHouse | 433,924,381 | 3,455 | alesapin | Why not `¯\_(ツ)_/¯`? | 2018-10-29T14:11:34 | 2018-10-29T14:11:34 | {} | MEMBER |
ClickHouse | ClickHouse | 434,025,190 | 3,380 | ztlpn | Fixed starting from 18.14.9 | 2018-10-29T18:24:09 | 2018-10-29T18:24:09 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 434,149,788 | 3,380 | BlankRain | yes ,it works . | 2018-10-30T02:33:00 | 2018-10-30T02:33:00 | {} | NONE |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.