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
426,609,881
3,279
proller
If you set interserver_http_host = hostname/ip from eth2 - other servers will connect to this ip . And you should have one listen_host = interserver_http_host or :: Now you can't define explicit pairs of listen_host and *some*_port . But you can use many listen_host .
2018-10-03T11:57:53
2018-10-03T11:57:53
{}
CONTRIBUTOR
ClickHouse
ClickHouse
427,057,813
3,210
alesapin
Ok, thank you. Now travis build fails (you can see error in "Details" link) and two ODBC integration tests fail. https://pastebin.com/FuMwJTy4
2018-10-04T15:14:48
2018-10-04T15:14:48
{}
MEMBER
ClickHouse
ClickHouse
427,221,783
3,230
alexey-milovidov
No, it is unrelated.
2018-10-05T01:47:56
2018-10-05T01:47:56
{}
MEMBER
ClickHouse
ClickHouse
427,738,608
3,310
healiseu
@filimonov I think I got it, it's a hash join cashed in memory. That's a great idea they made this feature separable too in clickhouse. Thank you also for the link on how to search for examples on its usage. May I also ask you about the other kind of engine the Dictionary and how this is compared with the Join Engi...
2018-10-08T07:09:28
2018-10-08T12:57:16
{}
NONE
ClickHouse
ClickHouse
427,954,588
3,299
filimonov
I think that it caused by the fact that empty string is not correct UInt16 value for`ndx` attribute. Try zero instead.
2018-10-08T19:37:47
2018-10-08T19:37:47
{}
CONTRIBUTOR
ClickHouse
ClickHouse
427,960,651
167
simPod
I see, you're right. Thanks! On Mon, Oct 8, 2018, 21:25 Alex Zatelepin <notifications@github.com> wrote: > Closed #167 <https://github.com/yandex/ClickHouse/issues/167>. > > — > You are receiving this because you are subscribed to this thread. > Reply to this email directly, view it on GitHub > <https://github.com/ya...
2018-10-08T19:59:45
2018-10-08T19:59:45
{}
CONTRIBUTOR
ClickHouse
ClickHouse
426,613,650
3,279
proller
Correct. Also you can use <listen_host>::</listen_host> for listening all possible interfaces.
2018-10-03T12:10:31
2018-10-03T12:10:31
{}
CONTRIBUTOR
ClickHouse
ClickHouse
426,614,152
3,279
MasterGroosha
Thank you @proller. Now this is clear to me and my issue is solved.
2018-10-03T12:12:10
2018-10-03T12:12:10
{}
NONE
ClickHouse
ClickHouse
426,746,805
3,276
andyvauliln
I had same problem and i did actions above, but still can't run Building CXX object dbms/src/IO/tests/CMakeFiles/write_buffer.dir/write_buffer.cpp.o FAILED: dbms/src/IO/tests/CMakeFiles/write_buffer.dir/write_buffer.cpp.o /usr/local/bin/ccache /Library/Developer/CommandLineTools/usr/bin/c++ -DBOOST_SYSTEM_NO_DEPR...
2018-10-03T18:30:54
2018-10-03T18:30:54
{}
NONE
ClickHouse
ClickHouse
427,120,326
3,271
rajadayalan
Please find the create table statements at https://gist.github.com/rajadayalan/2c669797e33cd19153ef20a4568c2939. We are not retaining source table data for more than couple of hours. We will increase the source table data retention for couple of days. Can provide you with count comparison by tomorrow evening (afte...
2018-10-04T18:21:06
2018-10-04T18:21:06
{}
NONE
ClickHouse
ClickHouse
427,140,067
3,295
alexey-milovidov
It's important to note, that `SHOW CREATE TABLE` query is output a resultset with single row and single column of type String. It's still a resultset, and you can get it in any available format. Examples: ``` clickhouse-client -q "SHOW CREATE TABLE test FORMAT TSV" clickhouse-client -q "SHOW CREATE TABLE test...
2018-10-04T19:24:54
2018-10-04T19:24:54
{}
MEMBER
ClickHouse
ClickHouse
427,147,254
3,296
alesapin
Yes, accident error.
2018-10-04T19:50:26
2018-10-04T19:50:26
{}
MEMBER
ClickHouse
ClickHouse
427,788,304
3,310
filimonov
> May I also ask you about the other kind of engine you have the Dictionary and how this is compared with the Join Engine. The biggest difference is that dictionaries have completely another syntax, and rather another usage scenarios. They are handier for scenarios like "get value by id", "create some derived colum...
2018-10-08T10:38:22
2018-10-08T10:39:28
{ "+1": 1 }
CONTRIBUTOR
ClickHouse
ClickHouse
427,809,405
3,308
healiseu
I wish I could, out of frustration I `rm -rf` the whole folder and I also deleted the `xml` dictionary configuration I have made inside the `/etc/clickhouse-server` folder and when I tried to create a new configuration for the dictionary I run into the problem I described in issue #3299. Once I finish playing with norm...
2018-10-08T12:10:48
2018-10-08T12:10:48
{}
NONE
ClickHouse
ClickHouse
427,915,119
3,307
vavrusa
That sounds correct.
2018-10-08T17:24:01
2018-10-08T17:24:01
{}
CONTRIBUTOR
ClickHouse
ClickHouse
427,962,838
2,383
alexey-milovidov
#3101
2018-10-08T20:08:07
2018-10-08T20:08:07
{}
MEMBER
ClickHouse
ClickHouse
426,940,046
3,210
alesapin
Build still fails with errors: https://pastebin.com/EeVHeViY
2018-10-04T08:58:56
2018-10-04T08:58:56
{}
MEMBER
ClickHouse
ClickHouse
427,041,646
3,289
champtar
There is a fix for #3144 that is not merged yet: #3223
2018-10-04T14:31:54
2018-10-04T14:31:54
{}
CONTRIBUTOR
ClickHouse
ClickHouse
427,044,088
3,290
4ertus2
Есть несколько похожих проблем с именованием колонок в выдаче. В ближайшее время перепишем там код.
2018-10-04T14:38:31
2018-10-04T14:38:31
{}
CONTRIBUTOR
ClickHouse
ClickHouse
427,063,178
3,210
alex-krash
Got it, thanks! Will spend more time on trying to launch integrations.
2018-10-04T15:29:31
2018-10-04T15:29:31
{}
CONTRIBUTOR
ClickHouse
ClickHouse
427,167,594
3,283
alexey-milovidov
Great!
2018-10-04T21:00:43
2018-10-04T21:00:43
{}
MEMBER
ClickHouse
ClickHouse
427,566,766
1,981
healiseu
## More about uninstalling and installing again clickhouse I noticed that `apt purge clickhouse-server` `apt purge clickhouse-client` are not enough you need also `apt autoremove` But even with these in place there is leftover of clickhouse from previous installation So I also did several `rm -rf .../clickho...
2018-10-06T11:33:23
2018-10-06T11:34:37
{}
NONE
ClickHouse
ClickHouse
427,662,328
3,310
iFoton
Try to change `ON BicycleTripStar.start_station_id = StartStation.id` to `ON BicycleTripStar.start_station_id = StartStationRAM.id`, and in group by clause same.
2018-10-07T15:34:17
2018-10-07T15:34:52
{}
NONE
ClickHouse
ClickHouse
427,919,800
3,299
healiseu
Following @filimonov examples and after creating again the dictionary above and the main facts table `describe table BicycleTripStar` | name | type | | --------| ------ | | ..//.. | ..//.. | | member_id | UInt8 | I managed to bypass the UInt8 conflict data type by applying type conversion function `toUInt...
2018-10-08T17:39:23
2018-10-08T17:52:57
{}
NONE
ClickHouse
ClickHouse
427,939,114
3,299
healiseu
OK I decided to add another source to my `bicycle_trips` dictionaries, so I modified the XML configuration as follows: ```xml <bicycle_trips> <dictionary> <name>MemberTypeLookup</name> <source> <clickhouse> <host>localhost</host> <port>9000</port> <u...
2018-10-08T18:42:06
2018-10-08T18:42:06
{}
NONE
ClickHouse
ClickHouse
427,961,588
3,310
filimonov
Pointers to external fact tables sounds pretty similar to well known [star schema](https://en.m.wikipedia.org/wiki/Star_schema). Am I right? Storing values which are not byte/word aligned can give better compression but can also be less CPU effective for processing. About qlick and hypergraphs - i don't feel i ha...
2018-10-08T20:03:13
2018-10-08T20:05:25
{}
CONTRIBUTOR
ClickHouse
ClickHouse
427,146,926
3,296
alexey-milovidov
Why we have `packager` and `packager.py` - mostly identical files? Maybe you have added `packager.py` accidentially?
2018-10-04T19:49:18
2018-10-04T19:49:18
{}
MEMBER
ClickHouse
ClickHouse
427,773,174
3,311
filimonov
Not really sure if it will work, but you can try to replace credentials with macros defined in config.
2018-10-08T09:34:33
2018-10-08T09:34:33
{}
CONTRIBUTOR
ClickHouse
ClickHouse
427,791,028
3,308
filimonov
For non-replicated tables removing corresponding files / folders in those directories should help to completely forget 'bad' table: ``` /var/lib/clickhouse/metadata /var/lib/clickhouse/data ```
2018-10-08T10:50:42
2018-10-08T10:50:42
{ "+1": 1 }
CONTRIBUTOR
ClickHouse
ClickHouse
426,620,977
3,277
4ertus2
Посмотрел по коду. Для строк малой кровью расширить не получится - алгоритм ожидает колонку на базе цифрового типа. Выглядит легким расширить на Data, DateTime, UUID, Decimal.
2018-10-03T12:34:53
2018-10-03T12:34:53
{}
CONTRIBUTOR
ClickHouse
ClickHouse
426,893,337
1,389
maxulan
+1
2018-10-04T05:50:43
2018-10-04T05:50:43
{}
CONTRIBUTOR
ClickHouse
ClickHouse
426,960,918
3,280
filimonov
Confirm. Selects on same table: ``` select count() from tst final; -- ok (2) select count() from tst final where timestamp>0; -- wrong result (0) select count() from tst final prewhere timestamp > 0; -- exeption: Not found column timestamp in block. There are only columns: 0. select count() from tst ...
2018-10-04T10:06:51
2018-10-04T10:06:51
{}
CONTRIBUTOR
ClickHouse
ClickHouse
427,038,394
3,289
filimonov
Maybe the simple extracting fragment of raw JSON would be a solution? Like in visitParamExtractRaw? input_format_import_nested_json_as_raw_strings or whatever?
2018-10-04T14:23:30
2018-10-04T14:24:54
{}
CONTRIBUTOR
ClickHouse
ClickHouse
427,134,236
3,295
4ertus2
It's cause output of your query is in escaped format. You can use FORMAT TabSeparatedRaw to avoid escaping. Then you'll get expected result. ``` CREATE TABLE IF NOT EXISTS test.t (s String) ENGINE = Memory; CREATE MATERIALIZED VIEW test.mv ENGINE = Memory AS SELECT toDateTime(splitByChar('\\', s)[1]) AS d FROM tes...
2018-10-04T19:05:10
2018-10-04T19:05:38
{}
CONTRIBUTOR
ClickHouse
ClickHouse
427,172,907
3,237
alexey-milovidov
Up.
2018-10-04T21:18:35
2018-10-04T21:18:35
{}
MEMBER
ClickHouse
ClickHouse
427,700,816
3,014
lightee2
@filimonov the debug mode is ok . so odd
2018-10-08T00:45:34
2018-10-08T00:45:34
{}
NONE
ClickHouse
ClickHouse
427,743,881
3,312
blinkov
At the moment there are no UNIQUE constraints in ClickHouse, even for what looks like primary key. So there's no way there would be constraint violations that could be handled differently.
2018-10-08T07:35:31
2018-10-08T07:41:31
{}
CONTRIBUTOR
ClickHouse
ClickHouse
427,848,847
3,310
filimonov
AFAIK there are some plans/ ideas of introducing 'some kind' of secondary indexes. Right now there is an option to have an extra copy of same data (or some subset) with another PK (it can take a lot of disk space, but if you really need it - it's doable) via the materialized view. > The most effective way to demo...
2018-10-08T14:08:50
2018-10-08T14:13:22
{}
CONTRIBUTOR
ClickHouse
ClickHouse
427,929,254
3,322
vavrusa
The latest master switched to ephemeral locks for block numbers, but I'm not sure whether they're cleared when abandoned, @ztlpn would know.
2018-10-08T18:10:39
2018-10-08T18:10:39
{}
CONTRIBUTOR
ClickHouse
ClickHouse
427,950,429
167
ztlpn
> I have a question. Shouldn't the UPDATE and DELETE propagate to materialized views as well? Well in theory that would be nice, but I don't see how this could be realized in practice. Materialized views often store aggregated data and deleting/modifying data inside aggregate function states is in general impossible...
2018-10-08T19:22:28
2018-10-08T19:22:28
{}
CONTRIBUTOR
ClickHouse
ClickHouse
428,086,948
3,310
filimonov
Not sure if I understood you correctly, but it sounds like for you the most important are key-value-alike features of the database (and now I understand why you were asking so much about JOIN engine and dictionaries). If so - I'm not sure if ClickHouse will suit that usage scenario. Engin=JOIN, Dictionaries, Engine=Set...
2018-10-09T07:19:31
2018-10-09T07:23:44
{}
CONTRIBUTOR
ClickHouse
ClickHouse
426,612,327
3,279
MasterGroosha
Do you mean that to make clickhouse-client / third-party clients work through `eth1 : 192.168.2.92` AND replication work through `eth2: 192.168.40.92` I simply have to add `<listen_host>192.168.40.92</listen_host>` so that config will be the following one? ```xml <interserver_http_port>9009</interserver_http_port> ...
2018-10-03T12:06:15
2018-10-03T12:18:47
{}
NONE
ClickHouse
ClickHouse
427,077,854
3,293
4ertus2
Can't reproduce. In current version in all cases I've got an error: ``` DateTime, parsed text: "1"ERROR: DateTime must be in YYYY-MM-DD hh:mm:ss or NNNNNNNNNN (unix timestamp, exactly 10 digits) format. ``` Code to reproduce ``` CREATE TABLE IF NOT EXISTS test.tsv (d DateTime, x UInt32) ENGINE = Memory; INS...
2018-10-04T16:10:35
2018-10-04T16:10:35
{}
CONTRIBUTOR
ClickHouse
ClickHouse
427,785,676
3,308
healiseu
**UPDATE:** The error might be related to my experimental setup with dictionaries (see #3299) although clickhouse server crashed immediately after setting the feature above and trying to restart the server. Unfortunately I could not find how to revert the process and make the clickhouse server start again. In that c...
2018-10-08T10:26:29
2018-10-08T10:26:29
{}
NONE
ClickHouse
ClickHouse
428,055,142
3,323
zhang2014
LGTM
2018-10-09T04:24:32
2018-10-09T04:24:32
{}
CONTRIBUTOR
ClickHouse
ClickHouse
426,606,424
3,279
MasterGroosha
@proller Okay, now quoting config file again: `This setting could be used to switch replication to another network interface`. So how exactly one can switch replication to another interface? What should be set as `interserver_http_host` value to make replication work through interface `eth2` (ip 192.168.40.92), not ...
2018-10-03T11:44:13
2018-10-03T11:45:14
{}
NONE
ClickHouse
ClickHouse
426,680,731
3,282
4ertus2
Hello, Could you print here sample queries to reproduce the issue?
2018-10-03T15:29:03
2018-10-03T15:29:31
{}
CONTRIBUTOR
ClickHouse
ClickHouse
427,038,990
3,269
4ertus2
There's a workaround for the problem: ``` set optimize_min_equality_disjunction_chain_length = 100; ``` There's some bug in disjunction optimisations, that starts from 3 OR by default. It's possible to disable it by setting option above to huge value.
2018-10-04T14:25:11
2018-10-04T14:25:11
{}
CONTRIBUTOR
ClickHouse
ClickHouse
427,801,828
3,308
healiseu
@filimonov I see, well actually I ended up doing this anyway because when I reinstalled clickhouse and attempted to attach my old tables I got this server startup error. So I had to remove the data/metadata folder that was causing the problem. But I am sure this whole experience of mine will help others to save some ti...
2018-10-08T11:37:54
2018-10-08T11:42:15
{}
NONE
ClickHouse
ClickHouse
427,951,087
167
ztlpn
ALTER UPDATE feature has been released in v18.12.14.
2018-10-08T19:24:59
2018-10-08T19:24:59
{ "+1": 2, "hooray": 2 }
CONTRIBUTOR
ClickHouse
ClickHouse
427,167,155
3,285
alexey-milovidov
Thank you!
2018-10-04T20:59:11
2018-10-04T20:59:11
{}
MEMBER
ClickHouse
ClickHouse
428,091,736
3,326
filimonov
> clickhouse ask password again Sounds like a bug connected to #2293. Just to ensure - you call clickhouse-client with `--ask-password`, do you? > And one more thing is that why upper-case "VERTICAL" is not allowed ? [By design](https://clickhouse.yandex/docs/en/query_language/syntax/#keywords): _Keywords (su...
2018-10-09T07:37:40
2018-10-09T07:37:40
{}
CONTRIBUTOR
ClickHouse
ClickHouse
426,631,621
3,277
filimonov
Добавление остальных нестроковых типов не особо нужно - в конце концов их без особых проблем можно кастовать туда-обратно и на уровне запроса (перечислил их просто для полноты картины). Добавление строк - да, может быть ощутимо более болезненным, но и польза ощутимая. Когда-нибудь нужно довести sumMap до ума (та...
2018-10-03T13:10:40
2018-10-03T13:11:58
{}
CONTRIBUTOR
ClickHouse
ClickHouse
426,890,829
3,271
filimonov
That should't happen. The most wierd that you have proper data for current day. Can you identify which MV have correct numbers and which - wrong by comparing the counts with directly counted from source data? Can you provide some datails like CREATE statements for source table and views?
2018-10-04T05:34:28
2018-10-04T05:34:28
{}
CONTRIBUTOR
ClickHouse
ClickHouse
426,967,753
3,286
4ertus2
Yup, it's known bug. It should be fixed soon. https://github.com/yandex/ClickHouse/blob/master/dbms/tests/queries/0_stateless/00722_inner_join.sql#L36
2018-10-04T10:30:18
2018-10-04T10:30:18
{}
CONTRIBUTOR
ClickHouse
ClickHouse
426,995,844
3,276
proller
Try build with gcc-8 from brew.
2018-10-04T12:14:31
2018-10-04T12:14:31
{}
CONTRIBUTOR
ClickHouse
ClickHouse
427,663,885
3,310
filimonov
The idea behind that is quite simple. For using some data as right part of join (or as the set of values for IN operator) data should be read from disk (with preprocessing if needed) and put in RAM in special memory structures (like hash tables). If you need to make a lot of requests with the same right part of the...
2018-10-07T15:55:36
2018-10-08T09:36:38
{}
CONTRIBUTOR
ClickHouse
ClickHouse
427,811,248
3,311
filimonov
hm... Maybe some special macros available only for create/attach table statements can be implemented.
2018-10-08T12:18:37
2018-10-08T12:18:37
{}
CONTRIBUTOR
ClickHouse
ClickHouse
427,833,383
3,311
filimonov
Also: you can connect to MySQL using ODBC. And reference connection only by DSN, while storing credentials in odbc.ini. See here: https://www.altinity.com/blog/2018/9/20/using-odbc-with-clickhouse
2018-10-08T13:30:13
2018-10-08T13:30:13
{}
CONTRIBUTOR
ClickHouse
ClickHouse
426,676,994
3,280
4ertus2
It's a bug. It's look like 'is not null' is not working on not nullable column with MergeTree engine at all. ``` drop table if exists tst; create table tst ( timestamp DateTime, val Nullable(Int8) ) engine MergeTree partition by toYYYYMM(timestamp) ORDER by (timestamp); insert into tst values ('2018-02-...
2018-10-03T15:19:24
2018-10-03T15:19:24
{}
CONTRIBUTOR
ClickHouse
ClickHouse
426,684,574
3,282
igor-sh8
Create table query: `create table TestDec (Date Date, Id Int32, Sum Decimal64(8)) Engine=SummingMergeTree(Date,(Id),8192,(Sum))` Optimize query: `optimize table TestDec` Stack Trace: ``` 2018.10.03 15:35:50.275602 [ 14 ] <Error> void DB::BackgroundProcessingPool::threadFunction(): Code: 48, e.displayText(...
2018-10-03T15:38:54
2018-10-03T15:38:54
{}
NONE
ClickHouse
ClickHouse
427,013,062
3,288
4ertus2
Выжимка: Method index is not supported for ColumnUnique. ``` tadig LowCardinality(String), country LowCardinality(String), ````
2018-10-04T13:12:27
2018-10-04T13:12:27
{}
CONTRIBUTOR
ClickHouse
ClickHouse
427,015,638
3,210
alex-krash
@alesapin , sorry, my bad. I don't have reproduce for this error while build.
2018-10-04T13:20:13
2018-10-04T13:20:13
{}
CONTRIBUTOR
ClickHouse
ClickHouse
427,424,125
3,304
igor-sh8
Oh, sorry. My fault. I made a mistake in the create table statement. Date column already included in the sort-key. `CREATE TABLE Table1 ( Date Date, Col1 Int32, Col2 In32, ... ,Col10 Int32) ENGINE = ReplicatedSummingMergeTree('/clickhouse/tables/{shard}/Table1', '{replica}', Date, (Date, Col1, Col2, ... ,Col5), 819...
2018-10-05T16:29:38
2018-10-05T16:29:38
{}
NONE
ClickHouse
ClickHouse
427,786,167
3,313
blinkov
@4ertus2 а можешь сразу и в английскую версию это дописать?
2018-10-08T10:28:50
2018-10-08T10:28:50
{}
CONTRIBUTOR
ClickHouse
ClickHouse
426,768,430
3,283
alesapin
Changes don't affect failed test.
2018-10-03T19:30:45
2018-10-03T19:30:45
{}
MEMBER
ClickHouse
ClickHouse
427,125,896
3,297
alexey-milovidov
No need to worry, it is in our task queue. Give us a few days pls.
2018-10-04T18:38:02
2018-10-04T18:38:02
{}
MEMBER
ClickHouse
ClickHouse
427,806,076
3,308
filimonov
> So I had to remove the data/metadata folder that was causing the problem. It's enough to remove one file (name = your_table_name.sql). And it would be nice if you will attach it's content here to allow developers to reproduce that problem :)
2018-10-08T11:56:33
2018-10-08T11:57:21
{}
CONTRIBUTOR
ClickHouse
ClickHouse
427,933,018
3,323
blinkov
@sundy-li @zhang2014
2018-10-08T18:22:55
2018-10-08T18:22:55
{}
CONTRIBUTOR
ClickHouse
ClickHouse
428,098,705
3,198
omigo
We meet this problem too. So sad!
2018-10-09T08:02:57
2018-10-09T08:02:57
{}
NONE
ClickHouse
ClickHouse
427,035,705
1,389
alex-krash
Dear maintainers, I may try do it by my own. There is a vendored Poco's implementation, but it operates std::*stream, so will not take it into consideration for now. Let's try https://github.com/aklomp/base64 as submodule? I assume that encode/decode functions will receive a ColumnString, and do the magic. The tri...
2018-10-04T14:15:58
2018-10-04T14:15:58
{ "+1": 1 }
CONTRIBUTOR
ClickHouse
ClickHouse
427,198,106
3,230
vavrusa
Is the workaround to compile with tcmalloc?
2018-10-04T23:11:21
2018-10-04T23:11:21
{}
CONTRIBUTOR
ClickHouse
ClickHouse
427,745,074
3,311
tungdam
+1 I have the same issue. And i believe we should hide this password somehow at least.
2018-10-08T07:41:00
2018-10-08T07:41:00
{ "+1": 2 }
NONE
ClickHouse
ClickHouse
427,806,919
3,299
healiseu
@filimonov about the error, you are right `last_exception` column displays it crystal clear `dictionary.structure.attribute.null_value, e.what() = Not found` I added `<null_value></null_value>` in the structure section of the configuration file and the problem is solved. Thank you also for the motivational links, I ...
2018-10-08T12:00:17
2018-10-08T12:00:17
{}
NONE
ClickHouse
ClickHouse
427,772,710
3,299
healiseu
**UPDATE:** I have read in more detail the documentation of clickhouse and it seems the mandatory UInt64 data type of the Dictionary I tried to set up above is due to the limitations imposed from the data structure layout of the dictionary in memory, i.e. [flat type unpacked array][1]. Sorry, I should have written the ...
2018-10-08T09:32:54
2018-10-08T17:55:19
{}
NONE
ClickHouse
ClickHouse
427,799,362
3,312
filimonov
> upsert You can try [ReplacingMergeTree engine](https://clickhouse.yandex/docs/en/operations/table_engines/replacingmergetree/) and use timestamp instead of versions. It has a lot of drawbacks but it removes duplicates during background merges (or if you add FINAL clause to your SELECTS).
2018-10-08T11:26:50
2018-10-08T11:26:50
{}
CONTRIBUTOR
ClickHouse
ClickHouse
428,050,437
3,323
sundy-li
This pr is ok, @lamber-ken we could translate more than one file in a single pr, it will be better.
2018-10-09T03:50:02
2018-10-09T03:50:02
{}
CONTRIBUTOR
ClickHouse
ClickHouse
426,665,081
3,277
4ertus2
> Добавление остальных нестроковых типов не особо нужно Ну, уже сделал. #3281
2018-10-03T14:46:54
2018-10-03T14:46:54
{}
CONTRIBUTOR
ClickHouse
ClickHouse
426,716,428
3,280
sarg
Actually, `IS NOT NULL` isn't strictly necessary. Any condition will do fine, try `where timestamp > '2017-01-01 00:00'`
2018-10-03T17:03:12
2018-10-03T17:03:12
{}
NONE
ClickHouse
ClickHouse
427,160,556
3,290
snail-fuji
Спасибо за оперативный ответ! А не подскажете, где можно это место найти в коде?)
2018-10-04T20:36:58
2018-10-04T20:36:58
{}
NONE
ClickHouse
ClickHouse
427,168,040
3,283
alexey-milovidov
I think, we can write a test... It can read metric from query_log. Another question: did you managed to show some performance difference in extreme cases?
2018-10-04T21:02:18
2018-10-04T21:02:18
{}
MEMBER
ClickHouse
ClickHouse
427,380,831
3,302
alexey-milovidov
Currently there is no way to do that. We have no classification of errors to server/user. If you really need it, you can manually collect a set of specific error codes and grep them from the error log.
2018-10-05T14:15:44
2018-10-05T14:16:32
{}
MEMBER
ClickHouse
ClickHouse
427,419,089
3,304
den-crane
Hi, one of the possible reason that you did not include date in your sort-key. ``` CREATE TABLE Table1 ( date Date, col1 UInt64, col2 Float64) ENGINE = SummingMergeTree(date, col1, 8192) insert into Table1 values('2018-10-01', 1, 6); insert into Table1 values('2018-10-02', 1, 6); insert into Table1 values('2018-...
2018-10-05T16:12:22
2018-10-05T16:12:22
{}
CONTRIBUTOR
ClickHouse
ClickHouse
427,420,228
3,304
den-crane
There is no any automatic resharding in CH. Old syntax (partition by month): SummingMergeTree(date, (col1,date), 8192) New syntax (partition by month): SummingMergeTree partition by toYYYYMM(date), order by (col1,date) New syntax (partition by day): SummingMergeTree partition by (date), order by (col1,date) and...
2018-10-05T16:16:06
2018-10-05T16:16:06
{}
CONTRIBUTOR
ClickHouse
ClickHouse
428,022,278
2,416
gabbifish
Are there any plans to support WITH clause for non-scalar values?
2018-10-09T00:39:17
2018-10-09T00:39:17
{ "+1": 17, "eyes": 1 }
NONE
ClickHouse
ClickHouse
428,089,783
3,325
blinkov
We might update it in future, but it has very low priority. https://clickhouse.yandex/benchmark.html used to be the only benchmark when ClickHouse has been opensourced, but nowadays there are a lot of independent benchmarks (some examples are linked on [official website front page](https://clickhouse.yandex/#indepen...
2018-10-09T07:30:09
2018-10-09T07:30:09
{}
CONTRIBUTOR
ClickHouse
ClickHouse
426,612,887
3,182
VadimPlh
I remember about tests.
2018-10-03T12:08:09
2018-10-03T12:08:09
{}
CONTRIBUTOR
ClickHouse
ClickHouse
427,797,923
3,311
PhantomPhreak
Fair enough, let's try: ``` my.awesome.clickhouse :) create table mysqltest (id UInt64, field String) ENGINE = MySQL('127.0.0.1:3306', 'chtest', 'test', 'testuser', '{supersecurepassword}'); CREATE TABLE mysqltest ( id UInt64, field String ) ENGINE = MySQL('12...
2018-10-08T11:20:24
2018-10-08T11:20:24
{}
NONE
ClickHouse
ClickHouse
427,875,992
3,318
zhang2014
:+1:
2018-10-08T15:21:24
2018-10-08T15:21:24
{}
CONTRIBUTOR
ClickHouse
ClickHouse
428,108,246
3,198
filimonov
1) can you try to consume kafka messages with [kafkacat](https://github.com/edenhill/kafkacat) from the same host where CH is installed? 2) Do you use clickhouse from packages, or build from sources? 3) can you also check clickhouse-server.log (not an err.log) for something suspicious? 4) how many nodes do you hav...
2018-10-09T08:35:39
2018-10-09T08:37:16
{}
CONTRIBUTOR
ClickHouse
ClickHouse
426,648,118
3,276
SlavaNovikov
Thanks, I ran.
2018-10-03T13:59:38
2018-10-03T13:59:38
{}
NONE
ClickHouse
ClickHouse
426,728,468
3,271
rajadayalan
dont see insert/materialization errors in clickhouse-server.err.log.
2018-10-03T17:37:55
2018-10-03T17:37:55
{}
NONE
ClickHouse
ClickHouse
427,032,598
3,289
4ertus2
JSONEachRow is a simple json-like format that optimized for perfomance. It doesn't support complex nested structs. It's possible to insert Nested type via array syntax: ``` CREATE TABLE test.json_each_row_nested( n Nested (s String, i Int32) ) ENGINE = Memory ``` { "n.i":[45, 67, 8], "n.s":["x", "y", "z"] } ...
2018-10-04T14:07:36
2018-10-04T14:07:36
{}
CONTRIBUTOR
ClickHouse
ClickHouse
427,037,039
3,289
filimonov
@4ertus2, PR #3144 introduced support for nested structs when subkeys are arrays, but i have no idea how to deal (in common case) with json structure @minaevmike show. May be @veloman-yunkan have some ideas regarding that?
2018-10-04T14:19:43
2018-10-04T14:21:51
{}
CONTRIBUTOR
ClickHouse
ClickHouse
427,065,195
3,210
alesapin
Maybe it's possible to recognize error from stack trace: ``` E QueryRuntimeException: Client failed! Return code: 232, stderr: Received exception from server (version 18.14.2): E Code: 1000. DB::Exception: Received from 172.18.0.3:9000. DB::Exception: Not found: dictionary.source.connection_strin...
2018-10-04T15:35:05
2018-10-04T15:35:05
{}
MEMBER
ClickHouse
ClickHouse
427,130,866
1,389
alexey-milovidov
> Let's try https://github.com/aklomp/base64 as submodule? Yes, this library is recommended. > I assume that encode/decode functions will receive a ColumnString, and do the magic. The trick here, is that the buffer for base64 output length should be 4/3 of input, and string length in column differs. I propose to...
2018-10-04T18:54:01
2018-10-04T18:55:08
{}
MEMBER
ClickHouse
ClickHouse
427,798,071
3,299
filimonov
> Is it always the case that id fields of a dictionary are of type UInt64 ? It depends on the chosen [memory layout](https://clickhouse.yandex/docs/en/query_language/dicts/external_dicts_dict_layout/). > Why is it compulsory to have a lookup field in a facts-main table that is set to UInt64 ? Just implementati...
2018-10-08T11:21:07
2018-10-08T11:21:07
{}
CONTRIBUTOR
ClickHouse
ClickHouse
427,809,858
3,307
filimonov
Theoretically: 1) Kafka engine should commit new offset only after the successful 'insert'. @vavrusa am I right? 2) if some MV insert is failed the main insert also fails (but inserts to other MV which finished successfully in a meanwhile are not rolled back). 3) MV insert should fail if insert_quorum is enabled a...
2018-10-08T12:12:42
2018-10-08T12:13:31
{}
CONTRIBUTOR
ClickHouse
ClickHouse
427,959,718
3,299
healiseu
Hi @filimonov, this is the [CSV file][1] I load on `StartStation` dictionary source table ```bash cat 2011-tripdata-start_station.csv | clickhouse-client --database=DemoDB --query="INSERT INTO StartStationRAM FORMAT CSV" ``` I thought the problem was with the order of columns, but it seems that is not the case. Not...
2018-10-08T19:56:27
2018-10-08T19:57:53
{}
NONE
ClickHouse
ClickHouse
428,124,535
3,299
healiseu
@filimonov I traced the error, it is caused by the attribute definition of `ndx` field: ```xml <attribute> <name>ndx</name> <type>UInt16</type> <null_value></null_value> </attribute> ``` If I change the <type>UInt16</type> to <type>String</type> it works. But I don't underst...
2018-10-09T09:28:49
2018-10-09T09:28:49
{}
NONE