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 | 529,088,565 | 6,858 | bodgergely | @alexey-milovidov You are totally right, my mistake! :) | 2019-09-07T08:56:12 | 2019-09-07T08:56:12 | {} | NONE |
ClickHouse | ClickHouse | 529,148,574 | 6,860 | simPod | Oh I see. Thanks for quick response! | 2019-09-07T21:33:25 | 2019-09-07T21:33:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,248,729 | 6,828 | alexey-milovidov | I will ask if it is possible to provide these repositories via symlinks or similar mechanics without actually duplicating files.
> Also it would be nice to have builds done in the appropriate debian release instead of a common one supposed to work on all releases.
We intentionally make portable packages that can ... | 2019-09-08T22:45:56 | 2019-09-08T22:45:56 | {} | MEMBER |
ClickHouse | ClickHouse | 529,677,034 | 6,870 | alexey-milovidov | It's difficult to figure out what aggregate function do you want.
Hardcoding `sum` aggregate function is not a good way to do. | 2019-09-09T21:36:51 | 2019-09-09T21:36:51 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 529,711,330 | 6,879 | volfco | ```
SELECT *
FROM system.merges
┌─database─┬─table───┬─────elapsed─┬───────────progress─┬─num_parts─┬─source_part_names───────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┬─result_part_name───────┬─partition_id─┬─is_mutation─┬─tot... | 2019-09-09T23:56:21 | 2019-09-09T23:56:21 | {} | NONE |
ClickHouse | ClickHouse | 528,896,565 | 6,841 | filimonov | 19.11 works. | 2019-09-06T15:17:51 | 2019-09-06T15:17:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,441,298 | 6,818 | joebling | > Your usage of `view of distributed table` looks bad. The idiomatic way is to setup a view for the local table on every node， then setup a distributed view for those local views. See [this](https://github.com/yandex/ClickHouse/issues/467) for reference.
thank you, it works | 2019-09-09T12:07:00 | 2019-09-09T12:07:00 | {} | NONE |
ClickHouse | ClickHouse | 529,659,130 | 6,855 | stavrolia | for #5371 issue. | 2019-09-09T20:44:51 | 2019-09-09T20:44:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,675,779 | 6,367 | filimonov | What about queries across few databases, like
```
select ... from db1.table1 left join db2.table2 left join db3.table3 where value in (select * from db4.table4)
``` | 2019-09-09T21:32:45 | 2019-09-09T21:32:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,681,249 | 4,105 | filimonov | Fixed in https://github.com/yandex/ClickHouse/commit/bf99e785c8cf8d120f2952d76b335466739f08db#diff-d1bbe1458d881f9059e260afeffdaae3
and https://github.com/yandex/ClickHouse/commit/777af5544b8259758875f3ff39738d957f9cacc8#diff-d1bbe1458d881f9059e260afeffdaae3 | 2019-09-09T21:50:31 | 2019-09-09T21:50:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,685,327 | 4,050 | filimonov | See #1510 #3174 #2811 #4199 #6720 | 2019-09-09T22:04:08 | 2019-09-09T22:04:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,710,403 | 6,879 | alexey-milovidov | BTW, OPTIMIZE runs in background using single CPU core. The reason of high load average may be because of high IO usage in presence of concurrent queries. | 2019-09-09T23:51:43 | 2019-09-09T23:51:43 | {} | MEMBER |
ClickHouse | ClickHouse | 528,728,588 | 6,818 | joebling | @yuzhichang
That is how i create table and bitmap table:
CREATE TABLE qmkg_abtest.fd_qmkg_abtest_kboss_user_locate
(`ftime` Date COMMENT '',
`env_type` String COMMENT '',
`key` String,
`platform` Int64 COMMENT '',
`account_source` Int64 COMMENT '',
`app_ver` String COMMENT '',
`app_ver_id` String,
... | 2019-09-06T06:28:16 | 2019-09-06T06:28:16 | {} | NONE |
ClickHouse | ClickHouse | 528,732,907 | 6,050 | fessmage | Oh, thanks, i understood reason now. On server start i have problem with `clickhouse-odbc-bridge` - forgot to update it with rest packages. And with setting `dictionaries_lazy_load` defaults true - because of that error with odbc dictionaries, it resulted in all dictionaries not loaded. | 2019-09-06T06:45:08 | 2019-09-06T06:45:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,051,951 | 6,729 | alexey-milovidov | Performance test is clear. | 2019-09-07T00:10:17 | 2019-09-07T00:10:17 | {} | MEMBER |
ClickHouse | ClickHouse | 529,111,063 | 6,844 | etah000 | The same metric query should return the same result no matter what network protocol it uses. DBeaver uses http-based jdbc and clickhouse-client uses tcp. When both clients connect to the server, the query should return "httpconnection: 1; tcpconnection:1", but the result for "clickhouse-client" gets "httpconnection: 0"... | 2019-09-07T14:02:55 | 2019-09-07T14:02:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,174,385 | 6,862 | nicelulu | https://github.com/yandex/ClickHouse/issues/6834 | 2019-09-08T06:30:44 | 2019-09-08T06:30:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,479,940 | 467 | Prakash9944 | asdad | 2019-09-09T13:40:50 | 2019-09-09T13:40:50 | {
"confused": 5,
"eyes": 1
} | NONE |
ClickHouse | ClickHouse | 529,708,837 | 3,520 | filimonov | It's looks like a duplicate of https://github.com/yandex/ClickHouse/issues/1861
It was fixed in 19.1 | 2019-09-09T23:43:51 | 2019-09-09T23:43:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,833,667 | 6,834 | abyss7 | Can you provide the numbers of lost rows against total rows? And sample queries on how you compare both tables? Also, please, provide CH version. | 2019-09-06T12:23:52 | 2019-09-06T12:24:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,054,051 | 5,294 | alexey-milovidov | It's abandoned, but will be continued in a different form. | 2019-09-07T00:27:46 | 2019-09-07T00:27:46 | {} | MEMBER |
ClickHouse | ClickHouse | 529,106,857 | 6,830 | filimonov | Mark files has quite simple structure, while bin files structure is quite complicated.
The ClickHouse is opensource.
Please read the sources, there is no better way or step-by-step instructions.
The best way to read clickhouse files is to read it with ClickHouse.
If you have some concrete questions - pleas... | 2019-09-07T13:08:35 | 2019-09-07T13:08:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,677,211 | 5,974 | filimonov | Actually it's a duplicate of https://github.com/yandex/ClickHouse/issues/4197 (which have full stacktraces), so let's continue there. | 2019-09-09T21:37:25 | 2019-09-09T21:37:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,703,916 | 3,541 | filimonov | I think it's the duplicate of https://github.com/yandex/ClickHouse/issues/2507
and reporter provide no useful information, so closing this one.
Let's continue in #2507 | 2019-09-09T23:20:45 | 2019-09-09T23:20:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,831,172 | 6,833 | abyss7 | Can you provide listing of `SHOW PROCESSLIST` while receiving these exceptions?
This exception means that you have exceeded the queue size of global thread pool which is 10k. | 2019-09-06T12:15:27 | 2019-09-06T12:15:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,928,016 | 6,830 | filimonov | https://stackoverflow.com/a/57813669/1555175 | 2019-09-06T16:44:54 | 2019-09-06T16:45:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,097,366 | 6,843 | 243f6a8885a308d313198a2e037 | > Maybe better to always use initial_user to calculate per-user limits?
Yeah, this can solve our current problem. However...
> BTW, initial_user can be easily fabricated by client.
If my understanding is correct, limiting the memory usage for initial_user ( `max_memory_usage_for_initial_user` ?) might be unexp... | 2019-09-07T10:56:29 | 2019-09-07T10:56:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,674,558 | 3,917 | filimonov | Can't reproduce at 19.14:
```
c39ddd31ab1c :) CREATE TABLE foo
(
`date` Date COMMENT 'Date when the event occurred',
`campaign_id` Int32 COMMENT 'Campaign identifier' CODEC(ZSTD(22)),
`event_type` LowCardinality(String) COMMENT 'Type of event'
)
ENGINE = MergeTree()
ORDER BY campaign_id
Ok.
... | 2019-09-09T21:28:59 | 2019-09-09T21:28:59 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,655,251 | 6,840 | alexey-milovidov | Ok.
BTW, we also support `./clickhouse-client.xml` and `~/.clickhouse-client/config.xml`. | 2019-09-06T00:14:06 | 2019-09-06T00:14:06 | {} | MEMBER |
ClickHouse | ClickHouse | 528,800,613 | 6,365 | filimonov | @vzakaznikov is working on that in our internal CI.
We will probably sync in back later.
I'm closing that for now, we will either reopen / either create new PR later. | 2019-09-06T10:22:15 | 2019-09-06T10:22:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,809,837 | 6,814 | vitlibar | What do you think about `JSONExtract` function? It seems it already can split JSON into parts as you want:
```
CREATE TABLE test.jsnp(p Tuple(a Int16, b Int16, c Array(String))) ENGINE=MergeTree ORDER BY tuple();
INSERT INTO test.jsnp SELECT JSONExtract('{"a":5, "c":["x","y","z"], "b":7}', 'Tuple(a Int16, b Int16,... | 2019-09-06T10:56:48 | 2019-09-06T10:57:12 | {} | MEMBER |
ClickHouse | ClickHouse | 529,002,993 | 6,273 | h-abinaya28 | @pachico What is the workaround that you implemented for this use case? | 2019-09-06T20:28:37 | 2019-09-06T20:28:37 | {} | NONE |
ClickHouse | ClickHouse | 529,050,208 | 6,858 | alexey-milovidov | But it was on line 38: https://github.com/yandex/ClickHouse/pull/6858/files#diff-44ad2ac86d7c14e6e7256095d1b43628R38 | 2019-09-06T23:58:39 | 2019-09-06T23:58:39 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 529,054,206 | 4,354 | alexey-milovidov | Continued in two separate PRs. | 2019-09-07T00:29:06 | 2019-09-07T00:29:06 | {} | MEMBER |
ClickHouse | ClickHouse | 529,068,233 | 6,812 | sssrank | same problem in version 19.3.6.
At first time it works well, but problem happens after a few days. When I restart clickhouse, it works well again , but things whill go wrong again after a few hours or days .
Logs:
````
2019.09.07 11:30:17.017370 [ 12 ] {} <Error> abtest._abtest_biz_himovie_ope_inner (StorageR... | 2019-09-07T03:34:30 | 2019-09-07T03:34:30 | {} | NONE |
ClickHouse | ClickHouse | 529,116,963 | 6,844 | alexey-milovidov | @etah000 What is `snowball-client` and where to obtain it? | 2019-09-07T15:16:06 | 2019-09-07T15:16:06 | {} | MEMBER |
ClickHouse | ClickHouse | 529,154,676 | 3,238 | den-crane | fixed https://github.com/yandex/ClickHouse/pull/6324 ?
Works without the issue 19.15.1.1285
```
SELECT *
FROM anonymous_replacing
┌─eventId─┬─subjectId─┬───────────timestamp─┬─eventType────┐
│ id1 │ user_1 │ 2018-09-23 12:16:16 │ user/added │
│ id2 │ user_1 │ 2018-09-23 12:17:16 │ user/upda... | 2019-09-07T23:25:32 | 2019-09-07T23:25:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,211,695 | 6,844 | etah000 | > @etah000 What is `snowball-client` and where to obtain it?
Sorry, just a typo, it should be "ClickHouse client" . | 2019-09-08T15:09:00 | 2019-09-08T15:09:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,695,897 | 5,230 | filimonov | Let's continue in #3981 | 2019-09-09T22:45:29 | 2019-09-09T22:45:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,800,374 | 6,760 | vzakaznikov |
The issue itself was introduced in https://github.com/yandex/ClickHouse/commit/c96fa2c08030ab0fdbd31f9b054bbaec51bbe13f (v19.13.1.1057-testing) and fixed in https://github.com/yandex/ClickHouse/commit/d09c5c871b16ee10aff6ec8e7b1125034bfe25b4 (v19.14.1.1095-testing). | 2019-09-06T10:21:27 | 2019-09-06T10:21:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,879,386 | 6,814 | hczhcz | Is it possible to have some interaction with ["Nested"](https://clickhouse.yandex/docs/en/data_types/nested_data_structures/nested/) type? I.e. making "SmallestJSON" derived from "Nested" type. "SmallestJSON" may do the same thing as what "Nested" do in type checking, member access, etc, while it has extra functionalit... | 2019-09-06T14:34:26 | 2019-09-06T14:45:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,929,429 | 6,814 | zhang2014 | > Is it possible to have some interaction with "Nested" type? I.e. making "SmallestJSON" derived from "Nested" type. "SmallestJSON" may do the same thing as what "Nested" do in type checking, member access, etc, while it has extra functionality on reading/writing.
@hczhcz Why do you need it to be compatible with nes... | 2019-09-06T16:49:33 | 2019-09-06T16:49:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,195,253 | 6,273 | pachico | @h-abinaya28 I used a materialized column to preprocess part of the data so I didn't have to use a subquery.
In this case, in the table `source` I created a materialized column storing `timeSlots(toDateTime(date1, 'Zulu'), toUInt32((dateDiff('day', date1, date2) - 1) * dayseconds), toUInt32(dayseconds))` and I arrayJ... | 2019-09-08T11:49:10 | 2019-09-08T11:49:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,204,886 | 6,273 | den-crane | BTW, Original issue already fixed in master. Works in 19.15.1.1285. | 2019-09-08T13:52:33 | 2019-09-08T13:52:33 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,687,280 | 6,311 | filimonov | Most probably duplicate of https://github.com/yandex/ClickHouse/issues/2755 | 2019-09-09T22:11:37 | 2019-09-09T22:11:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,689,421 | 6,457 | filimonov | @bessonov87 did you check all other options from the error message?
1. Using newer version of compression library after server update.
2. Using another compression method.
3. Non-deterministic compression algorithm (highly unlikely)
4. Non-deterministic merge algorithm due to logical error in code.
5. Data corru... | 2019-09-09T22:19:27 | 2019-09-09T22:19:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,710,869 | 6,879 | volfco | I'm not seeing abnormally high IO usage.
For context, I thought running an OPTIMIZE after I altered a column to add compression would cause the column to be compressed. I removed that fact from the initial... | 2019-09-09T23:54:00 | 2019-09-09T23:54:00 | {} | NONE |
ClickHouse | ClickHouse | 528,951,694 | 6,814 | hczhcz | @zhang2014
It could be a way to unify data accessing in a specific column (like `xxx.url.string` in your example) without adding another semantic (parsing a JSON string vs. locating a column) to JSONExtract* functions.
Besides, I met similar cases of handling dynamically expanded structures when I was writing some ... | 2019-09-06T17:52:57 | 2019-09-06T17:52:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,112,639 | 6,844 | den-crane | @etah000 Because HTTP connections are very short. Just execute long running SQL in DBeaver `select sleep(3) from numbers(20) settings max_block_size=1`
and check :
```SELECT value
FROM system.metrics
WHERE metric LIKE '%HTTPConnection%'
┌─value─┐
│ 1 │
└───────┘
``` | 2019-09-07T14:23:22 | 2019-09-07T14:24:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,314,550 | 6,273 | pachico | Awesome news, @den-crane! Great work! | 2019-09-09T06:12:12 | 2019-09-09T06:12:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,469,779 | 6,867 | nicelulu | > Glancing over the code, I don't see an infinite loop, but the nextSize() logic looks broken in that it leads to a quadratic behavior when doing alignAlloc() beyond the linear_growth_threshold: we only allocate the size we need, which might be threshold + 8 bytes, then threshold + 16 bytes, etc., copying the entire ch... | 2019-09-09T13:18:16 | 2019-09-09T13:41:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,707,102 | 3,416 | filimonov | Duplicate of https://github.com/yandex/ClickHouse/issues/1576 and #3964 | 2019-09-09T23:35:51 | 2019-09-09T23:35:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,662,643 | 5,395 | filimonov | #6825 | 2019-09-06T00:51:04 | 2019-09-06T00:51:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,787,112 | 6,829 | filimonov | Maybe you just have old latest in your local repository, and didn't update that?
```
$ docker images
REPOSITORY TAG IMAGE ID CREATED SIZE
yandex/clickhouse-server latest eb1259c282e0 7 weeks ago ... | 2019-09-06T09:39:32 | 2019-09-06T09:39:32 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,937,954 | 6,843 | alexey-milovidov | BTW, initial_user can be easily fabricated by client. | 2019-09-06T17:16:02 | 2019-09-06T17:16:02 | {} | MEMBER |
ClickHouse | ClickHouse | 529,662,476 | 6,825 | filimonov | @clemrodriguez can you create a PR to review the change? | 2019-09-09T20:53:51 | 2019-09-09T20:53:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,669,191 | 6,870 | filimonov | In common case it's impossible to calculate the sums (just imagine that field_a is `String` or `Array`).
There is related feature called extremes: https://clickhouse.yandex/docs/en/query_language/select/#extreme-values
And you can always calculate that with 2 queries.
1) `SELECT * FROM ... `
2) `SELECT count... | 2019-09-09T21:12:45 | 2019-09-09T21:13:04 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 482,119,226 | 4,977 | den-crane | https://github.com/yandex/ClickHouse/issues/3174#issuecomment-423074171
https://github.com/yandex/ClickHouse/issues/2811#issuecomment-414464589
#6720 | 2019-04-11T13:42:29 | 2019-09-09T22:21:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 526,045,768 | 6,720 | filimonov | Possible problems:
1. You insert too often (use bulk inserts instead of single row inserts, optimal tempo - up to several inserts per second per server, good bulks size - from hundeds of rows up to 1 million row). You can do single row inserts via Buffer table if needed.
2. You insert in bulks but too fast and merges... | 2019-08-29T06:35:04 | 2019-09-09T22:23:07 | {
"+1": 8,
"eyes": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,693,221 | 5,994 | filimonov | Let's continue in #3981 | 2019-09-09T22:34:35 | 2019-09-09T22:34:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,701,101 | 4,049 | filimonov | Fixed in https://github.com/yandex/ClickHouse/pull/6278 | 2019-09-09T23:08:02 | 2019-09-09T23:08:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,799,413 | 6,829 | janson-git | Ouch!
`docker-compose pull yandex/clickhouse` it just shows "image is up to date".
But manual `docker pull yandex/clickhouse` gets the **real** latest!
Thanks!
| 2019-09-06T10:18:09 | 2019-09-06T10:18:09 | {} | NONE |
ClickHouse | ClickHouse | 528,837,589 | 6,834 | den-crane | > @den-crane no this reason, tmp_all and final_all table is not the same table
Please, show the output ` grep "already exists; ignoring it" /var/log/clickhouse-server/clickhouse-server.log` | 2019-09-06T12:36:53 | 2019-09-06T12:36:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,659,225 | 6,876 | stavrolia | For #5371 issue. | 2019-09-09T20:45:04 | 2019-09-09T20:45:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,698,022 | 3,941 | filimonov | In general - you should check the logs to find the reason / exact error which blocked pushing data forward.
If you have huge stream of data and you're overloading your hardware limits - most probably the exception will be "Too many parts..." ( check #6720 )
Closing issue for now (it's old, and rather non-actual)... | 2019-09-09T22:54:45 | 2019-09-09T22:54:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,649,549 | 6,840 | filimonov | I've tested that manually, and it looks ok. | 2019-09-05T23:47:59 | 2019-09-05T23:47:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,705,915 | 6,834 | florida-hehe | @den-crane no this reason, tmp_all and final_all table is not the same table | 2019-09-06T04:43:21 | 2019-09-06T04:43:21 | {} | NONE |
ClickHouse | ClickHouse | 529,478,174 | 6,825 | clemrodriguez | This bug is indeed coming from #5395 .
MySQL connection is closed at end of readImpl method even if there are still Block to come after the current one is full in object dbm>Formats>MySQLBlockInputStream.cpp
Removing connection closure at lines 133 & 134, fix the issue (locally reproduced and fixed).
Proposed fix st... | 2019-09-09T13:36:45 | 2019-09-09T13:36:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,687,144 | 2,339 | filimonov | Duplicate https://github.com/yandex/ClickHouse/issues/2755 | 2019-09-09T22:11:05 | 2019-09-09T22:11:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,889,610 | 6,843 | alexey-milovidov | Maybe better to always use initial_user to calculate per-user limits? | 2019-09-06T15:00:15 | 2019-09-06T15:00:15 | {} | MEMBER |
ClickHouse | ClickHouse | 528,923,536 | 6,814 | zhang2014 | > How do you plan to implement getting parts of SmallestJSON? Maybe it's possible to make function JSONExtract* work with this new data type? I think it would be useful to have the same interface for user.
Yes, it will reuse the current JSON functions. I'm developing this part : )
| 2019-09-06T16:31:08 | 2019-09-06T16:31:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,987,133 | 6,822 | Jack012a | very helpful.
Thank you! | 2019-09-06T19:36:42 | 2019-09-06T19:36:42 | {} | NONE |
ClickHouse | ClickHouse | 529,101,194 | 6,830 | kainever7 | > There were some examples in this talk (slides are in description), but most of it is in Russian: https://www.youtube.com/watch?v=7pyNmackUTw
Russian..oh...no... Do you have it in English? | 2019-09-07T11:51:09 | 2019-09-07T11:51:09 | {} | NONE |
ClickHouse | ClickHouse | 529,126,933 | 1,522 | alexey-milovidov | It should be fixed in master and in 19.14 (prestable). | 2019-09-07T17:12:14 | 2019-09-07T17:12:14 | {} | MEMBER |
ClickHouse | ClickHouse | 529,148,433 | 6,860 | den-crane | Yes, blocks. Try:
select number from numbers(10);
select number from numbers(10) settings max_block_size=1;
select number from numbers(10) settings max_block_size=1 format PrettyCompactMonoBlock;
| 2019-09-07T21:30:53 | 2019-09-07T21:30:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,458,599 | 6,867 | akuzm | Glancing over the code, I don't see an infinite loop, but the nextSize() logic looks broken in that it leads to a quadratic behavior when doing alignAlloc() beyond the linear_growth_threshold: we only allocate the size we need, which might be threshold + 8 bytes, then threshold + 16 bytes, etc., copying the entire chun... | 2019-09-09T12:52:03 | 2019-09-09T12:52:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,478,239 | 6,873 | akuzm | This has three commits, first two are cosmetic changes, and the last one introduces internal iteration. | 2019-09-09T13:36:56 | 2019-09-09T13:36:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,658,025 | 6,050 | filimonov | > I got NOT_LOADED state for every dictionary after update of clickhouse-server from 18.* to 19.* and only thing that worked for me - advice from @vitlibar. `system reload dictionary <dictionary_name>` for every dictionary - fixes problem.
Try disabling lazy loading of dictionaries if you need them preinitialized. | 2019-09-06T00:26:55 | 2019-09-06T00:26:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,774,070 | 6,818 | joebling | Another question:
#1.
select
ftime,
bitmapCardinality(uid_bitmap) dau
from
(
select
ftime,
groupBitmapState(toUInt64(uid)) as uid_bitmap
from qmkg_abtest.fd_qmkg_abtest_kboss_user_locate
group by ftime
) t
---
2019-08-26 41767618
2019-08-27 42019433
2019-08... | 2019-09-06T09:00:05 | 2019-09-06T09:00:50 | {} | NONE |
ClickHouse | ClickHouse | 528,834,994 | 6,818 | yuzhichang | Your usage of `view of distributed table` looks bad. The idiomatic way is to setup a view for the local table on every node， then setup a distributed view for those local views. See [this](https://github.com/yandex/ClickHouse/issues/467) for reference. | 2019-09-06T12:28:28 | 2019-09-06T12:28:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,836,982 | 6,844 | abyss7 | Because command-line doesn't use HTTP interface? Please describe your scenario better, in steps. | 2019-09-06T12:34:57 | 2019-09-06T12:34:57 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,051,626 | 6,815 | alexey-milovidov | Ok. | 2019-09-07T00:07:45 | 2019-09-07T00:07:45 | {} | MEMBER |
ClickHouse | ClickHouse | 529,100,302 | 4,736 | borismol | @filimonov i reproduced issue with 19.13.2.19 | 2019-09-07T11:38:25 | 2019-09-07T11:38:25 | {} | NONE |
ClickHouse | ClickHouse | 529,105,481 | 4,736 | filimonov | > @filimonov i reproduced issue with 19.13.2.19
It means nothing.
The **only** stable release having latest Kafka fixes is 19.11.9. It is pull requests #6805 and #6597 (merged 3 days ago).
You can also pick latest 19.14 version, for example v19.14.1.1274-prestable, or wait few days till Kafka fixes will be por... | 2019-09-07T12:49:57 | 2019-09-07T12:50:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,448,731 | 6,864 | abyss7 | Fixed in 19.14+ with https://github.com/yandex/ClickHouse/pull/6748 | 2019-09-09T12:26:37 | 2019-09-09T12:26:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,549,437 | 5,387 | akuzm | MongoDB URI support was added in Poco release 1.8.0. I'll try to merge it to our fork. | 2019-09-09T16:03:07 | 2019-09-09T16:03:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,782,196 | 6,814 | kevintop | Jsonstring performance is not good, if you can use jsonstring at write time and expand storage, it will be very good; this function will solve similar scenarios like Time Series tags storage. Great function, strong support.
| 2019-09-06T09:24:32 | 2019-09-06T09:24:32 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 528,792,392 | 1,827 | 631068264 | The same error in docker
```
clickhouse_1 | Poco::Exception. Code: 1000, e.code() = 0, e.displayText() = Not found: http_port (version 19.13.1.11 (official build))
clickhouse_1 | Poco::Exception. Code: 1000, e.code() = 0, e.displayText() = Not found: path (version 19.13.1.11 (official build))
clickhouse_1 | Poco... | 2019-09-06T09:55:33 | 2019-09-06T09:55:33 | {} | NONE |
ClickHouse | ClickHouse | 529,670,652 | 6,868 | filimonov | Try
```sql
select col1, col2, groupArray(col3) as col3
from tbl
group by col1, col2
``` | 2019-09-09T21:17:06 | 2019-09-09T21:17:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,696,689 | 3,981 | filimonov | That is caused by client side disconnect on http protocol, for example due to read timeout on client side. So server can't write anymore to the socket after client close that.
Test case:
```bash
wget 'http://localhost:8123/?query=SELECT%201,sleepEachRow(0.01)%20FROM%20numbers(200)' -qO- --read-timeout=1
```
T... | 2019-09-09T22:49:10 | 2019-09-09T22:49:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,873,139 | 5,231 | daniccan | Upgraded to the latest version which fixed the issue. Closing the issue. | 2019-09-06T14:18:21 | 2019-09-06T14:18:21 | {} | NONE |
ClickHouse | ClickHouse | 528,915,657 | 6,814 | vitlibar | > Yes , `JSONExtract` is a very useful function, but it must know the structure of JSON before using it.
Ok, seems I understand it now. Looks great!
How do you plan to implement getting parts of SmallestJSON? Maybe it's possible to make function JSONExtract* work with this new data type? I think it would be usefu... | 2019-09-06T16:07:39 | 2019-09-06T16:07:39 | {} | MEMBER |
ClickHouse | ClickHouse | 529,676,344 | 6,367 | alexey-milovidov | We can add info about default database at the time of query execution. | 2019-09-09T21:34:39 | 2019-09-09T21:34:39 | {} | MEMBER |
ClickHouse | ClickHouse | 528,871,939 | 6,814 | zhang2014 | Yes , `JSONExtract` is a very useful function, but it must know the structure of JSON before using it. | 2019-09-06T14:15:01 | 2019-09-06T14:15:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,212,610 | 6,844 | etah000 | @den-crane I am a little confused. You mean that even a jdbc-based GUI client keeps a connection session to a Clickhouse server, but actually it has no lasting long connection to the server. The connection is valid/establed only when it submits a query. After it gets the result, it disconnects the connection.
I... | 2019-09-08T15:20:05 | 2019-09-08T15:20:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,217,354 | 6,844 | den-crane | >but actually it has no lasting long connection to the server
yes.
>Is my understanding correct?
JDBC creates a new HTTP connection for every query.
But there are CH sessions (even JDBC) https://clickhouse.yandex/docs/en/interfaces/http/ . You can set through JDBC the session_id=.... and have one session ... | 2019-09-08T16:14:01 | 2019-09-08T16:20:48 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,451,203 | 6,868 | alexey-milovidov | It's named `groupArray`. | 2019-09-09T12:33:07 | 2019-09-09T12:33:07 | {} | MEMBER |
ClickHouse | ClickHouse | 529,706,098 | 3,446 | filimonov | Simpler test-case:
```
select notEmpty(toUUID('00000000-0000-0000-0000-000000000000'));
``` | 2019-09-09T23:30:51 | 2019-09-09T23:30:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 529,711,152 | 6,879 | alexey-milovidov | What is the result of `SELECT * FROM system.merges`? | 2019-09-09T23:55:22 | 2019-09-09T23:55:22 | {} | MEMBER |
ClickHouse | ClickHouse | 529,768,809 | 6,063 | tonal | Also mysql db don`t drop if not connect to it:
```
host2 :) CREATE DATABASE mysql_db ENGINE = MySQL('yandex.ru:3306', 'test_db', 'yandex_admin', '1234');
CREATE DATABASE mysql_db
ENGINE = MySQL('yandex.ru:3306', 'test_db', 'yandex_admin', '1234')
Ok
0 rows in set. Elapsed: 0.064 sec.
host2 :) show databas... | 2019-09-10T04:53:24 | 2019-09-10T04:53:24 | {} | NONE |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.