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 | 562,494,169 | 7,860 | nikitamikhaylov | No need to change submodule in documentation PR. | 2019-12-06T09:12:40 | 2019-12-06T09:12:40 | {} | MEMBER |
ClickHouse | ClickHouse | 562,495,799 | 7,646 | jsding | Only 1 table /server to read that topic.
We have to move to sinker program to consume that big topics(3000 per sec now) to clickhouse and works now.
As we saw, Kafka engine only polls for 5 seconds and leave the consume group, that makes the lag bigger and bigger.
| 2019-12-06T09:17:34 | 2019-12-06T09:17:34 | {} | NONE |
ClickHouse | ClickHouse | 562,372,100 | 7,792 | vzakaznikov | @KochetovNicolai, could you review the changes? I have managed to make subqueries work with live view tables. The implementation is not ideal due to a bug (https://github.com/ClickHouse/ClickHouse/issues/8057). I had to create a dummy context and use a temporary "external" table.
Here is the summary of the changes... | 2019-12-05T23:59:37 | 2019-12-06T16:06:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,743,464 | 8,069 | den-crane | Hi, to reproduce this we need an example of your parquet file and query/API you used to ingest. | 2019-12-06T21:28:25 | 2019-12-06T21:28:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,881,121 | 8,073 | nvartolomei | Might be useful to return something like table below? Especially for cases with multiple volumes.
|shadow|part name|path to part|
|--------|----------|------------|
|***increment** or **name** (if WITH NAME)*|*part name*|*path*|
| 2019-12-07T19:53:52 | 2019-12-07T19:53:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,922,920 | 8,051 | alesapin | Performance tests flap. | 2019-12-08T08:10:19 | 2019-12-08T08:10:19 | {} | MEMBER |
ClickHouse | ClickHouse | 563,015,644 | 8,080 | alexey-milovidov | Please review the https://github.com/ClickHouse/ClickHouse/pull/8086 | 2019-12-09T00:21:25 | 2019-12-09T00:21:25 | {} | MEMBER |
ClickHouse | ClickHouse | 561,938,820 | 3,929 | zhang2014 | @alesapin Yep, but I have some other tasks to complete that have higher priority in our plan | 2019-12-05T02:31:14 | 2019-12-05T02:31:14 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 561,708,889 | 7,837 | filimonov | > No exacly-once guarantee.
Kafka by default can work only in 'at-least-once' or 'at-most-once' models. To have guarantees you need to use exactly-once semantics (EOS), the producer also needs to support that on its end to make that work.
Support for EOS semantics is still in progress on rdkafka side, see https:/... | 2019-12-04T15:58:10 | 2019-12-05T06:09:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,049,068 | 7,859 | excitoon | Separated information about paths to #8043 | 2019-12-05T09:37:35 | 2019-12-05T09:37:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,091,711 | 7,665 | 4ertus2 | > @4ertus2 can you check my addition to the pr description? Is it correct?
Yup. It's OK.
> Also, now `t1 ANY LEFT JOIN t2` equals `t2 ANY RIGHT JOIN t1`. Did it change previous behaviour somehow? What have changed if so?
```ANY LEFT JOIN``` haven't been changed. ```ANY RIGHT JOIN``` with ```any_join_distinct_r... | 2019-12-05T11:33:56 | 2019-12-05T11:33:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,106,128 | 8,028 | nikitamikhaylov | Conflicts with 19.16
| 2019-12-05T12:20:39 | 2019-12-05T12:20:39 | {} | MEMBER |
ClickHouse | ClickHouse | 562,369,247 | 8,021 | den-crane | There is no other way with this definition of the task.
I believe you don't need join at all. Use a wide table or external dictionaries. | 2019-12-05T23:47:47 | 2019-12-05T23:48:01 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,450,808 | 8,039 | toby080508050805 | I try to doing some test about parameter "--max_insert_block_size 33554432". I change it to 1481089. Memory growth is decrease from 1GB to 100Mb from every insert query and insert the same file. | 2019-12-06T06:30:46 | 2019-12-06T06:50:46 | {} | NONE |
ClickHouse | ClickHouse | 562,563,689 | 8,064 | filimonov | How can we reproduce such a problem? What is 10.0.19.30 ? Why it open/close connections to ClickHouse all the time?
Does the crash repeats? | 2019-12-06T13:05:17 | 2019-12-06T13:05:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,742,728 | 7,785 | nvartolomei | Workaround
```sql
ALTER TABLE system.query_log
FREEZE WITH NAME 'random-string-xyz'
```
```txt
Freezing part 201912_41_41_0 snapshot will be placed at /home/nv/clickhouse-testbed/srv1/data/shadow/random%2Dstring%2Dxyz
``` | 2019-12-06T21:25:52 | 2019-12-06T21:25:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,104,307 | 5,126 | 4ertus2 | New `ANY JOIN` logic would be default starting from 19.18. Old logic is default before 19.14. Versions 19.14 - 19.17 has no default logic for `ANY INNER` and `ANY RIGHT` joins (it's possible to enable old logic using `any_join_distinct_right_table_keys`) | 2019-12-05T12:14:27 | 2019-12-05T12:14:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 561,933,097 | 8,027 | jiugem | > Why do you need tcmalloc instead of jemalloc?
If jemalloc is not used, the error is as follows：
[ 3%] Building CXX object dbms/src/Common/StringUtils/CMakeFiles/string_utils.dir/StringUtils.cpp.o
[ 3%] Building C object contrib/jemalloc-cmake/CMakeFiles/jemalloc.dir/__/jemalloc/src/pages.c.o
/home/clickho... | 2019-12-05T02:06:14 | 2019-12-06T08:59:55 | {} | NONE |
ClickHouse | ClickHouse | 562,560,031 | 8,061 | filimonov | The thing is that max doesn't return null here.
```
SELECT max(number) FROM numbers(0)
┌─max(number)─┐
│ 0 │
└─────────────┘
```
In recent ClickHouse versiosn you can use `maxOrNull` instead max to get nulls in case of empty dataset. | 2019-12-06T12:53:19 | 2019-12-06T12:53:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,592,505 | 8,064 | shuliakovsky | > > https://gitweb.gentoo.org/repo/gentoo.git/tree/dev-db/clickhouse/clickhouse-19.15.2.2.ebuild
> > we build clickhouse from that ebuild
>
> I'm afraid that could be the reason. Didn't you consider using ClickHouse binary from official yandex builds? It should be portable between linux distros.
>
> If you want ... | 2019-12-06T14:29:06 | 2019-12-06T14:29:06 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 562,091,618 | 7,953 | amosbird | Discussed offline that this isn't suitable. Developers should have dedicated tools to format their PRs only, instead of reformatting the codebase | 2019-12-05T11:33:32 | 2019-12-05T11:33:32 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 562,199,977 | 8,047 | alesapin | `00094_union_race_conditions_5` not related to changes. | 2019-12-05T16:15:24 | 2019-12-05T16:15:24 | {} | MEMBER |
ClickHouse | ClickHouse | 562,414,765 | 8,039 | toby080508050805 | Following attachment is whole log. In the very last page has record server restart moment. OOM dose not worked at this time so server crash at 2019.12.02 02:10:05.211791. I manually restart a clickhouse server at 16:47:32.725068. Attachment has a full log.
> 2019.12.02 02:01:27.948796 [ 94 ] {c6d771f5-35e7-4ec1-86... | 2019-12-06T03:14:56 | 2019-12-06T03:15:47 | {} | NONE |
ClickHouse | ClickHouse | 562,553,606 | 8,037 | nikitamikhaylov | Conflicts with 19.16 | 2019-12-06T12:30:43 | 2019-12-06T12:30:43 | {} | MEMBER |
ClickHouse | ClickHouse | 562,574,262 | 8,064 | shuliakovsky | > What kind of build do you use?
>
> ```
> 2019.12.06 12:37:43.695434 [ 155 ] {} <Fatal> BaseDaemon: ########################################
> 2019.12.06 12:37:43.695508 [ 155 ] {} <Fatal> BaseDaemon: (version 19.15.2.2) (from thread 143) Received signal Broken pipe (13).
> 2019.12.06 12:37:43.695566 [ 155 ] {} ... | 2019-12-06T13:37:18 | 2019-12-06T13:37:18 | {} | NONE |
ClickHouse | ClickHouse | 562,743,614 | 7,785 | alexey-milovidov | Ok. BTW, It's also possible to get the info with `SET send_logs_leve = 'trace'`. | 2019-12-06T21:28:55 | 2019-12-06T21:29:10 | {} | MEMBER |
ClickHouse | ClickHouse | 562,829,839 | 8,069 | filimonov | Looks like your data just contains some fragments of postgres queries.
ClickHouse can't generate postgres SQL :) So it definitely cames from outside of clickhouse. | 2019-12-07T08:47:06 | 2019-12-07T08:47:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 561,937,033 | 8,038 | alexey-milovidov | @Vxider is going to implement it. | 2019-12-05T02:23:20 | 2019-12-05T02:23:20 | {} | MEMBER |
ClickHouse | ClickHouse | 562,574,108 | 8,064 | shuliakovsky | > How can we reproduce such a problem? What is 10.0.19.30 ? Why it open/close connections to ClickHouse all the time?
it's our service IP which write/read to/from clickhouse
>
> Does the crash repeats?
yes often, few times per day
| 2019-12-06T13:36:49 | 2019-12-06T13:36:49 | {} | NONE |
ClickHouse | ClickHouse | 562,826,577 | 8,070 | filimonov | Looks like duplicate of https://github.com/ClickHouse/ClickHouse/issues/7960 | 2019-12-07T08:02:21 | 2019-12-07T08:02:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,832,325 | 8,043 | excitoon | It's not Ok | 2019-12-07T09:17:12 | 2019-12-07T09:17:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,885,538 | 8,073 | nvartolomei | See https://github.com/nvartolomei/ClickHouse/pull/1/files for example. | 2019-12-07T20:54:14 | 2019-12-07T20:54:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,992,806 | 8,080 | alexey-milovidov | Ok, we should rename it to `geoDistance`. | 2019-12-08T20:57:25 | 2019-12-09T00:08:18 | {} | MEMBER |
ClickHouse | ClickHouse | 562,066,964 | 7,665 | KochetovNicolai | @4ertus2 can you check my addition to the pr description? Is it correct?
Also, now `t1 ANY LEFT JOIN t2` equals `t2 ANY RIGHT JOIN t1`. Did it change previous behaviour somehow? What have changed if so?
Seems like `ANY RIGHT JOIN` was changed. Before `any` was for keys from right table, now - for left. Is it corre... | 2019-12-05T10:23:58 | 2019-12-05T10:52:15 | {} | MEMBER |
ClickHouse | ClickHouse | 562,182,944 | 7,482 | alesapin | Fix https://github.com/ClickHouse/ClickHouse/pull/8051. | 2019-12-05T15:37:04 | 2019-12-05T15:37:04 | {} | MEMBER |
ClickHouse | ClickHouse | 562,370,512 | 8,020 | den-crane | Check this https://github.com/ClickHouse/ClickHouse/issues/7961#issuecomment-559833272 | 2019-12-05T23:53:03 | 2019-12-05T23:53:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,410,784 | 8,041 | kaijianding | though test_view_local is not replicated, the based table test_local is ReplicatedMergeTree.
the test_view_local only sees the local inserts into test_local, but can't see the inserts from another test_local replica? | 2019-12-06T02:55:10 | 2019-12-06T02:55:10 | {} | NONE |
ClickHouse | ClickHouse | 562,425,242 | 8,041 | den-crane | MV never reads a source table. It just catches inserted rows from inserts. So it does not get anything from inserts at another node. You should create MV with ReplicatedAggregatingMergeTree engine. | 2019-12-06T04:15:04 | 2019-12-06T04:15:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,570,669 | 7,962 | filimonov | Hm.
AFAIK roaring bitmap support only 32bit numbers.
https://github.com/RoaringBitmap/roaring/issues/136
> generate random numbers in shell with multiple threads
cool script. :) bit in clickhouse you can do much simpler:
```
create table test Engine=MergeTree() ORDER BY tuple() SETTINGS index_granularity ... | 2019-12-06T13:26:35 | 2019-12-06T13:26:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,575,547 | 8,061 | ghuname | Well, actually it returns '0000-00-00 00:00:00'. Is it null datetime or not? | 2019-12-06T13:41:17 | 2019-12-06T13:41:17 | {} | NONE |
ClickHouse | ClickHouse | 562,950,585 | 8,080 | hombit | Comment to the new `greatCircleDistance` says that the origin of the code is
https://github.com/sphinxsearch/sphinx/blob/409f2c2b5b2ff70b04e38f92b6b1a890326bad65/src/sphinxexpr.cpp#L3825.
Tests of this project compare the output of this code with [Vincenty's formula](https://en.wikipedia.org/wiki/Vincenty%27s_form... | 2019-12-08T13:51:09 | 2019-12-08T14:00:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 561,940,335 | 8,014 | zhang2014 | GRANULARITY is too big. Can it be adjusted to 3 or 5? detailed information see: https://github.com/ClickHouse/ClickHouse/issues/7517#issuecomment-549224927
For `granularity = 5`, bloom filter will build an index for the adjacent 5 granlules data block(one granlule size is specified when creating the MergeTree, the d... | 2019-12-05T02:37:55 | 2019-12-05T02:49:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,414,430 | 6,943 | nicelulu | > It would be nice to reduce pre-allocated column buffer size when number of columns is large.
Do you have any practices? How much is more appropriate. Will the reduction result in more frequent IO requests? | 2019-12-06T03:13:10 | 2019-12-06T03:13:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,764,001 | 8,067 | den-crane | How much memory your server has?
>'max_memory_usage': 312500000,
>'max_bytes_before_external_group_by': 156250000,
>would use 300.50 MiB / maximum: 298.02 MiB
>would use 312.69 MiB / maximum: 309.94 MiB
I don't see any contradiction here. max_memory_usage is a limit.
A query failed because it needs mor... | 2019-12-06T22:35:23 | 2019-12-06T22:36:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,804,872 | 8,068 | Vxider | My compiler version is `gcc version 8.2.1 20180905 (Red Hat 8.2.1-3) (GCC)` | 2019-12-07T02:23:20 | 2019-12-07T02:23:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 563,103,045 | 7,962 | kaijianding | @filimonov It would be wonderful if the 64-bit bitmap is supported, it would be very useful for a scenario like order_id, user_id, product_id | 2019-12-09T07:32:11 | 2019-12-09T07:32:11 | {} | NONE |
ClickHouse | ClickHouse | 562,235,876 | 8,014 | volfco | @zhang2014 Thanks for the information.
I've tried to recreate the OOM situation in my QA cluster and cannot replicate it. It might have just been a perfect storm of pending merges when the overly granular index was applied. | 2019-12-05T17:41:49 | 2019-12-05T17:41:49 | {} | NONE |
ClickHouse | ClickHouse | 562,385,917 | 7,812 | yuzhichang | @Akazz Any comments on this PR? | 2019-12-06T00:59:32 | 2019-12-06T00:59:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,401,820 | 8,046 | amosbird | btw, the problematic aligned_alloc symbol is from contrib/capnproto-cmake/libkj.a
| 2019-12-06T02:11:40 | 2019-12-06T02:11:40 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 560,528,951 | 2,764 | 4ertus2 | Implemented in #7665 | 2019-12-02T18:49:26 | 2019-12-05T12:15:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,360,410 | 8,017 | byx313 | thx you so much! | 2019-12-05T23:13:17 | 2019-12-05T23:13:17 | {} | NONE |
ClickHouse | ClickHouse | 562,366,914 | 8,041 | den-crane | test_view_local should replicated as well (ReplicatedAggregatingMergeTree)
`MV test_view_local .... FROM test_local` sees only local inserts into test_local.
So each replica has own partial in test_view_local.
If wanted exactly this then you need to describe one more cluster `cluster02` and place each CH node as s... | 2019-12-05T23:37:49 | 2019-12-05T23:41:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,072,306 | 8,029 | nikitamikhaylov | Conflicts with 19.17 and 19.16
| 2019-12-05T10:37:39 | 2019-12-05T12:23:35 | {} | MEMBER |
ClickHouse | ClickHouse | 562,184,089 | 8,049 | filimonov | Мне кажется что скорее вариант с огруглением должен быть по умолчанию, а вариант увеличивающий кол-во знаков после запятой как-то типа accurateMultiply( ... ). Но наверное можно и наоборот.
Надо перепроверить как это выглядит в других базах, и в частности какое именно огругление там обычно применяется (типичный use... | 2019-12-05T15:39:26 | 2019-12-05T15:41:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,416,800 | 8,021 | k1my0 | I'm sorry. It seems my question was not good.
I just created a temporary table because I couldn't use my own table.
**In short, I want to put context information in each src ip(key), dst ip(key). If some IP does not have any context, I'll put 'etc' instead of.**
This mean I thought, First I have to join with src... | 2019-12-06T03:25:45 | 2019-12-06T03:28:22 | {} | NONE |
ClickHouse | ClickHouse | 562,551,743 | 7,637 | nikitamikhaylov | Conflicts with 19.16 and 19.17
| 2019-12-06T12:24:23 | 2019-12-06T12:27:01 | {} | MEMBER |
ClickHouse | ClickHouse | 562,589,813 | 8,064 | filimonov | > https://gitweb.gentoo.org/repo/gentoo.git/tree/dev-db/clickhouse/clickhouse-19.15.2.2.ebuild
> we build clickhouse from that ebuild
I'm afraid that could be the reason. Didn't you consider using ClickHouse binary from official yandex builds? It should be portable between linux distros.
If you want to build by ... | 2019-12-06T14:21:43 | 2019-12-06T14:27:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,596,474 | 8,064 | alexey-milovidov | > /usr/lib64/libPocoNet.so.60
This indicates wrong, broken build.
We don't use shared linking and we don't use libraries from the system. | 2019-12-06T14:40:08 | 2019-12-06T14:40:14 | {} | MEMBER |
ClickHouse | ClickHouse | 562,600,678 | 7,646 | filimonov | > polls for 5 seconds and leave the consume group
it should not leave the consumer group. It explicitly exits the group in case of error (need to be checked what kind of error was that).
Kafka engine can consume about 10k-200k rows per second (depending on used format, size of the row and hardware). | 2019-12-06T14:50:57 | 2019-12-06T14:54:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 563,024,804 | 7,307 | alexey-milovidov | Now it's about three times. | 2019-12-09T01:25:15 | 2019-12-09T01:25:15 | {} | MEMBER |
ClickHouse | ClickHouse | 563,084,178 | 8,022 | ghost | @alesapin
I fixed some codes to pass the tests.
Let me know if there is any other things in order to start code review.
Thank you. | 2019-12-09T06:27:33 | 2019-12-09T06:27:33 | {} | NONE |
ClickHouse | ClickHouse | 562,106,445 | 8,039 | alesapin | Hi, thank you for the report. Can you provide logs from `/var/log/clickhouse-server/clickhouse-server.log` at the moment when server restarts (killed by OOM)? | 2019-12-05T12:21:34 | 2019-12-05T12:21:34 | {} | MEMBER |
ClickHouse | ClickHouse | 562,180,279 | 8,049 | 4ertus2 | Я вижу это как отдельную функцию `scaledMultiply(a, b, result_scale)`. Если мы обрезаем точность операции, стоит задавать ее явно. Функция может принимать 2 аргумента, тогда result_scale принимается scale-ом первого аргумента. | 2019-12-05T15:31:02 | 2019-12-05T15:31:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,296,128 | 7,120 | filimonov | > Is it abandoned?
Paused for now, but I want to continue / finish that. Most probably i will have some time for that in January. | 2019-12-05T20:08:30 | 2019-12-05T20:08:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,377,992 | 8,015 | den-crane | > <max_block_size>259840308</max_block_size>
Don't do it. 65k is optimal for the most usecases.
Check a size of duration column
```
select formatReadableSize(sum(column_data_compressed_bytes) as x) size,
formatReadableSize(sum(column_bytes_on_disk)),
formatReadableSize(sum(column_data_uncompre... | 2019-12-06T00:23:55 | 2019-12-06T00:29:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,385,666 | 8,030 | den-crane | Workarounds (copy from Telegram):
1. Insert through Null table https://gist.github.com/den-crane/f7382cd4f1f859ff6ac46afe7dc9925a
2. SimpleAggregateFunction
```
create table test_table
( x Int64,
cnt UInt32,
min_ts SimpleAggregateFunction(min, UInt64) Codec(LZ4),
max_ts SimpleAggregateFuncti... | 2019-12-06T00:58:29 | 2019-12-06T00:58:29 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,516,168 | 7,468 | alesapin | Maximum part size `max_bytes_to_merge_at_max_space_in_pool` doesn't related to mutations. Actually mutations can only decrease the size of the part, so they don't perform any size checks for result part size (only checks for available space on disk).
Changing maximum parts size is not recommended, but you can do it... | 2019-12-06T10:18:24 | 2019-12-06T10:18:24 | {} | MEMBER |
ClickHouse | ClickHouse | 562,587,927 | 8,061 | filimonov | It is NOT null. It is default value for DateTime (it's zero for ints, '0000-00-00 00:00:00' for DateTime, emty string for string etc) | 2019-12-06T14:16:38 | 2019-12-06T14:16:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,950,567 | 8,080 | hombit | Comparison new and old implementation with [WGS web calculator](https://www.cqsrg.org/tools/GCDistance/)
**Current implementation**
```sql
SELECT greatCircleDistance(0., 0., 0., 1.)
```
```text
┌─greatCircleDistance(0., 0., 0., 1.)─┐
│ 110567.33 │
└───────────────────────────────────... | 2019-12-08T13:50:54 | 2019-12-08T20:54:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 563,140,552 | 6,263 | hustnn | @alexey-milovidov @CurtizJ
Is any more detail about polymorphic parts? Thanks. | 2019-12-09T09:22:18 | 2019-12-09T09:22:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,097,880 | 7,799 | jarhed | I too am getting this error with:
`DB::Exception: Method createColumn() is not implemented for data type Set (version 19.17.4.11 (official build))`
```sql
SELECT
placement,
sum(impressions) as impressionSum,
sum(validPOI) as insidePOI,
sum(validDeviceId) as validDeviceId,
sum(cellularCentroid) as cellu... | 2019-12-05T11:54:31 | 2019-12-05T11:59:37 | {} | NONE |
ClickHouse | ClickHouse | 562,118,333 | 8,040 | nvartolomei | @SreekanthMannari few settings to keep in mind
* parallel_view_processing
* max_threads (affects number of views to which you can write in parallel if parallel_view_processing is set, should be set for the user who initiates INSERT).
Does it scale? There is nothing better than benchmarks with real data on real s... | 2019-12-05T12:57:52 | 2019-12-05T12:57:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,341,448 | 7,646 | filimonov | > Here is the kafka-consumer-group output, I checked and confirmed all kafka clickhouse table use
> kafka_num_consumers = 1 settings
How many tables / servers do you have which try to read that topic? Is it really only one table on one server, as `kafka-consumer-groups.sh` output suggests? or you just have a single... | 2019-12-05T22:10:32 | 2019-12-05T22:11:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,963,113 | 8,069 | shawel | OK thank you. i will investigate outside of clickhouse | 2019-12-08T15:58:42 | 2019-12-08T15:58:42 | {} | NONE |
ClickHouse | ClickHouse | 562,061,549 | 8,031 | avointsev | Den, thanks for the answer.
Could you say if this possible to control this behavior? For me prefer to get an exception instead of an incorrect result, if someone just does a mistake with date.
Unfortunately, I couldn't find these settings in the documentation ( | 2019-12-05T10:10:37 | 2019-12-05T10:13:50 | {} | NONE |
ClickHouse | ClickHouse | 562,073,163 | 7,979 | nikitamikhaylov | Conflicts with 19.17
| 2019-12-05T10:39:50 | 2019-12-05T10:39:50 | {} | MEMBER |
ClickHouse | ClickHouse | 551,275,340 | 7,665 | 4ertus2 | #5126 #7665 | 2019-11-07T21:34:30 | 2019-12-05T12:16:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,185,595 | 8,051 | alesapin | Now we will throw an exception with some trash locales like `en-US-u-va-posix`, but some users may use them, because ICU fallback to some default locale in case of strange name. | 2019-12-05T15:42:50 | 2019-12-05T15:42:50 | {} | MEMBER |
ClickHouse | ClickHouse | 562,185,795 | 8,051 | alesapin | Good examples https://github.com/mozilla/gecko-dev/blob/27ded6834ef8b61fa52838acd59fe617bf44c61c/js/src/builtin/intl/SharedIntlData.cpp. | 2019-12-05T15:43:18 | 2019-12-05T15:43:18 | {} | MEMBER |
ClickHouse | ClickHouse | 562,383,913 | 8,015 | den-crane | Example with MV
```
CREATE TABLE web_access_logs (
date Date,
duration UInt64)
Engine = MergeTree order by date
create materialized view web_access_logs_mv Engine=AggregatingMergeTree partition by date order by date
as select date, sumState(toUInt64(1)) cntduration, avgState(duration) avgduration, medianStat... | 2019-12-06T00:49:51 | 2019-12-06T00:51:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 563,144,190 | 8,045 | onine007 | I have the same question . | 2019-12-09T09:31:38 | 2019-12-09T09:31:38 | {} | NONE |
ClickHouse | ClickHouse | 563,151,667 | 8,067 | WesleyBatista | > How much memory your server has?
**20 GB**
@den-crane, thank you for your time :)
The examples were to show that the memory control feels like not working, and when it does it works by throwing error instead of keeping the query running with limited resources.
For instance, the same query runs if I provide ... | 2019-12-09T09:50:53 | 2019-12-09T09:50:53 | {} | NONE |
ClickHouse | ClickHouse | 562,136,971 | 8,048 | den-crane | duplicate: https://github.com/ClickHouse/ClickHouse/issues/7195 | 2019-12-05T13:51:31 | 2019-12-05T13:51:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,334,453 | 8,016 | filimonov | OK. I've tested all formats.
Formats tested:
CSV CSVWithNames JSONEachRow Native RowBinary RowBinaryWithNamesAndTypes TSKV TSV TSVWithNames TSVWithNamesAndTypes Values CapnProto ORC Parquet Protobuf Template
'Values' format still broken, I was trying to fix that, but without success for now (tricky buffers). No... | 2019-12-05T21:50:22 | 2019-12-05T21:59:38 | {
"hooray": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,907,708 | 7,998 | huang--yijian | > Merged, we will backport this fix to 19.17, 19.16 and 19.15 if possible.
Thank you for the quick support.
When the fix is released in the yum packages to install, please notify us. | 2019-12-08T03:01:41 | 2019-12-08T03:01:41 | {} | NONE |
ClickHouse | ClickHouse | 562,042,218 | 8,042 | excitoon | @alexey-milovidov | 2019-12-05T09:19:46 | 2019-12-05T09:19:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,105,641 | 7,747 | nikitamikhaylov | Conflicts with 19.16
| 2019-12-05T12:18:56 | 2019-12-05T12:18:56 | {} | MEMBER |
ClickHouse | ClickHouse | 562,382,283 | 7,773 | vzakaznikov | This issue should fixed by https://github.com/ClickHouse/ClickHouse/pull/7792.
Here is a manual test:
```
Ubuntu-1804-bionic-64-minimal :) CREATE TABLE A (id Int32) Engine=Memory;
Ubuntu-1804-bionic-64-minimal :) CREATE TABLE B (id Int32, name String) Engine=Memory;
Ubuntu-1804-bionic-64-minimal :) SET allow_e... | 2019-12-06T00:42:27 | 2019-12-06T00:46:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,548,032 | 6,269 | Mefydze | we also have this trouble with version 19.16.2.2
```
2019.12.06 10:07:23.784032 [ 25 ] {} <Error> default.rawlog_shard: DB::StorageReplicatedMergeTree::queueTask()::<lambda(DB::StorageReplicatedMergeTree::LogEntryPtr&)>: Code: 173, e.displayText() = DB::ErrnoException: Allocator: Cannot rea
lloc from 1.00 MiB to ... | 2019-12-06T12:09:51 | 2019-12-06T12:14:28 | {} | NONE |
ClickHouse | ClickHouse | 562,564,282 | 8,064 | filimonov | What kind of build do you use?
```
2019.12.06 12:37:43.695434 [ 155 ] {} <Fatal> BaseDaemon: ########################################
2019.12.06 12:37:43.695508 [ 155 ] {} <Fatal> BaseDaemon: (version 19.15.2.2) (from thread 143) Received signal Broken pipe (13).
2019.12.06 12:37:43.695566 [ 155 ] {} <Fatal> Base... | 2019-12-06T13:07:09 | 2019-12-06T13:07:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,583,560 | 8,016 | filimonov | Test used is here: https://github.com/filimonov/ch-kafka-consume-perftest/blob/master/test_formats.sh
@vzakaznikov have some plans for integarting those tests. | 2019-12-06T14:04:16 | 2019-12-06T14:04:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,696,440 | 8,068 | alexey-milovidov | What compiler do you use?
(Why we don't see this issue in our CI?) | 2019-12-06T18:59:52 | 2019-12-06T18:59:52 | {} | MEMBER |
ClickHouse | ClickHouse | 562,879,075 | 8,043 | excitoon | Now it's Ok | 2019-12-07T19:23:59 | 2019-12-07T19:23:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 563,134,100 | 8,011 | alesapin | CI checks don't start because we cannot clone submodules:
```
fatal: reference is not a tree: 5666c94dc90adf1aa8cb66527b322af1ec85bcf6
fatal: reference is not a tree: 6cd01c101e233691adb27d7a55b267436673940a
fatal: reference is not a tree: d601f7b4949f6fd64a84e88a126359b734aa56d8
Unable to checkout '5666c94dc90adf... | 2019-12-09T09:05:41 | 2019-12-09T09:05:41 | {} | MEMBER |
ClickHouse | ClickHouse | 562,060,627 | 7,942 | godfreyd | @hczhcz , Fixed | 2019-12-05T10:08:06 | 2019-12-05T10:08:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,072,894 | 7,981 | nikitamikhaylov | Conflicts with 19.17
| 2019-12-05T10:39:10 | 2019-12-05T10:39:10 | {} | MEMBER |
ClickHouse | ClickHouse | 562,745,593 | 8,070 | den-crane | >КХ выдаёт ОК и 0 рядов вставлено
А что когда-то КХ выдавал что-то другое????
КХ не знает сколько строк вставил и никогда не знал.
18.14.18. -- версия 2018 года
create table x (A Int64) Engine=Memory;
insert into x select * from remote('127.0.0.1', system, one);
**0 rows in set. Elapsed: 0.002 sec.**
... | 2019-12-06T21:35:36 | 2019-12-06T21:38:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,797,880 | 8,062 | den-crane | Это все пахнет cost-based оптимизатором который учитывает стоимость random i/o и знает что миллион раз сделать лукап (nested-loop) по индексу в A дешевле чем full scan + HJ.
КМК абсолютно бесмысленная трудоемкая затея, чуть более чем полностью для OLAP.
Я не уверен что результаты 1 и 2 всегда равны, с учетом возмож... | 2019-12-07T01:10:48 | 2019-12-07T01:17:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 562,832,469 | 8,070 | sgolod | > > КХ выдаёт ОК и 0 рядов вставлено
>
> А что когда-то КХ выдавал что-то другое????
> КХ не знает сколько строк вставил и никогда не знал.
>
> 18.14.18. -- версия 2018 года
>
> create table x (A Int64) Engine=Memory;
> insert into x select * from remote('127.0.0.1', system, one);
> **0 rows in set. Elapsed... | 2019-12-07T09:19:02 | 2019-12-08T07:36:36 | {} | NONE |
ClickHouse | ClickHouse | 567,740,141 | 8,300 | alexey-milovidov | `unknown warning option '-Wno-ctad-maybe-unsupported'`
Fixed in https://github.com/ClickHouse/ClickHouse/commit/cea2903e72d4378b382f9b8ae616c01f512483dc#diff-eabb2b894359aa83d33168e3117959dcR55 | 2019-12-20T00:45:59 | 2019-12-20T00:45:59 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.