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 | 471,564,979 | 4,023 | karakum | @alexey-milovidov, What about `SELECT formatDateTime(toDate('2018-12-31'), '%Y%V')` -> `201801` ?
It seems that new modificator "year of %V week" required. | 2019-03-11T14:37:18 | 2019-03-11T14:37:18 | {} | NONE |
ClickHouse | ClickHouse | 472,054,771 | 4,663 | pbuxtonccuk | Just shy of 200m rows. What I would like is to be able to do an incremental update. Now, currently, I am pulling this data from MSSQL, and putting it in the ELK stack, using Logstash . I am aware of the ClickhouseMigrator project on Github, but it does't work. I am also aware of the logstash-output-clickhouse pl... | 2019-03-12T15:42:31 | 2019-03-12T15:42:31 | {} | NONE |
ClickHouse | ClickHouse | 472,191,909 | 4,658 | silviucpp | Seems the above condition get broken on the ParquetBlockInputStream.cpp line 262 where creating this global variable
```c++
const std::unordered_map<arrow::Type::type, std::shared_ptr<IDataType>> arrow_type_to_internal_type = {
...
{arrow::Type::DATE64, std::make_shared<DataTypeDateTime>()},
{arrow... | 2019-03-12T21:39:31 | 2019-03-12T21:39:31 | {
"+1": 3
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 472,665,331 | 4,654 | alexey-milovidov | But your test is slightly incorrect:
```
milovidov-Latitude-E7470 :) CREATE VIEW test.t ( next_number UInt64) AS SELECT number + 1 FROM system.numbers
CREATE VIEW test.t
(
next_number UInt64
) AS
SELECT number + 1
FROM system.numbers
Ok.
0 rows in set. Elapsed: 0.011 sec.
milovidov-Latitude-E... | 2019-03-14T01:10:20 | 2019-03-14T01:10:20 | {} | MEMBER |
ClickHouse | ClickHouse | 473,106,934 | 4,689 | alexey-milovidov | The issue is related to the usage of poco-1.9.0-2 (non standard build variant).
You need to run clickhouse with gdb,
type `catch throw` and `run`,
reproduce the issue and type
`bt full` to obtain full stack trace. | 2019-03-14T23:45:33 | 2019-03-14T23:45:52 | {} | MEMBER |
ClickHouse | ClickHouse | 471,912,076 | 1,821 | filimonov | Related: https://github.com/yandex/ClickHouse/issues/520 | 2019-03-12T08:57:23 | 2019-03-12T08:57:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 472,328,363 | 4,047 | mikeeremin | 19.3.7 - still present. | 2019-03-13T08:37:13 | 2019-03-13T08:37:13 | {
"confused": 2
} | NONE |
ClickHouse | ClickHouse | 472,444,931 | 3,611 | den-crane | @amosbird everything works fine without final.
And I execute optimize final at the non-leader host localhost and error reports about the leader XXXX
Received from localhost:9000, ::1. DB::Exception: Received from XXXX DB::Exception: Unknown log entry type: 0.
```
| => cat test_final.sql |clickhouse-client -mn... | 2019-03-13T14:28:33 | 2019-03-13T14:28:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 472,452,351 | 3,611 | den-crane | The same
```
| => cat test.sql
truncate table opt_test;
insert into dw.opt_test select 999 from numbers(20);
insert into dw.opt_test select 999 from numbers(20);
optimize table opt_test final;
| => cat test.sql |clickhouse-client -mn
Received exception from server (version 18.14.18):
Code: 49. DB::Exceptio... | 2019-03-13T14:45:13 | 2019-03-13T14:45:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 472,652,826 | 4,683 | alexey-milovidov | A test? | 2019-03-14T00:02:34 | 2019-03-14T00:02:34 | {} | MEMBER |
ClickHouse | ClickHouse | 472,665,436 | 4,683 | amosbird | updated. Not sure if this fixes the original issue but at least it covers what @den-crane had reproduced. | 2019-03-14T01:10:58 | 2019-03-14T01:10:58 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 473,327,319 | 4,660 | BayoNet | Починил вроде. | 2019-03-15T15:19:53 | 2019-03-15T15:19:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 472,648,410 | 4,671 | alexey-milovidov | PREWHERE is disabled for Buffer engine for reason.
When you query Buffer table, it will read from two places:
- from the destination table;
- from its buffers in memory;
When a query with PREWHERE is executed, filtering by PREWHERE expression must be done entirely inside Storage. In contrast to WHERE, it is not... | 2019-03-13T23:40:43 | 2019-03-13T23:40:43 | {} | MEMBER |
ClickHouse | ClickHouse | 472,996,268 | 1,861 | Farfaday | Hi,
it seems that we are not affected anymore with clickhouse version 19.3.6 :)
Thanks ! | 2019-03-14T18:13:17 | 2019-03-14T18:13:17 | {} | NONE |
ClickHouse | ClickHouse | 473,327,704 | 4,703 | filimonov | > Besides when I login in to xxx.xx.xx.32,I can't connect to xxx.xx.xx.35 using clickhouse-client -host xxx.xx.xx.35 --password xxx
So you have some connectivity issues.
Please check your network/ports/listen configuration (for example ensure that you listen needed network interfaces), also firewalls/iptable etc... | 2019-03-15T15:20:50 | 2019-03-15T15:20:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 471,605,157 | 4,023 | karakum | @alexey-milovidov, yes, but what about
```
$ date -d "2018-12-31 00:00:00" +"%G%V"
201901
```
help:
```
%g последние две цифры года для номера недели ISO (см. %G)
%G год для номера недели ISO (см. %V); обычно бывает полезно только с %V
``` | 2019-03-11T16:04:51 | 2019-03-11T16:04:51 | {} | NONE |
ClickHouse | ClickHouse | 472,174,189 | 4,658 | silviucpp | I found the problem:
When setting the value from config we have:
```c++
/// This must be done before any usage of DateLUT. In particular, before any logging.
if (config().has("timezone"))
{
if (0 != setenv("TZ", config().getString("timezone").data(), 1))
throw Poco::Exception("Can... | 2019-03-12T20:46:06 | 2019-03-12T20:46:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 472,654,152 | 4,674 | alexey-milovidov | A test?
You can use `.sh` tests, see `dbms/tests/queries/0_stateless` directory.
You can pipe test Cap'n'Proto message as blob embedded in `.sh` script. | 2019-03-14T00:09:03 | 2019-03-14T00:09:03 | {} | MEMBER |
ClickHouse | ClickHouse | 472,664,754 | 4,671 | alexey-milovidov | Nevertheless, you can complete this PR in a proper way.
To do it, you should add interpretation of PREWHERE (create ExpressionBlockInputStream FilterBlockInputStream) inside StorageBuffer::read for processing of buffered data.
| 2019-03-14T01:07:14 | 2019-03-14T01:07:14 | {} | MEMBER |
ClickHouse | ClickHouse | 472,774,554 | 4,685 | fancyqlx | @alexey-milovidov
Hi, I found POCO has fixed this bug in their repo, Sun -> Sunday. But clickhouse's extra poco lib has not update newest repo. Can you update extra poco submodule as soon as possible. Thanks.
Instead of `select distinct(MYCOLUMN) from t`
low cardinality types have an optimization to use low cardinality nature for GROUPBY by_single_LowC_column | 2019-03-14T13:04:38 | 2019-03-14T13:05:42 | {
"+1": 5
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 472,862,889 | 3,529 | madm1keru | Are MV made as Replicated*MergeTree engine? If not, it won't work properly, since MV will be applied only for data came from direct inserts, but not for replicated data.
If yes - check error log for errors in MV inner table. I hope It'll be about ZooKeeper, and next step should be ZK tuning according CH docs recommend... | 2019-03-14T13:50:32 | 2019-03-14T13:50:32 | {} | NONE |
ClickHouse | ClickHouse | 473,163,662 | 4,690 | hczhcz | Seems that it only happens on `UInt16` and `UInt32`.
```sql
select arrayReduce('median', [0, 255]); -- returns Float64
select arrayReduce('median', [0, 257]); -- returns UInt16
select arrayReduce('median', [0, -257]); -- returns Float64
select arrayReduce('median', [0, 4294967295]); -- returns Uint32
select arr... | 2019-03-15T05:28:55 | 2019-03-15T05:36:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 473,197,310 | 4,687 | SevaCode | > What does this change mean?
> Why it doesn't work without this change?
Description of **--keyserver** option from `man gpg`
```
--keyserver name
This option is deprecated - please use the --keyserver in ‘dirmngr.conf’ instead.
Use name as your keyserver. This is the se... | 2019-03-15T08:19:04 | 2019-03-15T08:21:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 471,449,348 | 4,533 | ruslangm | @kindred77 did you find another solution to this problem? The only option that I see at the moment is to decompose the entire json into a flat table. But it confuses me that nested structures should be just for such cases (IMHO) | 2019-03-11T08:38:33 | 2019-03-11T08:38:33 | {} | NONE |
ClickHouse | ClickHouse | 472,432,953 | 4,679 | AdmiralUA | >
>
> > P.S. This error is not logged in /var/log/clickhouse-server/clickhouse-server.err.log.
>
> Hmmm. Please ensure that oomkiller didn't kill the ClickHouse. (check `dmesg`)
Yes he did.
> [2488483.597724] Out of memory: Kill process 20686 (clickhouse-serv) score 857 or sacrifice child
> [2488483.597959]... | 2019-03-13T13:59:13 | 2019-03-13T13:59:13 | {} | NONE |
ClickHouse | ClickHouse | 472,646,135 | 4,654 | alexey-milovidov | The implementation looks almost Ok.
The flag `or_replace_view` looks too specific but Ok for now.
Does it work for materialized views? (I don't see a check for it in the test.)
If it is permitted for materialized views, there will be an issue with replicated materialized views (a typical case). When we drop replic... | 2019-03-13T23:30:08 | 2019-03-13T23:32:28 | {} | MEMBER |
ClickHouse | ClickHouse | 471,534,772 | 4,629 | bgranvea | there are 2 failing tests, is it really related to my PR? | 2019-03-11T13:19:22 | 2019-03-11T13:19:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 471,657,859 | 4,023 | alexey-milovidov | @karakum
Yes, we need to add `%g` and `%G` for `formatDateTime`. That's very easy task. Does anyone want to do? | 2019-03-11T18:12:13 | 2019-03-11T18:12:13 | {} | MEMBER |
ClickHouse | ClickHouse | 472,664,301 | 4,671 | alexey-milovidov | Here is a test that will provide incorrect results with your PR:
```
USE test;
DROP TABLE IF EXISTS dest;
DROP TABLE IF EXISTS buffer;
CREATE TABLE dest (s String, payload DEFAULT arrayMap(x -> s, range(1000))) ENGINE = MergeTree ORDER BY tuple();
CREATE TABLE buffer AS dest ENGINE = Buffer(test, dest, 1, 10, 60,... | 2019-03-14T01:04:41 | 2019-03-14T01:04:41 | {} | MEMBER |
ClickHouse | ClickHouse | 472,996,703 | 4,499 | nikvas0 | Comparison for insert
\# insert into ... select * from datasets.hits_v1
no index: 0 rows in set. Elapsed: 20.463 sec. Processed 8.87 million rows, 8.46 GB (433.65 thousand rows/s., 413.45 MB/s.)
3 x ngrambf(3, 512, 2, 0) (URLDomain, SearchPhrase, Title): 0 rows in set. Elapsed: 42.752 sec. Processed 8.87 ... | 2019-03-14T18:14:29 | 2019-03-14T18:14:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 472,647,072 | 4,654 | alexey-milovidov | I'll add it. | 2019-03-13T23:34:21 | 2019-03-13T23:34:21 | {} | MEMBER |
ClickHouse | ClickHouse | 472,651,176 | 4,672 | alexey-milovidov | `allow_shORt_circuit_logic_expressions`
It's too hacky, maybe name it "allow_short_circuit_or"?
(Or it's just accidential search-and-replace?) | 2019-03-13T23:54:20 | 2019-03-13T23:57:33 | {} | MEMBER |
ClickHouse | ClickHouse | 472,668,445 | 4,683 | amosbird | oops, do you mind revert this. I just pushed a refinement... | 2019-03-14T01:27:23 | 2019-03-14T01:27:23 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 473,009,641 | 4,687 | alexey-milovidov | 1. Need to make it consistent with the instruction on https://clickhouse.yandex/
2. Check that it works on Ubuntu Precise (Docker).
3. Explanation required. What does this change mean? Why it doesn't work without this change? Why it was working fine on some systems? | 2019-03-14T18:47:47 | 2019-03-14T18:49:01 | {} | MEMBER |
ClickHouse | ClickHouse | 473,320,675 | 4,703 | xxxxzr | > Please try to connect from xxx.xx.xx.32 to xxx.xx.xx.35 using same credentials with `clickhouse-client`
Besides when I login in to xxx.xx.xx.32,I can't connect to xxx.xx.xx.35 using
`clickhouse-client -host xxx.xx.xx.35 --password xxx` | 2019-03-15T15:02:31 | 2019-03-15T15:02:31 | {} | NONE |
ClickHouse | ClickHouse | 472,482,301 | 4,675 | alesapin | Rest tests are inconsistent with changes. | 2019-03-13T15:50:29 | 2019-03-13T15:50:29 | {} | MEMBER |
ClickHouse | ClickHouse | 472,568,487 | 4,681 | 4ertus2 | related #4218 | 2019-03-13T19:21:13 | 2019-03-13T19:21:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 472,665,123 | 4,654 | alexey-milovidov | Ocassionally, we have no simple way to test for presence of syntax errors. | 2019-03-14T01:09:16 | 2019-03-14T01:09:16 | {} | MEMBER |
ClickHouse | ClickHouse | 472,672,869 | 4,671 | LiuYangkuan | > Your own test has also failed.
Yes, if I execute my test sql wholly in multi-commands mode, test will be failed. I will take your advice and add interpretation of PREWHERE. Thank you for your advice! | 2019-03-14T01:51:28 | 2019-03-14T01:51:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 472,773,828 | 4,686 | zhang2014 | > About your solution: whether it can be considered as the beginning of the new DataType Json? After parsing incoming JSON, it could store it it some more efficient and searchable format than ColumnString? Is it possible to store JSON in a way that the column store nature of Clickhouse remains so that working with it i... | 2019-03-14T09:37:45 | 2019-03-14T09:48:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 473,319,534 | 4,703 | xxxxzr | > Please try to connect from xxx.xx.xx.32 to xxx.xx.xx.35 using same credentials with `clickhouse-client`
It didn't work. I tried to use `clickhouse-client` on xxx.xx.xx.32 and xxx.xx.xx.35 ,when I run
`SELECT count(1) FROM ontime_local` and got count=0 | 2019-03-15T14:59:41 | 2019-03-15T14:59:41 | {} | NONE |
ClickHouse | ClickHouse | 472,349,776 | 3,611 | amosbird | @den-crane can you reproduce it without `final`? The only difference I can see here is `final` will select the only part to merge which might not exist on the leader server. @alexey-milovidov This doesn't look like a transaction issue as we are still creating only one log entry. | 2019-03-13T09:43:11 | 2019-03-13T09:43:20 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 472,447,364 | 3,611 | amosbird | can you try doing twice insert to generate two parts?
```
cat test.sql
truncate table opt_test;
insert into dw.opt_test select 999 from numbers(20);
insert into dw.opt_test select 999 from numbers(20);
optimize table opt_test final;
``` | 2019-03-13T14:34:02 | 2019-03-13T14:34:02 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 472,456,686 | 4,517 | filimonov | > Does anyone know of a workaround for this if bits level of cuts is needed?
That string just contain a binary represenation of ipv6.
You can process that string by parts, but it's a bit tricky because of endianness.
Please try to follow:
```sql
WITH
IPv6StringToNum('2001:0DB8:AC10:FE01:FEED:BABE:CAFE:... | 2019-03-13T14:54:33 | 2019-03-13T15:16:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 472,813,743 | 3,529 | madm1keru | Since MV works as triggers (processing asynchronously after initial query) I think there will be no any difference.
In my case all issues gone after ZK tuning to avoid huge ZK session timeouts. | 2019-03-14T11:32:31 | 2019-03-14T11:32:31 | {} | NONE |
ClickHouse | ClickHouse | 473,230,304 | 4,663 | pbuxtonccuk | Thanks for the pointer. Never used NiFi, so going to be interesting to look at.... if there are any gotchas, i would appreciate any heads up!!
| 2019-03-15T10:12:50 | 2019-03-15T10:12:50 | {} | NONE |
ClickHouse | ClickHouse | 473,254,783 | 4,701 | alexey-milovidov | 00376_shard_group_uniq_array_of_int_array with UBSan - TODO. | 2019-03-15T11:41:22 | 2019-03-15T11:41:22 | {} | MEMBER |
ClickHouse | ClickHouse | 473,314,183 | 4,678 | alexey-milovidov | Thank you! | 2019-03-15T14:46:16 | 2019-03-15T14:46:16 | {} | MEMBER |
ClickHouse | ClickHouse | 473,320,197 | 4,660 | blinkov | сборка сломалась | 2019-03-15T15:01:18 | 2019-03-15T15:01:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 471,694,098 | 4,655 | alexey-milovidov | Build is broken. | 2019-03-11T19:45:55 | 2019-03-11T19:45:55 | {} | MEMBER |
ClickHouse | ClickHouse | 472,339,018 | 3,611 | moxiaofan9 | Having the same issue.
Clickhouse verision 18.14.13 & 18.14.19.
I can reproduce it with 100% probability. | 2019-03-13T09:11:28 | 2019-03-13T09:11:28 | {} | NONE |
ClickHouse | ClickHouse | 472,454,023 | 3,611 | den-crane | Hm. Magic sleep.
```
| => cat test.sql
truncate table opt_test;
insert into dw.opt_test select 999 from numbers(20);
select sleep(1);
insert into dw.opt_test select 999 from numbers(20);
optimize table opt_test final;
```
```
| => cat test.sql |clickhouse-client -mn
0
| => cat test.sql |clickhouse-c... | 2019-03-13T14:48:53 | 2019-03-13T14:48:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 472,751,881 | 4,686 | mfridental | Nice job too. I was afraid introducing a new dependency because I don't possess a good understanding of the build system. I also didn't discuss the new API in the group, because I wasn't sure whether I am able to implement it properly.
About your solution: whether it can be considered as the beginning of the new Da... | 2019-03-14T08:32:13 | 2019-03-14T08:32:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 472,772,024 | 3,529 | goranc | I have same problem with missing data in materialized view table on each node.
Not sure about zookeeper errors, but finally have missing data in replicated MV tables.
CH version 18.14.19. | 2019-03-14T09:32:35 | 2019-03-14T09:32:51 | {} | NONE |
ClickHouse | ClickHouse | 473,015,965 | 4,678 | alexey-milovidov | For this case, you just need to build ClickHouse locally - that should be easy.
There is also `docker/builder` directory - it will build ClickHouse in Docker. | 2019-03-14T19:04:55 | 2019-03-14T19:04:55 | {} | MEMBER |
ClickHouse | ClickHouse | 473,249,232 | 2,797 | filinger | I'm also interested | 2019-03-15T11:19:01 | 2019-03-15T11:19:01 | {} | NONE |
ClickHouse | ClickHouse | 473,343,881 | 4,708 | kvap | cc @bocharov @twalwyn @Dorokhov | 2019-03-15T16:02:16 | 2019-03-15T16:02:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 471,451,616 | 4,533 | kindred77 | @ruslangm yes, and I find the nested structure is not for this situation, it is for nested table(array+tupple), I have changed the code to parse json nestly to solve this problem.
thanks for comment. | 2019-03-11T08:46:18 | 2019-03-11T08:46:18 | {} | NONE |
ClickHouse | ClickHouse | 471,520,727 | 4,618 | abyss7 | Need to test more fixes in this release - before closing PR | 2019-03-11T12:33:41 | 2019-03-11T12:33:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 471,578,744 | 4,023 | filimonov | > @alexey-milovidov, What about `SELECT formatDateTime(toDate('2018-12-31'), '%Y%V')` -> `201801` ?
Other tools will work exactly the same:
```bash
➜ date -d "2018-12-31 12:00:00 CST" +"%Y%V"
201801
```
> It seems that new modificator "year of %V week" required.
Looks like it should be `W`.
```
➜ date ... | 2019-03-11T15:08:42 | 2019-03-11T15:08:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 471,759,917 | 4,631 | simPod | Another observations showed that the time period when the lag is increasing has random length. | 2019-03-11T22:12:18 | 2019-03-11T22:12:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 471,997,125 | 3,605 | alexey-milovidov | @igron99 has changed his name to @ObjatieGroba. He will definitely do this task. | 2019-03-12T13:20:27 | 2019-03-12T13:20:27 | {
"hooray": 2
} | MEMBER |
ClickHouse | ClickHouse | 472,150,217 | 4,665 | alexey-milovidov | A test:
```
milovidov@example-perftest03j:~/ClickHouse/build_gcc8$ sudo -u clickhouse bash -c "ulimit -v 1000000; dbms/programs/clickhouse server -C /etc/clickhouse-server/config.xml"
Include not found: clickhouse_compression
Logging trace to /var/log/clickhouse-server/clickhouse-server.log
Logging errors to /va... | 2019-03-12T19:36:17 | 2019-03-12T19:36:17 | {} | MEMBER |
ClickHouse | ClickHouse | 472,440,313 | 4,679 | filimonov | So your query is to big too be executed in memory. Did you enable on disk group by & order by?
Check `max_bytes_before_external_group_by` & `max_bytes_before_external_order_by`. | 2019-03-13T14:17:37 | 2019-03-13T14:17:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 472,454,993 | 3,611 | den-crane | ```
| => cat test.sql
truncate table opt_test;
insert into dw.opt_test select 999 from numbers(20);
select sleep(1);
insert into dw.opt_test select 666 from numbers(2);
optimize table opt_test final;
```
| => cat test.sql |clickhouse-client -mn
0
| => cat test.sql |clickhouse-client -mn
0 | 2019-03-13T14:50:54 | 2019-03-13T14:50:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 472,456,598 | 3,611 | amosbird | So it seems when the table is initially empty and when the first part gets created and immediately being remerged, the leader get confused of an empty table having a part to merge. | 2019-03-13T14:54:20 | 2019-03-13T14:54:20 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 472,523,580 | 3,611 | amosbird | @den-crane can you try this pr https://github.com/yandex/ClickHouse/pull/4683 | 2019-03-13T17:24:20 | 2019-03-13T17:24:20 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 472,688,637 | 4,672 | sundy-li | Yes, it's just accidental search-and-replace. I have considered other names like `allow_constant_folding_logic_expressions`, `allow_logic_expressions_constant_folding`
| 2019-03-14T03:15:31 | 2019-03-14T03:15:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 472,727,251 | 4,679 | AdmiralUA | Thanks for your answer. | 2019-03-14T06:57:46 | 2019-03-14T06:57:46 | {} | NONE |
ClickHouse | ClickHouse | 472,810,459 | 3,529 | goranc | Anybody tried impact of setting **_insert_distributed_sync = 1_**
on this issue with materialized vies ? | 2019-03-14T11:23:37 | 2019-03-14T11:23:37 | {} | NONE |
ClickHouse | ClickHouse | 473,115,794 | 4,701 | alexey-milovidov | More info:
```
==================
WARNING: ThreadSanitizer: lock-order-inversion (potential deadlock) (pid=134888)
Cycle in lock order graph: M401237648678584448 (0x000000000000) => M812 (0x7b8400000030) => M401237648678584448
Mutex M812 (Context) acquired here while holding mutex M401237648678584448 (ext_di... | 2019-03-15T00:31:09 | 2019-03-15T00:31:09 | {} | MEMBER |
ClickHouse | ClickHouse | 473,232,721 | 4,678 | simPod | I'm able to build on macos but will try docker. Thanks | 2019-03-15T10:20:23 | 2019-03-15T10:20:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 472,317,518 | 4,669 | silviucpp | Building with -DENABLE_PARQUET=0 can be used as workaround for https://github.com/yandex/ClickHouse/issues/4658 in case you don't need parquet. | 2019-03-13T07:57:39 | 2019-03-13T07:57:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 472,666,485 | 4,683 | alexey-milovidov | Ok.
PS. It seems difficult to provide a good test.
(Integration test is possible but not worth.) | 2019-03-14T01:16:37 | 2019-03-14T01:16:37 | {} | MEMBER |
ClickHouse | ClickHouse | 472,763,311 | 4,685 | fancyqlx | I finally found this is a bug in POCO. DateTimeFormat has a wrong word typo which leads to a mismatch of regular expression. | 2019-03-14T09:07:06 | 2019-03-14T09:07:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 472,032,640 | 4,663 | alex-krash | What amount of data (in terms of row count and data volume) you need to transfer? | 2019-03-12T14:49:37 | 2019-03-12T14:49:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 472,394,250 | 4,517 | toddyocum | Does anyone know of a workaround for this if bits level of cuts is needed?
I figured it out for ipv4,using bitAnd, but am not sure for ipv6, which returns an Fixed(16) string for
IPv6StringToNum()
Here is what I do for IPv4
`IPv4NumToString(bitAnd(IPv4StringToNum(Destination_IPv4_Address), reinterpretAsUInt... | 2019-03-13T12:03:08 | 2019-03-13T12:03:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 472,666,067 | 4,678 | alexey-milovidov | There is undefined configuration macro, you should update `librdkafka-cmake`.
(See `ClickHouse build check`) | 2019-03-14T01:14:22 | 2019-03-14T01:14:22 | {} | MEMBER |
ClickHouse | ClickHouse | 472,758,214 | 4,686 | hczhcz | @mfridental Good questions!
This PR is not intended to introduce a JSON datatype. Instead, it extracts structured data from JSON strings.
The scenario we meet in production is handling tables imported from somewhere else that contains JSON strings. Generally, we extracts the data from JSON strings. After that, we... | 2019-03-14T08:52:19 | 2019-03-14T08:52:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 472,785,901 | 4,688 | alex-krash | @hczhcz , how it differs from this?
```
SELECT toTypeName(('hello', [1, 2, 3])) AS name
┌─name────────────────────────┐
│ Tuple(String, Array(UInt8)) │
└─────────────────────────────┘
``` | 2019-03-14T10:10:44 | 2019-03-14T10:10:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 472,822,629 | 3,529 | goranc | I have simple structure very similar to yours, 4 shards, 2 replicas, cluster with 4 + 4 nodes (for replicas)
ReplicatedMergeTree for master table and for underlying table in materialized views.
Materialized tables are just for different indexes (order by) and should have same data as master.
For all those tables hav... | 2019-03-14T11:59:30 | 2019-03-14T11:59:30 | {} | NONE |
ClickHouse | ClickHouse | 473,235,200 | 4,689 | apisarenco | I'm a bit of a noob here when it comes to gdb. I made it attach to the server PID because I couldn't make it `run`, it says "Exited with exit code 1".
But I did get a vague exception stack trace:
```
#0 0x00007ff753c24afc in pthread_cond_wait@@GLIBC_2.3.2 () from /usr/lib/libpthread.so.0
No symbol table info a... | 2019-03-15T10:28:59 | 2019-03-15T10:28:59 | {} | NONE |
ClickHouse | ClickHouse | 471,416,576 | 4,645 | engccq | I have met the same issue. | 2019-03-11T06:13:52 | 2019-03-11T06:13:52 | {} | NONE |
ClickHouse | ClickHouse | 472,084,581 | 4,663 | TH-HA | Hello , i've also used logstash to do that , but it lack of flexbility on the clickhouse export.Also logstash is very heavy and not really user friendly when developping.
Instead to handle both case , i would suggest to use Apache NIFI.
Nifi is able to query on MSSQL transform your data and push them to clickhouse ... | 2019-03-12T16:48:52 | 2019-03-12T16:48:52 | {} | NONE |
ClickHouse | ClickHouse | 472,192,990 | 4,448 | silviucpp | @alexey-milovidov @proller this feature creates a new bug: https://github.com/yandex/ClickHouse/issues/4658 | 2019-03-12T21:42:52 | 2019-03-12T21:42:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 472,429,055 | 4,679 | filimonov | > P.S. This error is not logged in /var/log/clickhouse-server/clickhouse-server.err.log.
Hmmm. Please ensure that oomkiller didn't kill the ClickHouse. (check `dmesg`) | 2019-03-13T13:48:55 | 2019-03-13T13:48:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 472,646,549 | 4,654 | alexey-milovidov | I see that it is already checked: https://github.com/yandex/ClickHouse/pull/4654/files#diff-b7f2297ee277fecc56911861bae15818R442
The only remaining step is to add a negative test... | 2019-03-13T23:31:57 | 2019-03-13T23:31:57 | {} | MEMBER |
ClickHouse | ClickHouse | 472,685,042 | 4,571 | alexey-milovidov | It works in 19.4 (as you mentioned in chat).
I've added your test case to the test suite. | 2019-03-14T02:54:55 | 2019-03-14T02:54:55 | {} | MEMBER |
ClickHouse | ClickHouse | 472,750,188 | 4,047 | maskshell | Me too,
After upgrade from 18.14 to 19.3.5.
Change log mode to trace:
```
2019.03.14 16:21:21.533589 [ 35 ] {} <Debug> report.amg_wms_stock_flow_local (ReplicatedMergeTreeQueue): Pulling 1 entries to queue: log-0003080762 - log-0003080762
2019.03.14 16:21:21.538431 [ 35 ] {} <Debug> report.amg_wms_stock_flow_... | 2019-03-14T08:26:22 | 2019-03-14T09:02:54 | {} | NONE |
ClickHouse | ClickHouse | 472,687,313 | 4,686 | hczhcz | I have seen @mfridental 's pull request #4680 . Nice job!
Here is our solution to the limitation of exist JSON functions. Hopefully it can bring some help in API design, implementation, etc, no matter if it will be merged.
---
Examples:
```sql
select jsonHas('{"a": "hello", "b": [-100, 200.0, 300]}', 'b');... | 2019-03-14T03:07:42 | 2019-03-14T09:43:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 472,815,556 | 4,678 | simPod | @alexey-milovidov is there a way I can launch CI locally / myself somehow? | 2019-03-14T11:37:54 | 2019-03-14T11:37:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 473,235,344 | 4,703 | filimonov | Please try to connect from xxx.xx.xx.32 to xxx.xx.xx.35 using same credentials with `clickhouse-client` | 2019-03-15T10:29:25 | 2019-03-15T10:29:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 473,258,305 | 3,529 | goranc | Thanks for feedback.
Master table is sharded on 4 nodes and replicated on another 4 nodes.
MVs are created using in advance created table using syntax CREATE MATERIALIZED VIEW xx TO yy.
Node underlying tables for MVs are replicated to another counterpart node on replica server.
So this is something standard rep... | 2019-03-15T11:55:47 | 2019-03-15T11:55:47 | {} | NONE |
ClickHouse | ClickHouse | 471,904,389 | 2,987 | filimonov | Nothing is wrong actually. Just misleading error message (ipv6 is disabled, so it it uses only ipv4).
Duplicate of #4406 | 2019-03-12T08:33:15 | 2019-03-12T08:33:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 472,664,541 | 4,671 | alexey-milovidov | Your own test has also failed. | 2019-03-14T01:06:04 | 2019-03-14T01:06:04 | {} | MEMBER |
ClickHouse | ClickHouse | 472,750,735 | 4,654 | bgranvea | I've excluded materialized views because it is like this in other databases and I think it makes sense to consider MV more like a standard table because of the associated storage.
Anyway, thank you for taking care of the changes. | 2019-03-14T08:28:08 | 2019-03-14T08:28:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 472,455,648 | 3,611 | den-crane | ```
cat test.sql
truncate table opt_test;
select sleep(1);
insert into dw.opt_test select 999 from numbers(20);
optimize table opt_test final;
| => cat test.sql |clickhouse-client -mn
0
Received exception from server (version 18.14.18):
Code: 49. DB::Exception: Received from localhost:9000, ::1. DB::Exceptio... | 2019-03-13T14:52:22 | 2019-03-13T14:52:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 472,786,694 | 4,688 | hczhcz | @alex-krash Oops.. I thought there should be one, but I did not found this one. Problem solved. | 2019-03-14T10:12:58 | 2019-03-14T10:12:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 473,106,032 | 4,690 | alexey-milovidov | Looks like interpolation was lost. | 2019-03-14T23:40:57 | 2019-03-14T23:40:57 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.