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 | 544,305,551 | 3,885 | den-crane | . | 2019-10-20T23:52:35 | 2019-10-20T23:52:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,308,664 | 1,866 | stale[bot] | This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
| 2019-10-21T00:25:19 | 2019-10-21T00:25:19 | {} | NONE |
ClickHouse | ClickHouse | 544,377,183 | 6,720 | Aricliang | <Error> executeQuery: Code: 252, e.displayText() = DB::Exception: Too many parts (6009). Merges are processing significantly slower than inserts. (version 19.7.5.27) (from 1.1.1.1:59162) (in query: INSERT INTO track.array_rtb_15 (arr_64, arr_int, arr_str, hashed_jda, click_time, dt) VALUES) | 2019-10-21T07:00:26 | 2019-10-21T07:00:26 | {} | NONE |
ClickHouse | ClickHouse | 544,497,905 | 7,406 | 4ertus2 | What do you mean saying "query works with usual join"? It should not work for all join except CROSS JOIN. ClickHouse support equi-join algorithm that means you need columns from different tables in each ON clause.
There's related discussion on stackoverflow that says PG executes such JOINS as CROSS JOIN and some spe... | 2019-10-21T12:46:04 | 2019-10-21T12:46:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,501,214 | 5,818 | millin | Pls keep | 2019-10-21T12:54:49 | 2019-10-21T12:54:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,504,819 | 7,406 | dpogorelov | > What do you mean saying "query works with usual join"?
ah, sorry, my fail, actually it's being rewritten to:
`SELECT t0.key, t0.name, t1.key, t1.name FROM demo.abc2 as t0, demo.abc2 as t1`
> For now ClickHouse supports rewrite INNER JOIN with inequality into CROSS JOIN. It's not clear if we need support descri... | 2019-10-21T13:03:59 | 2019-10-21T13:04:19 | {} | NONE |
ClickHouse | ClickHouse | 544,746,611 | 7,005 | jpiper | Yes, in the end I figured out there was a `detached` and `format_version.txt` in the distributed data folder, even though for a distributed table these shouldn't be there. Once I removed them everything booted fine, but it took a long time to identify the table at fault, I have to move the metadata, start the server, p... | 2019-10-21T23:14:34 | 2019-10-21T23:14:34 | {
"+1": 5
} | NONE |
ClickHouse | ClickHouse | 544,777,207 | 7,400 | den-crane | >But I am able to see the same table in CH2 why??
able or unable? | 2019-10-22T01:55:53 | 2019-10-22T01:55:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,904,240 | 7,316 | Enmk | @den-crane the #7170 will not add FULL support of dates outside of given range, that was planned for phase 2 of DateTime64 implementation.
Even more to that, now there is no way to lift that limitation completely in simple and efficient manner, since current calendar-related features (getting date/weekday/month) rel... | 2019-10-22T10:46:53 | 2019-10-22T10:47:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,745,796 | 7,005 | SaltTan | The exception happens when CH tries to attach a distributed table, and there is a directory that does not comply with the naming convention for distributed table's directories.
```
CREATE TABLE system.dt AS system.settings
ENGINE = Distributed(test_shard_localhost, system, settings)
Ok.
DETACH TABLE system.dt
... | 2019-10-21T23:11:00 | 2019-10-21T23:11:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,748,759 | 7,396 | nvartolomei | > Even if we had the alter, it's unlikely that we'll ever have an atomic alter that touches several tables, so some predictable way of ignoring the discrepancies will still be required, so maybe ignoring the missing columns makes sence in the long run as well.
Ideally I'd prefer ignoring missing columns in destinati... | 2019-10-21T23:22:59 | 2019-10-22T12:01:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,975,606 | 7,402 | alexey-milovidov | It does not respect `external_table_functions_use_nulls` setting. | 2019-10-22T13:58:04 | 2019-10-22T13:58:04 | {} | MEMBER |
ClickHouse | ClickHouse | 545,166,060 | 7,316 | den-crane | @Enmk OK good to know. Thanks.
>now there is no way to lift that limitation
Lazy cache. Do not see a problem to calculate and put into cache several days with i.e. 8000 year.
```
pg> select to_char(to_date('01-01-0000','DD-MM-YYYY'), 'day');
to_char
-----------
saturday
(1 row)
pg> select to_char(t... | 2019-10-22T21:34:57 | 2019-10-22T21:35:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 545,179,088 | 7,422 | den-crane | @llj098 Try to make reproducible example.
```
create table MS(time UInt64, price Decimal128(12), side UInt16) Engine=Memory;
insert into MS select 1554439015000001000+number, number%1997/100, 1 from numbers(100000);
insert into MS select 1571782299000000000+number, number%1997/100, 1 from numbers(100000);
SE... | 2019-10-22T22:14:42 | 2019-10-22T22:14:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 545,201,093 | 7,417 | minkymorgan | I spoke too soon. Although cmake completed on the stable branch, the next step when doing ninja failed. Details are:
I'm using macOS 10.14.6
```
$ ninja --version
1.9.0
$ cmake --version
cmake version 3.15.4
$ gcc-8 --version
gcc-8 (Homebrew GCC 8.3.0) 8.3.0
$ git branch
* stable
$ cmake .. -DCMAKE_CXX_CO... | 2019-10-22T23:40:34 | 2019-10-22T23:40:34 | {} | NONE |
ClickHouse | ClickHouse | 544,415,242 | 7,382 | xmariachi | Thanks, it works! I see, we just need a "dummy" value here to bypass the limitation, used for not breaking calculation but that will never be actually used. | 2019-10-21T08:49:02 | 2019-10-21T08:49:02 | {} | NONE |
ClickHouse | ClickHouse | 544,883,147 | 7,414 | excitoon | > Why not moved?
#7424 | 2019-10-22T09:41:07 | 2019-10-22T09:41:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,884,302 | 7,414 | excitoon | > Mutation
#7425 | 2019-10-22T09:44:33 | 2019-10-22T09:44:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 545,005,394 | 7,401 | mobydi | I also noticed that after ```OPTIMIZE TABLE test.ht_repl FINAL ```
It doesn't delete parts and I see them in ```SELECT partition FROM system.parts ...```
| 2019-10-22T15:00:26 | 2019-10-22T15:00:26 | {} | NONE |
ClickHouse | ClickHouse | 545,261,972 | 7,439 | alexey-milovidov | PS. I have checked that we have no such typos in other places. | 2019-10-23T04:35:53 | 2019-10-23T04:36:00 | {} | MEMBER |
ClickHouse | ClickHouse | 544,308,632 | 2,271 | stale[bot] | This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
| 2019-10-21T00:25:07 | 2019-10-21T00:25:07 | {} | NONE |
ClickHouse | ClickHouse | 544,308,671 | 1,644 | stale[bot] | This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
| 2019-10-21T00:25:21 | 2019-10-21T00:25:21 | {} | NONE |
ClickHouse | ClickHouse | 544,514,985 | 6,996 | BayoNet | @alexey-milovidov Look at this one more time, please. | 2019-10-21T13:29:40 | 2019-10-21T13:29:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,303,594 | 2,485 | stale[bot] | This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
| 2019-10-20T23:25:20 | 2019-10-20T23:25:20 | {} | NONE |
ClickHouse | ClickHouse | 544,333,110 | 7,399 | den-crane | CH is a columnar database. Such databases have another nature. They don't require indexes for each query. But you should chose the main index (PRIMARY KEY) (TABLE ORDER) wisely before you create a table. When the table is created the index cannot be changed.
https://clickhouse.yandex/docs/en/operations/table_engines... | 2019-10-21T03:18:31 | 2019-10-21T03:18:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,346,287 | 3,950 | vasyaabr | I will up this issue | 2019-10-21T04:42:27 | 2019-10-21T04:42:27 | {} | NONE |
ClickHouse | ClickHouse | 544,595,155 | 7,409 | Tasselmi | > It will not.
> A current session's state of max_memory_usage (and all other settings) will be propagated through distributed from a node initiator.
anyway, you suggest us not to specify password in plaintext, but this config method explode password of default user to all of the employees. This is not a safe appro... | 2019-10-21T16:26:26 | 2019-10-21T16:26:26 | {} | NONE |
ClickHouse | ClickHouse | 544,599,937 | 7,409 | den-crane | I suggest nothing.
You have options:
* plain password in remote_servers for Distributed table
* passwordless auth with IP address restrictions for passwordless user.
https://github.com/ClickHouse/ClickHouse/issues/5823
https://github.com/ClickHouse/ClickHouse/issues/5441
https://github.com/ClickHouse/ClickHo... | 2019-10-21T16:37:01 | 2019-10-21T16:37:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,734,657 | 7,410 | filimonov | `<disks>` and `<policies>` should all go to `<storage_configuration>` subsection.
Here is the working sample: https://github.com/ClickHouse/ClickHouse/blob/4575b6e4ffb35c9d613fd3fe8b6a6cabd2bbaec6/dbms/tests/integration/test_multiple_disks/configs/config.d/storage_configuration.xml | 2019-10-21T22:27:47 | 2019-10-21T22:28:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,949,404 | 7,413 | CurtizJ | Please, can you run queries:
```
SELECT formatReadableSize(sum(data_uncompressed_bytes))
FROM system.parts
WHERE (database = 'brain') AND (table = 'brain_mb')
```
and
```
SELECT count()
FROM system.parts
WHERE (database = 'brain') AND (table = 'brain_mb')
```
| 2019-10-22T12:58:45 | 2019-10-22T12:58:45 | {} | MEMBER |
ClickHouse | ClickHouse | 545,186,112 | 7,263 | den-crane | works
```
echo 'select fullHostName()' |clickhouse-benchmark -c100 -i 1000
Loaded 1 queries.
Queries executed: 1000.
localhost:9000, queries 1000, QPS: 13633.102, RPS: 13633.102, MiB/s: 0.013, result RPS: 13633.102, result MiB/s: 0.442.
``` | 2019-10-22T22:39:18 | 2019-10-22T22:39:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 545,321,277 | 7,443 | vpanfilov | Have you tried `-MergeState` modifier?
```
select d, uniqMergeState(us) FROM mvz group by d order by d
``` | 2019-10-23T08:03:25 | 2019-10-23T08:03:25 | {} | NONE |
ClickHouse | ClickHouse | 545,329,870 | 7,446 | akonyaev90 | opened by mistake | 2019-10-23T08:25:35 | 2019-10-23T08:25:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,345,328 | 1,644 | filimonov | Adaptive index granularity is avaliable since 19.7 and is default since 19.11 | 2019-10-21T04:37:10 | 2019-10-21T04:37:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,439,241 | 7,398 | akuzm | Wrong repository? | 2019-10-21T09:50:40 | 2019-10-21T09:50:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,721,759 | 7,414 | filimonov | Mutation
```
3 [ 13 ] {} <Trace> Aggregator: Aggregated. 0 to 1 rows (from 0.000 MiB) in 0.000 sec. (0.000 rows/sec., 0.000 MiB/sec.)
2019.10.21 21:41:35.339432 [ 13 ] {} <Trace> Aggregator: Merging aggregated data
2019.10.21 21:41:35.339472 [ 13 ] {} <Trace> virtual DB::MergingAndConvertingBlockInputStream::~Mergi... | 2019-10-21T21:46:11 | 2019-10-21T21:46:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,790,936 | 7,417 | GOGOYAO | I build it following the step in https://clickhouse.yandex/docs/en/development/build_osx/ but failed with the error message. Can anybody give me a favor? | 2019-10-22T03:24:03 | 2019-10-22T03:24:45 | {} | NONE |
ClickHouse | ClickHouse | 544,816,150 | 7,392 | yuzhichang | Great job! Tested with product data, found no problem. | 2019-10-22T05:58:05 | 2019-10-22T05:58:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,934,491 | 7,344 | yd1261137568 | Well, thank you | 2019-10-22T12:18:42 | 2019-10-22T12:18:42 | {} | NONE |
ClickHouse | ClickHouse | 545,055,877 | 7,438 | champtar | Actually clickhouse match the Enum8 string to Protobuf enum string, so not sure how to handle that except just using uint32 | 2019-10-22T16:54:39 | 2019-10-22T16:54:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 545,058,074 | 6,243 | akuzm | `Functional stateless tests (debug)` failure is out of disk space in the CI. | 2019-10-22T17:00:11 | 2019-10-22T17:00:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,538,031 | 7,409 | den-crane | It will not.
A current session's state of max_memory_usage (and all other settings) will be propagated through distributed from a node initiator. | 2019-10-21T14:22:43 | 2019-10-21T14:22:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,611,786 | 6,586 | 4ertus2 | #7186 | 2019-10-21T17:05:36 | 2019-10-21T17:05:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,765,759 | 7,392 | amosbird | perf test shows promising results | 2019-10-22T00:46:15 | 2019-10-22T00:46:15 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 545,063,012 | 7,413 | reflection | ```
SELECT formatReadableSize(sum(data_uncompressed_bytes))
FROM system.parts
WHERE (database = 'brain') AND (table = 'brain_mb')
┌─formatReadableSize(sum(data_uncompressed_bytes))─┐
│ 3.52 MiB │
└──────────────────────────────────────────────────┘
1 rows in set. Elaps... | 2019-10-22T17:12:48 | 2019-10-22T17:12:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 545,258,423 | 7,423 | alexey-milovidov | Missing integration test. | 2019-10-23T04:17:43 | 2019-10-23T04:17:43 | {} | MEMBER |
ClickHouse | ClickHouse | 544,499,152 | 5,367 | stavrolia | Fix #5281 | 2019-10-21T12:49:18 | 2019-10-21T12:49:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,537,040 | 7,409 | Tasselmi | > Does not matter. Works as documented. Not a bug.
I think this might be a problem.
If I set user a with profile a(`max_memory_usage=x`), and user b with profile b(`max_memory_usage=y`).
If we use user b to communicate between cluster machines and if we login with user a.
Then user a got the setting of user b , ... | 2019-10-21T14:20:21 | 2019-10-21T14:20:21 | {} | NONE |
ClickHouse | ClickHouse | 544,543,539 | 7,387 | vincenzon | That did it! Thanks.
| 2019-10-21T14:34:47 | 2019-10-21T14:34:47 | {} | NONE |
ClickHouse | ClickHouse | 544,624,555 | 7,397 | marklit | You're right. I'll try out the materialized table approach and re-open this ticket if anything related to the above comes up. | 2019-10-21T17:36:22 | 2019-10-21T17:36:22 | {} | NONE |
ClickHouse | ClickHouse | 545,383,496 | 5,417 | akuzm | This PR was fully merged as a series of patches ending with #6243. | 2019-10-23T10:43:35 | 2019-10-23T10:43:35 | {
"hooray": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,320,432 | 7,391 | Tasselmi | > Shards and replicas are not connected. If you have sharded tables there is no replication between shards--the data are disjoint. You don't need zookeeper in this case unless you plan to add replicas later.
>
> Note that if you don't use replication and a host drops out for some reason, any data on that host will b... | 2019-10-21T02:01:07 | 2019-10-21T02:01:07 | {} | NONE |
ClickHouse | ClickHouse | 545,192,984 | 7,417 | minkymorgan | I had the same issue today and this resolved it (I think):
1st, I deleted the whole ClickHouse folder causing issues, as switching to stable branch required some files to be deleted.
Then I cloned ONLY the stable branch using:
`git clone --recursive -b stable https://github.com/yandex/ClickHouse.git`
Then comp... | 2019-10-22T23:07:13 | 2019-10-22T23:07:13 | {} | NONE |
ClickHouse | ClickHouse | 545,217,615 | 7,363 | ghost | @zhang2014
Hello~
I appreciate your previous code review #6984.
And nice to met you again.
I fixed codes to pass tests, expecting your review. :) | 2019-10-23T01:02:02 | 2019-10-23T01:02:02 | {} | NONE |
ClickHouse | ClickHouse | 545,259,783 | 7,412 | alexey-milovidov | `No space left on device.` - infrastructure issue, @alesapin | 2019-10-23T04:24:40 | 2019-10-23T04:24:40 | {} | MEMBER |
ClickHouse | ClickHouse | 545,352,637 | 7,316 | Enmk | Yep, but current lookup table is pre-computed once the timezone is created, that means if we are to replace LUT with lazy cache, we'll get a performance degradation on first access on any particular time value.
Sure, we can have both precomputed LUT and lazy-cache, but this makes solution not quite simple and slight... | 2019-10-23T09:21:00 | 2019-10-23T09:21:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,304,997 | 6,287 | den-crane | still relevant. Still wrong query result. | 2019-10-20T23:44:37 | 2019-10-20T23:44:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,336,733 | 7,380 | hczhcz | > Maybe it's better to negate condition?
Sure.
> If you consider some alternative names to this function, what comes to mind?
We may consider `arrayPopulate` as a synonym if the word "fill" sounds too general.
`arrayBackFill` is more popular than `arrayReverseFill` among data analysts. But I would choose `a... | 2019-10-21T03:40:24 | 2019-10-21T03:40:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,425,227 | 7,397 | marklit | Cheers for that @zombiemonkey
@josepowera I've tried 19.14.7.15, 19.14.3.3, 19.13.6.51, 19.11.12.69 and none of them will read from Kafka. I'm not sure if I'll be able to prove the regression is behind all of this without a old version of CH I can get to work. I'll paste my steps below to see if I'm missing somethi... | 2019-10-21T09:14:57 | 2019-10-21T09:15:39 | {} | NONE |
ClickHouse | ClickHouse | 544,518,273 | 7,409 | den-crane | >where user default come from?
Distributed table uses the user/password you described with cluster definition
https://clickhouse.yandex/docs/en/operations/table_engines/distributed/
```
<remote_servers>
<test>
<shard>
<replica>
<host>host</host>
... | 2019-10-21T13:37:13 | 2019-10-21T13:37:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,550,990 | 6,243 | akuzm | The ubsan failure is in `preciseExp10`, not related to this PR. | 2019-10-21T14:50:50 | 2019-10-21T14:50:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,604,381 | 7,409 | Tasselmi | > I suggest nothing.
>
> You have options:
>
> * plain password in remote_servers for Distributed table
> * passwordless auth with IP address restrictions for passwordless user.
>
> #5823
> #5441
> #2290
ok, thanks. I will choose the second one. | 2019-10-21T16:47:33 | 2019-10-21T16:47:33 | {} | NONE |
ClickHouse | ClickHouse | 544,691,474 | 6,822 | stale[bot] | This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
| 2019-10-21T20:23:15 | 2019-10-21T20:23:15 | {} | NONE |
ClickHouse | ClickHouse | 544,892,125 | 6,814 | zhang2014 | Continue in #7426 .
| 2019-10-22T10:07:56 | 2019-10-22T10:07:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 545,001,143 | 7,427 | den-crane | No. CH does not support changing of table engines.
You should create a new table and copy data with `insert into new select from old` | 2019-10-22T14:51:31 | 2019-10-22T14:51:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 545,189,997 | 7,442 | den-crane | dup: https://github.com/ClickHouse/ClickHouse/issues/4165 | 2019-10-22T22:55:10 | 2019-10-22T22:55:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 545,201,724 | 4,250 | den-crane | @Rahuljais098 MV traces only inserts into left table (Test.Employee in your case)
```
insert into Test.User values(1, 'xyz', '55-5');
insert into Test.MV_Emp_detailss select number%2, toString(number%2), 10 from numbers(10);
SELECT *
FROM Test.MV_Emp_detailss
┌─Emp_id─┬──s─┬─Emp_name─┬─Emp_address─┐
│ ... | 2019-10-22T23:43:23 | 2019-10-22T23:44:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 545,231,030 | 7,422 | llj098 | fixed with sql:
```
SELECT
'okex-quarter' AS ex,
'btc_usdt' AS sym,
intDiv(time, 100000000000) AS time10,
price,
sum(amount),
side
FROM trade
WHERE (exchange = 'okex-quarter') AND (symbol = 'btc_usdt')
GROUP BY (time10, price, side)
ORDER BY time10 ASC
LIMIT 10
``` | 2019-10-23T02:07:31 | 2019-10-23T02:07:42 | {} | NONE |
ClickHouse | ClickHouse | 544,315,091 | 7,378 | introspection3 | I'm so sorry,my question is Does ClickHouse need us to create the **index** manually
| 2019-10-21T01:24:11 | 2019-10-21T01:24:11 | {} | NONE |
ClickHouse | ClickHouse | 544,330,939 | 7,399 | den-crane | MergeTree table can have only one index (primary key). Basically you cannot create a table without index. So when you create a MergeTree table you have to mention ORDER BY (primary key) section. This section sets rows order (all columns is sorted in that order). The same time this creates sparsed index.
https://cli... | 2019-10-21T03:04:35 | 2019-10-21T03:06:57 | {
"+1": 3
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,331,855 | 7,399 | introspection3 | Thansk sir,I mean:
does ClickHouse need us create all index manually just like every mysql query (if we didn't create index for mysql query ,the speed will be every slow....) | 2019-10-21T03:10:47 | 2019-10-21T03:14:11 | {} | NONE |
ClickHouse | ClickHouse | 544,564,938 | 7,005 | jpiper | I'm also seeing this in `19.15.2.2`, and am having a hard time working out what the issue is. | 2019-10-21T15:20:38 | 2019-10-21T15:20:38 | {} | NONE |
ClickHouse | ClickHouse | 544,606,462 | 7,397 | marklit | I made the above adjustments to the Kafka table in my CH 19.11.12.69 setup.
```sql
CREATE TABLE trips_kafka_csv (
...
) ENGINE = Kafka SETTINGS kafka_broker_list = 'localhost:9092',
kafka_topic_list = 'trips5',
kafka_group_name = 'group1',
... | 2019-10-21T16:52:52 | 2019-10-21T16:52:52 | {} | NONE |
ClickHouse | ClickHouse | 544,305,456 | 6,417 | den-crane | still relevant
SELECT *
FROM ES
LIMIT 10
FORMAT Null
↘ Progress: 0.00 rows, 0.00 B (0.00 rows/s., 0.00 B/s.) Received exception from server (version 19.16.1):
Code: 160. DB::Exception: Received from localhost:9000. DB::Exception: Estimated query execution time (3999.9725568 seconds) is too long. Maximum: 1. E... | 2019-10-20T23:51:13 | 2019-10-20T23:51:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,438,704 | 7,328 | akuzm | > @akuzm @alexey-milovidov what about `arrayCompress` ?
Let's leave it `arrayCompact`, it's OK.
The stress test failure is not related to your PR, so you don't have to do anything about it. | 2019-10-21T09:49:12 | 2019-10-21T09:49:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,507,082 | 6,243 | akuzm | ```
SELECT MobilePhoneModel, uniq(UserID) AS u FROM hits_100m_single WHERE MobilePhoneModel != '' GROUP BY MobilePhoneModel ORDER BY u DESC LIMIT 10
┏━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━┳━━━━━━━┳━━━━━━━┳━━━━━━━┳━━━━━━━━━┓
┃ version ┃ count() ┃ min ┃ med ┃ max ┃ q95-q05 ┃
┡━━━━━━━━━━━━━━━━━━━━━━━╇━━... | 2019-10-21T13:10:03 | 2019-10-21T13:10:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,849,128 | 7,420 | amosbird | > I hereby not agree
lol | 2019-10-22T08:03:12 | 2019-10-22T08:03:12 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 545,260,997 | 7,396 | alexey-milovidov | > Do we want to also implement second behaviour?
Yes.
> Do we want to have a SETTING section on MV where this behaviour could be opt-in/opt-out?
No.
> Do we want to remove current behaviour and generate errors when people try to write more columns than target table has?
No. | 2019-10-23T04:30:42 | 2019-10-23T04:30:42 | {} | MEMBER |
ClickHouse | ClickHouse | 544,308,623 | 2,245 | stale[bot] | This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
| 2019-10-21T00:25:02 | 2019-10-21T00:25:02 | {} | NONE |
ClickHouse | ClickHouse | 544,320,499 | 7,391 | Tasselmi | > > hi，I want to make it clear:
> > if I use 11 shards each with 1 replica, to use `on cluster` I have to deploy a standalone zookeeper and set `internal_replication = true`, right? ` internal_replication = true` is necessary？or I only need a standalone zookeeper and set `internal_replication = false` also be fine?
>... | 2019-10-21T02:01:45 | 2019-10-21T02:01:45 | {} | NONE |
ClickHouse | ClickHouse | 544,395,932 | 6,081 | abraithwaite | Still watching this. :-) | 2019-10-21T07:58:15 | 2019-10-21T07:58:15 | {} | NONE |
ClickHouse | ClickHouse | 544,449,123 | 7,397 | josepowera | Just a few remarks that I have seen at first sight
Sometimes last year there was a change that required to add kafka_row_delimiter = '\n' to kafka create table (for "Exception: expected end of line".
```
CREATE TABLE trips_kafka_csv5 (
...
)
ENGINE = Kafka
SETTINGS kafka_broker_list = 'xxx.com', kafka_... | 2019-10-21T10:16:56 | 2019-10-21T10:19:30 | {} | NONE |
ClickHouse | ClickHouse | 544,481,236 | 7,404 | amosbird | https://github.com/yandex/ClickHouse/blob/master/dbms/src/Functions/FunctionsLogical.h#L87
And perhaps you want `bitXor`? | 2019-10-21T11:57:48 | 2019-10-21T11:57:48 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 544,527,519 | 7,409 | den-crane | Does not matter. Works as documented. Not a bug. | 2019-10-21T13:58:40 | 2019-10-21T13:58:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,987,969 | 6,562 | MetikovVadim | I have reduced rollup rules from 255 to 15 and get only 1-3 parralel merge process. Also amount of active parts reduced from 100+ to 30-40.
Im closing the ticket. | 2019-10-22T14:23:24 | 2019-10-22T14:23:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 545,220,008 | 7,422 | llj098 | I will try... I imported 3 dateset, the issue only occurs after 3rd import. | 2019-10-23T01:13:46 | 2019-10-23T01:13:46 | {} | NONE |
ClickHouse | ClickHouse | 545,270,782 | 7,400 | anilxoxo | Unable. | 2019-10-23T05:17:34 | 2019-10-23T05:17:34 | {} | NONE |
ClickHouse | ClickHouse | 545,390,906 | 6,610 | CurtizJ | You can set `STEP` for `WITH FILL` modifier:
```... WITH FILL FROM toDate('2019-06-05') TO toDate('2019-06-16') STEP 3``` .
Now you have to convert your interval to seconds or days manually. It's possible to implement support of [interval syntax](https://clickhouse.yandex/docs/en/query_language/functions/type_conver... | 2019-10-23T11:05:44 | 2019-10-23T11:05:44 | {} | MEMBER |
ClickHouse | ClickHouse | 544,620,225 | 7,397 | den-crane | You mean you did not use materialized view?
>SELECT * FROM trips_kafka_csv;
It's a special debug mechanism it's not meant to use to ingest data. It uses max_block_size=1 setting.
Check examples with materialized view at https://clickhouse.yandex/docs/en/operations/table_engines/kafka/ | 2019-10-21T17:26:16 | 2019-10-21T17:26:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,719,360 | 7,414 | filimonov | ```
alter table test move partition tuple() TO volume 'default';
ALTER TABLE test
MOVE PARTITION tuple() TO VOLUME 'default'
Ok.
0 rows in set. Elapsed: 148.339 sec.
ip-172-31-47-236.eu-central-1.compute.internal :) select database, table, disk_name,formatReadableSize(sum(bytes_on_disk)) from system... | 2019-10-21T21:38:51 | 2019-10-21T21:38:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,791,220 | 7,290 | GOGOYAO | I got this problem too. | 2019-10-22T03:25:49 | 2019-10-22T03:25:49 | {} | NONE |
ClickHouse | ClickHouse | 544,931,676 | 6,243 | akuzm | With the latest dispatch improvement (`switch((sz - 1) >> 3`) there is even less regression on string_hash_map test:
```
┏━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━┳━━━━━━━┳━━━━━━━┳━━━━━━━━┳━━━━━━━━━┳━━━━━━━━━━━━┓
┃ file ┃ version ┃ count() ┃ min ┃ med ┃ max ┃ q95-q05 ┃ mem ┃
┡━━━━━━━━╇━... | 2019-10-22T12:11:02 | 2019-10-22T12:11:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 542,822,589 | 7,360 | alesapin | TODO:
- [x] `Engine = Dictionary()` table for each dictionary inside database.
- [x] system.dictionaries
- [x] More tests (also integeration for MySQL)
- [x] Bug fixes | 2019-10-16T18:03:45 | 2019-10-22T13:00:53 | {} | MEMBER |
ClickHouse | ClickHouse | 545,259,429 | 7,380 | alexey-milovidov | `arrayFill` name is Ok.
Now we have to review the code... | 2019-10-23T04:22:38 | 2019-10-23T04:22:38 | {} | MEMBER |
ClickHouse | ClickHouse | 544,333,388 | 7,399 | introspection3 | thank you sir. | 2019-10-21T03:20:16 | 2019-10-21T03:20:16 | {} | NONE |
ClickHouse | ClickHouse | 544,522,875 | 7,409 | Tasselmi | > > where user default come from?
>
> Distributed table uses the user/password you described with cluster definition
>
> https://clickhouse.yandex/docs/en/operations/table_engines/distributed/
>
> ```
> <remote_servers>
> <test>
> <shard>
> <replica>
> ... | 2019-10-21T13:47:56 | 2019-10-21T13:47:56 | {} | NONE |
ClickHouse | ClickHouse | 544,597,929 | 7,314 | 4ertus2 | The problem is here: https://github.com/ClickHouse/ClickHouse/blob/master/dbms/src/Interpreters/CrossToInnerJoinVisitor.cpp#L124
It's not easy to fix. It's hack to push IN clause down to subquery. And it looks like it depends on tables order. Related to #4731 | 2019-10-21T16:32:27 | 2019-10-21T16:32:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 544,634,809 | 7,412 | 4ertus2 | closes #5034 | 2019-10-21T18:01:32 | 2019-10-21T18:01:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 545,066,693 | 7,392 | 4ertus2 | I've updated PR description. We use template strings in scripts. PR should have "Category" and "Short description" lines present and unchanged. | 2019-10-22T17:22:02 | 2019-10-22T17:24:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 545,488,356 | 6,661 | abyss7 | Should be fixed in v19.17+ | 2019-10-23T15:04:30 | 2019-10-23T15:04:30 | {} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.