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 | 568,958,750 | 8,405 | alexey-milovidov | Your test is slightly incorrect.
1. default is lz4 or the value specified in config.xml in `compression` section.
2. cityHash64 of sequential numbers won't compress at all (neither by lz4 or zstd).
3. The compression ratio depends on whether your data parts were merged. You have four identical data parts, each... | 2019-12-26T03:25:22 | 2019-12-26T03:26:13 | {} | MEMBER |
ClickHouse | ClickHouse | 568,521,989 | 8,347 | den-crane | ```
# alter MV with implicit storage .inner (without TO)
1. stop ingestion
2. detach table MVX
3. alter table `.inner.MVX` add column .....
4. attach MATERIALIZED VIEW MVX as select new_SELECT_.....
5. start ingestion
# how to convert MV implicit storage .inner (without TO) to explicit storage (with TO)
... | 2019-12-23T16:41:01 | 2019-12-23T16:41:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 568,778,053 | 8,374 | alexey-milovidov | Please check if this helps: https://github.com/ClickHouse/ClickHouse/pull/8380 | 2019-12-24T16:53:48 | 2019-12-24T16:53:48 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 568,857,462 | 8,328 | qoega | If you will reproduce this bug in 19.15+ reopen the issue please. | 2019-12-25T07:34:14 | 2019-12-25T07:34:14 | {} | MEMBER |
ClickHouse | ClickHouse | 568,892,837 | 8,375 | KochetovNicolai | Looks fine, but some tests are failed. | 2019-12-25T12:14:48 | 2019-12-25T12:14:48 | {} | MEMBER |
ClickHouse | ClickHouse | 568,916,498 | 8,290 | CurtizJ | Nothing is broken with wide parts (current format).
`test_merge_doesnt_work_without_zookeeper` - flap | 2019-12-25T17:05:09 | 2019-12-25T17:05:09 | {} | MEMBER |
ClickHouse | ClickHouse | 568,944,164 | 8,242 | ghost | @alexey-milovidov
Hello, Did you have a merry Christmas?~
I passed all tests and now waiting for your review.
Please let me know If there is any other problem. | 2019-12-26T01:20:38 | 2019-12-26T01:20:38 | {} | NONE |
ClickHouse | ClickHouse | 568,948,167 | 8,401 | alexey-milovidov | Demo:
```
SELECT arrayStringConcat(arrayMap(x -> char(intDiv((rand(x) % 255) * 95, 256) + 32), range(100)))
┌─arrayStringConcat(arrayMap(lambda(tuple(x), char(plus(intDiv(multiply(modulo(rand(x), 255), 95), 256), 32))), range(100)))─┐
│ +P0\?rgPZ:iMVo7xHGQ>v-_>-2?6R6PEfe._VT<9<d$4Cv#FE9!:;;JSVL4r6!r "brtUb\"8FO... | 2019-12-26T01:59:29 | 2019-12-26T01:59:29 | {} | MEMBER |
ClickHouse | ClickHouse | 568,948,525 | 8,366 | ghost | @den-crane
FYI.
I shared my test result.
When it comes to CPU : **The slower the disk, the better the compression.**
To Disk : **Significant disk space efficiency.**
But, the `default` is the best in my case.
| compression method | disk | rows/s |
| --------------------- | ----- | ------- |
| default | 766... | 2019-12-26T02:02:36 | 2019-12-26T02:02:36 | {} | NONE |
ClickHouse | ClickHouse | 568,518,210 | 8,354 | blitzarx1 | It makes it clear 😊 | 2019-12-23T16:25:11 | 2019-12-23T16:25:11 | {} | NONE |
ClickHouse | ClickHouse | 568,578,858 | 8,333 | den-crane | @amosbird
one more test
```
select 1
where 0=(select number from numbers(2) order by number limit 1)
and
1=(select number from numbers(2) order by number desc limit 1)
19.19.1.1929
0 rows in set. Elapsed: 0.001 sec.
19.13.6.51
┌─1─┐
│ 1 │
└───┘
``` | 2019-12-23T20:52:39 | 2019-12-23T20:53:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 568,586,905 | 7,799 | den-crane | > doesn't reproduce on `19.19.1.1956` testing release installed from official Yandex RPM repo
```
ClickHouse client version 19.19.1.1957 (official build).
Connecting to localhost:9000 as user default.
Connected to ClickHouse server version 19.19.1 revision 54430.
SELECT count()
FROM system.query_log
ANY LEFT... | 2019-12-23T21:33:22 | 2019-12-23T21:33:22 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 568,149,969 | 7,813 | listenerri | > it seems that you can't create a view upon a MATERIALIZED view.
> I try to create MATERIALIZED view upon a MATERIALIZED view, but the upper one never updates.
@Tasselmi What you have encountered should be #7195 | 2019-12-21T04:00:47 | 2019-12-24T03:09:28 | {} | NONE |
ClickHouse | ClickHouse | 568,685,666 | 8,275 | vitlibar | There is no need to backport it to 19.16 because 19.16 doesn't have dictionary DDL support. | 2019-12-24T07:55:54 | 2019-12-24T07:55:54 | {} | MEMBER |
ClickHouse | ClickHouse | 568,692,668 | 7,477 | Vinatorul | @alexey-milovidov Hi! Everything seems fine, I'll try to build it today and if everything will be ok - open a PR. Full build takes about 7 hours on my pc now :( | 2019-12-24T08:29:59 | 2019-12-24T08:29:59 | {} | NONE |
ClickHouse | ClickHouse | 568,710,810 | 8,345 | azat | FWIW: AFAICS the failures are not related to this patchset | 2019-12-24T09:53:21 | 2019-12-24T09:53:21 | {} | MEMBER |
ClickHouse | ClickHouse | 568,712,549 | 8,287 | alex-zaitsev | All failures from master:
```
$ ninja -j 2 -k 20 | grep error
../contrib/libcxx/src/filesystem/filesystem_common.h:413:9: error: no member named 'utimensat' in the global namespace
1 error generated.
../contrib/libcxx/src/filesystem/filesystem_common.h:413:9: error: no member named 'utimensat' in the global namesp... | 2019-12-24T10:01:24 | 2019-12-24T10:01:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 568,794,994 | 8,369 | alexey-milovidov | We cannot just skip NaNs for all aggregate functions. NaNs make sense for `any`, `argMax`, `groupArray`, `uniq`.
Skipping NaNs just for some aggregate functions looks like unnesessary complexity.
Note that SQL standard doesn't have NaNs (and any other demormals) at all.
MySQL returns NULL for `0 / 0` and PostgreS... | 2019-12-24T19:35:35 | 2019-12-24T19:35:45 | {} | MEMBER |
ClickHouse | ClickHouse | 569,031,781 | 8,408 | alexey-milovidov | https://github.com/ClickHouse/ClickHouse/issues/1562 | 2019-12-26T10:19:11 | 2019-12-26T10:19:11 | {} | MEMBER |
ClickHouse | ClickHouse | 568,456,594 | 7,781 | 4ertus2 | Related #8169 | 2019-12-23T12:01:45 | 2019-12-23T12:01:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 568,517,969 | 8,354 | alexey-milovidov | CityHash has changed the algorithm after we have added it into ClickHouse.
CityHash documentation specifically notes that the user should not rely to specific hash values and should not save it anywhere or use it as sharding key.
But as we exposed this function to the user, we had to fix the version of CityHash (... | 2019-12-23T16:24:11 | 2019-12-23T16:24:11 | {
"+1": 3,
"hooray": 1
} | MEMBER |
ClickHouse | ClickHouse | 568,644,844 | 8,366 | ghost | @den-crane
Yes, I mean `is uncompressed` as `without LZ4 or ZSTD`.
Thank you for your advice.
And can I get some more detailed benchmark about total data read speed(decompression + disk read) comparison with `LZ4` and `Uncompressed form`. | 2019-12-24T03:58:42 | 2019-12-24T04:22:09 | {} | NONE |
ClickHouse | ClickHouse | 568,683,657 | 8,082 | vitlibar | Conflict with 19.16. | 2019-12-24T07:45:37 | 2019-12-24T07:45:37 | {} | MEMBER |
ClickHouse | ClickHouse | 568,847,740 | 8,373 | yehiaelbehery | > It returns default value for value type -- String: '', Int* : 0, Date: Date(0), Float: 0.0
This is what I was missing. Awesome, Thank you very much. | 2019-12-25T06:22:09 | 2019-12-25T06:22:30 | {} | NONE |
ClickHouse | ClickHouse | 568,908,096 | 8,396 | den-crane | entity_agg_count is repeated -- rows are deduplicated
https://github.com/ClickHouse/ClickHouse/issues/5274
Check logs at replicas for messages
```
<Information> ....... (Replicated OutputStream): Block with ID
.... already exists; ignoring it.
``` | 2019-12-25T15:10:32 | 2019-12-25T15:10:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 568,983,423 | 8,373 | yehiaelbehery | +1 | 2019-12-26T06:18:20 | 2019-12-26T06:18:20 | {} | NONE |
ClickHouse | ClickHouse | 568,496,306 | 8,347 | byx313 | thx for the reply!
About the Q2,I can not get further information about `TO` syntax in official document,so I guess you create `db.table_view` and drop it is aim to cut the link between `db.table` and `db.origin-table` ? After that you create the table again is aim to create the link again? | 2019-12-23T14:55:00 | 2019-12-23T15:02:09 | {} | NONE |
ClickHouse | ClickHouse | 568,704,365 | 7,477 | alexey-milovidov | Ok. BTW, you can send PR in advance and it will be built in our CI - we can fix all issues quickly. | 2019-12-24T09:23:24 | 2019-12-24T09:23:24 | {} | MEMBER |
ClickHouse | ClickHouse | 568,834,434 | 8,315 | yingfeng | We've made some experiments using Ceph as a Block device. The results show that ClickHouse instance running on Block Devices is only a bit slower (less than 10%) than that running on local disk for cold start queries, and the results are even more close for warm start queries because data in Ceph could also be cached w... | 2019-12-25T04:13:31 | 2019-12-25T04:13:31 | {
"+1": 7
} | NONE |
ClickHouse | ClickHouse | 568,890,334 | 8,334 | maxulan | `flatbuffers` added to contrib. | 2019-12-25T11:45:44 | 2019-12-25T16:40:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 568,914,949 | 8,361 | alesapin | test_parts_delete_zookeeper/test.py::test_merge_doesnt_work_without_zookeeper -- flap. | 2019-12-25T16:42:31 | 2019-12-25T16:42:38 | {} | MEMBER |
ClickHouse | ClickHouse | 568,929,864 | 8,397 | alexey-milovidov | The branch with TurboBase64 can be found here: https://github.com/ClickHouse/ClickHouse/compare/base64-update | 2019-12-25T20:22:56 | 2019-12-25T20:22:56 | {} | MEMBER |
ClickHouse | ClickHouse | 569,049,496 | 8,369 | vpanfilov | @ghuname You can also use `-If` aggregate function combinator (https://clickhouse.yandex/docs/en/query_language/agg_functions/combinators/) to ignore NaN values.
Here is a simple example:
```sql
WITH arrayJoin([1, 2, 3, nan, 4, 5]) AS data
SELECT
avg(data),
avgIf(data, isFinite(data))
┌─avg(data)─┬─a... | 2019-12-26T12:21:06 | 2019-12-26T12:22:16 | {
"+1": 8
} | NONE |
ClickHouse | ClickHouse | 568,455,257 | 8,300 | qoega | Should be fixed in #8297 | 2019-12-23T11:55:47 | 2019-12-23T11:55:47 | {} | MEMBER |
ClickHouse | ClickHouse | 568,570,242 | 8,297 | kolen | This PR fixed problem with building openssl on my machine (mac os 10.13.6, clang 10.0.0 (clang-1000.10.44.4)), but similar problem still persists with ICU:
```
[1681/4437] Building ASM object contrib/icu-cmake/CMakeFiles/icudata.dir/__/icudata/icudt66l_dat.S.o
FAILED: contrib/icu-cmake/CMakeFiles/icudata.dir/__/ic... | 2019-12-23T20:09:29 | 2019-12-23T20:09:29 | {} | NONE |
ClickHouse | ClickHouse | 568,607,965 | 8,344 | alexey-milovidov | It looks like RT-signals are not delivered inside Sandbox (Yandex internal job running system that is used in our CI). | 2019-12-23T23:47:31 | 2019-12-23T23:47:31 | {} | MEMBER |
ClickHouse | ClickHouse | 568,683,703 | 8,313 | svladykin | @filimonov
You mean Kafka engine in Clickhouse? We do not use it, since it looks pretty buggy and unstable right now.
We have an in-house solution for storing data from Kafka to Clickhouse, I'm not saying it is better, but at least we can quickly fix bugs in it ourselves when needed.
Yes, I was thinking about Kaf... | 2019-12-24T07:45:55 | 2019-12-24T07:45:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 568,791,142 | 8,385 | 4ertus2 | #8169 #7781 | 2019-12-24T18:54:11 | 2019-12-24T18:54:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 568,817,321 | 8,366 | den-crane | By default data is compressed by LZ4
https://clickhouse.yandex/docs/en/operations/server_settings/settings/#compression
https://www.altinity.com/blog/2017/11/21/compression-in-clickhouse
Let's test with RANDOM numbers.
# Nothing to compress.
```
create table compressionTest
(A Int64, col_lz4 Int64 Codec(LZ4... | 2019-12-25T01:19:49 | 2019-12-25T04:07:15 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 568,909,002 | 8,397 | alexey-milovidov | We usually encode and decode huge amount of short strings (instead of one large string), so the relevance of benchmarks of these libraries are in question. | 2019-12-25T15:24:40 | 2019-12-25T15:24:40 | {} | MEMBER |
ClickHouse | ClickHouse | 569,023,678 | 8,326 | kononencheg | I’ll fix it on weekend! | 2019-12-26T09:33:09 | 2019-12-26T09:33:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,029,353 | 8,382 | alesapin | It's slightly sad, that now we have to pass `Context` around, even more, when we need just a fixed global `path`. But it's OK for now, we will fix it when `Context` will be split into different entities. | 2019-12-26T10:05:32 | 2019-12-26T10:05:32 | {} | MEMBER |
ClickHouse | ClickHouse | 568,536,806 | 8,362 | azat | FWIW with `experimental_use_processors` (v19.19.1.1929-testing-163-g8711f98c30) I got `Pipeline stuck` error from time to time, I will try to come back with minimal reproducer (or maybe this is not an issues, and caused by for example network errors or similar) | 2019-12-23T17:41:36 | 2019-12-23T17:41:58 | {} | MEMBER |
ClickHouse | ClickHouse | 568,564,391 | 7,477 | alexey-milovidov | @Vinatorul Hi! How is it going? | 2019-12-23T19:41:55 | 2019-12-23T19:41:55 | {} | MEMBER |
ClickHouse | ClickHouse | 568,573,631 | 8,270 | vitlibar | There is no need to backport it to 19.16, 19.17, 19.18 because [19.18](https://github.com/ClickHouse/ClickHouse/tree/19.18/dbms/src/DataTypes) doesn't have the data type `DateTime64`. | 2019-12-23T20:25:57 | 2019-12-23T20:25:57 | {} | MEMBER |
ClickHouse | ClickHouse | 568,514,363 | 8,159 | weiqxu | @gubinjie
sorry, maybe I forget something. you need keep the configuration file name as 'hdfs-site.xml'. the LIBHDFS3_CONF should be set to /etc/clickhouse-server for you .
If you still get error, maybe you need give the error info. | 2019-12-23T16:09:36 | 2019-12-23T16:09:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 568,858,109 | 8,366 | qoega | @achimbab, do you have any more question left about compressed/uncompressed scenario? If everything is clear close the issue, please. | 2019-12-25T07:38:35 | 2019-12-25T07:38:35 | {} | MEMBER |
ClickHouse | ClickHouse | 568,965,175 | 8,405 | ghost | @alexey-milovidov
Cool!
Thank you a lot. :) | 2019-12-26T04:21:20 | 2019-12-26T04:27:27 | {} | NONE |
ClickHouse | ClickHouse | 568,429,952 | 8,159 | gubinjie | @weiqxu
1. The clickhouse host is a separate server and is not on any host in the hadoop cluster.
2. I copied hdfs-site.xml to the / etc / clickhouse-server directory and renamed the file hdfs-client.xml.
3.I configured LIBHDFS3_CONF = /etc/clickhouse-server/hdfs-client.xml in the / etc / profile file of the host o... | 2019-12-23T10:12:49 | 2019-12-23T10:12:49 | {} | NONE |
ClickHouse | ClickHouse | 568,441,899 | 8,339 | ishirav | The last upgrade was from 19.8.3.8 to 19.16.2.2
The server logs were already rotated since the upgrade.
An example of a corrupted directory:
```
-rw-r----- 1 clickhouse clickhouse 1396 Dec 23 09:00 checksums.txt
-rw-r----- 1 clickhouse clickhouse 424 Dec 23 09:00 columns.txt
-rw-r----- 1 clickhouse clickhous... | 2019-12-23T10:57:03 | 2019-12-23T10:57:03 | {} | NONE |
ClickHouse | ClickHouse | 568,457,606 | 8,347 | nvartolomei | 1. When you insert into `db.origin-table` ClickHouse will automatically apply the SELECT statement from MV over the same batch of data and write synchronously data to `db.table`. By default this is sequential, but you can enable `parallel_view_processing` setting (Enables pushing to attached views concurrently instead ... | 2019-12-23T12:06:11 | 2019-12-23T12:06:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 568,569,266 | 8,300 | kolen | While it fixes problem with `contrib/openssl`, this problem persists with ICU, in `contrib/icu-cmake`
```
[1681/4437] Building ASM object contrib/icu-cmake/CMakeFiles/icudata.dir/__/icudata/icudt66l_dat.S.o
FAILED: contrib/icu-cmake/CMakeFiles/icudata.dir/__/icudata/icudt66l_dat.S.o
/Library/Developer/CommandLine... | 2019-12-23T20:04:46 | 2019-12-23T20:04:46 | {} | NONE |
ClickHouse | ClickHouse | 568,738,949 | 8,374 | abyss7 | It works for me. | 2019-12-24T12:27:44 | 2019-12-24T12:27:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 568,792,147 | 8,373 | den-crane | Also joinGet does not support Nulls. It is by design as well.
```
SELECT joinGet('join_table', 'value_to_update_main_table', ID) AS x
FROM main_table
┌──────────x─┐
│ 0000-00-00 │
│ 0000-00-00 │
│ 0000-00-00 │
│ 0000-00-00 │
└────────────┘
```
You can use `nullIf` or any other conditional functions.
`... | 2019-12-24T19:05:02 | 2019-12-24T19:05:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 568,793,459 | 8,373 | den-crane | > I understand, but how can I query these empty values after I ran this query that returned no errors?
`SELECT * FROM main_table where value_to_get_from_join = Date(0);`
joinGet does not return errors if a key does not exists.
It returns `default` value for value type -- String: '', Int* : 0, Date: Date(0), F... | 2019-12-24T19:19:19 | 2019-12-24T19:24:17 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,024,368 | 8,116 | godfreyd | Поправил | 2019-12-26T09:37:03 | 2019-12-26T09:37:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 568,428,585 | 8,339 | alesapin | Which version did you use before the update? Can you provide server logs when the server was started for the first time after the update? Also, I'd like to see the contents of corrupted parts directories `$ ls /data-2/data/ramen/systeminfohistory/201906_21_21_0_67/`. | 2019-12-23T10:07:31 | 2019-12-23T10:07:31 | {} | MEMBER |
ClickHouse | ClickHouse | 568,497,906 | 8,066 | KochetovNicolai | 1 stateless test has failed because there is no column `parts_to_do_names` in system.mutations yet | 2019-12-23T15:01:52 | 2019-12-23T15:01:52 | {} | MEMBER |
ClickHouse | ClickHouse | 568,517,800 | 8,354 | blitzarx1 | This helped! Thanks | 2019-12-23T16:23:29 | 2019-12-23T16:23:29 | {} | NONE |
ClickHouse | ClickHouse | 568,678,924 | 8,300 | tisonkun | Similar problem reported.
[output.txt](https://github.com/ClickHouse/ClickHouse/files/3997517/output.txt)
| 2019-12-24T07:21:07 | 2019-12-24T07:21:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 568,685,291 | 8,252 | vitlibar | Conflict with 19.16 and 19.17. | 2019-12-24T07:54:11 | 2019-12-24T09:56:45 | {} | MEMBER |
ClickHouse | ClickHouse | 568,849,554 | 8,369 | ghuname | I have extended the select:
`select avg(val), sum(val), count(val), any(val), groupArray(val), uniq(val), uniqExact(val), max(val), min(val)`
This is what I get:
```
┌─avg(val)─┬─sum(val)─┬─count(val)─┬─any(val)─┬─groupArray(val)───┬─uniq(val)─┬─uniqExact(val)─┬─max(val)─┬─min(val)─┐
│ nan │ nan │ ... | 2019-12-25T06:35:40 | 2019-12-25T06:36:44 | {} | NONE |
ClickHouse | ClickHouse | 568,851,328 | 8,369 | ghuname | Just to add how it looks in practice.
I have to replace this
`sum(weight * di * workingSec24h * errorsUpXday) / (sum(sum_weight) * sum(di * workingSec24h)) `
with
`if(isNan(sum(weight * di * workingSec24h * errorsUpXday) / (sum(sum_weight) * sum(di * workingSec24h))), null, sum(weight * di * workingSec24h * ... | 2019-12-25T06:49:11 | 2019-12-25T06:49:11 | {} | NONE |
ClickHouse | ClickHouse | 568,947,861 | 8,401 | alexey-milovidov | "Ideal" solution (if you will not propose better) would be:
- use `pcg64` to generate at least 64bits in one iteration;
- for every byte do the following in unrolled fashion:
- multiply by 95 in uint16 and right shift by 8 bits, look here for details: https://lemire.me/blog/2016/06/27/a-fast-alternative-to-the-modul... | 2019-12-26T01:56:58 | 2019-12-26T01:56:58 | {} | MEMBER |
ClickHouse | ClickHouse | 569,028,767 | 8,382 | alesapin | ```2019-12-25 22:01:05 00995_exception_while_insert: [ FAIL ] 600.01 sec. - Timeout!```
Doesn't seem to be related to changes. | 2019-12-26T10:02:27 | 2019-12-26T10:02:27 | {} | MEMBER |
ClickHouse | ClickHouse | 565,147,759 | 8,169 | 4ertus2 | It's should be quite simple to make last table in multiple joins qualified too here:
https://github.com/ClickHouse/ClickHouse/blob/master/dbms/src/Interpreters/JoinToSubqueryTransformVisitor.cpp#L157
Update: it does not help. | 2019-12-12T19:23:24 | 2019-12-23T13:35:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 568,525,282 | 8,336 | den-crane | Check that all shards have `<internal_replication>true</internal_replication>` in remote servers cluster description https://github.com/ClickHouse/ClickHouse/blob/master/dbms/tests/integration/test_distributed_ddl/configs/config.d/clusters.xml | 2019-12-23T16:54:59 | 2019-12-23T16:54:59 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 568,573,590 | 8,332 | vitlibar | There is no need to backport it to 19.16, 19.17, 19.18 because [19.18](https://github.com/ClickHouse/ClickHouse/tree/19.18/dbms/programs/server) has no support for Prometheus endpoint. | 2019-12-23T20:25:47 | 2019-12-23T20:25:47 | {} | MEMBER |
ClickHouse | ClickHouse | 568,686,256 | 8,298 | vitlibar | Conflict with 19.16 and 19.17. | 2019-12-24T07:58:51 | 2019-12-24T09:58:02 | {} | MEMBER |
ClickHouse | ClickHouse | 568,758,505 | 8,374 | alexey-milovidov | @zhang2014 This PR completely removes the usage of `libc-headers` - that's not what we need. | 2019-12-24T14:31:38 | 2019-12-24T14:31:38 | {} | MEMBER |
ClickHouse | ClickHouse | 568,781,242 | 7,603 | alexey-milovidov | This is caused by unhandled exception in descrutor. Exception may happen when a client has disconnected and we cannot send data. The exception is not handled correctly. | 2019-12-24T17:20:16 | 2019-12-24T17:20:16 | {} | MEMBER |
ClickHouse | ClickHouse | 568,881,011 | 8,366 | ghost | @den-crane
Thank you very much.
@qoega
I will close this issue. | 2019-12-25T10:21:08 | 2019-12-25T10:21:08 | {} | NONE |
ClickHouse | ClickHouse | 569,020,801 | 7,611 | frankieci | > @RaulGuo I ran into the same problem you had, and how did you solve it, or clickhouse distributed how to deploy using MergeTree table
Happy! I found its method to deploy for clickhouse distributed, then it works well
| 2019-12-26T09:18:33 | 2019-12-26T09:18:33 | {} | NONE |
ClickHouse | ClickHouse | 568,500,531 | 8,275 | filimonov | > We don't check consistency between types in table and dictionary even for ClickHouse source. So we can create a dictionary with `Date` types and table with `DateTime` types and get very confusing behavior.
Maybe that's also the reason of https://github.com/ClickHouse/ClickHouse/issues/8352 | 2019-12-23T15:13:11 | 2019-12-23T15:13:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 568,511,258 | 8,354 | den-crane | ```
SELECT farmHash64('Moscow')
┌─farmHash64('Moscow')─┐
│ 5992710078453357409 │
└──────────────────────┘
``` | 2019-12-23T15:57:39 | 2019-12-23T15:57:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 568,517,692 | 8,347 | nvartolomei | @byx313 exactly. | 2019-12-23T16:23:06 | 2019-12-23T16:23:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 568,685,004 | 8,234 | vitlibar | Conflict with 19.16. | 2019-12-24T07:52:39 | 2019-12-24T07:52:39 | {} | MEMBER |
ClickHouse | ClickHouse | 568,688,620 | 8,343 | Felixoid | I'm implementing integration tests here | 2019-12-24T08:09:55 | 2019-12-24T08:09:55 | {} | MEMBER |
ClickHouse | ClickHouse | 568,876,488 | 8,336 | godintsov | @den-crane this setting works! Thank you very much! | 2019-12-25T09:50:09 | 2019-12-25T09:50:09 | {} | NONE |
ClickHouse | ClickHouse | 568,431,612 | 8,288 | YiuRULE | Test concerning the `SYSTEM RELOAD DICTIONARY ON CLUSTER` has been fixed and resolve the merge conflict.
The tests who currently fail are not related to the new feature | 2019-12-23T10:18:44 | 2019-12-23T10:18:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 568,614,788 | 8,366 | den-crane | What do you mean "is uncompressed"?
Without LZ4 or ZSTD data compression? No such feature available. Compression is set at server / column level.
Uncompressed data will not make a system faster, because LZ4 is CPU lightweight compression but it saves a lot of disk I/O. So usually LZ4 compression makes a system fast... | 2019-12-24T00:40:53 | 2019-12-24T00:41:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 568,908,855 | 8,397 | alexey-milovidov | https://github.com/aklomp/base64 has poor design of initialization
and at least two new competitive libraries exist:
https://github.com/lemire/fastbase64
https://github.com/powturbo/TurboBase64 | 2019-12-25T15:22:00 | 2019-12-25T15:22:11 | {} | MEMBER |
ClickHouse | ClickHouse | 568,950,046 | 8,376 | vitlibar | Conflict with 19.16. | 2019-12-26T02:15:55 | 2019-12-26T02:15:55 | {} | MEMBER |
ClickHouse | ClickHouse | 568,970,620 | 7,611 | frankieci | @RaulGuo I ran into the same problem you had, and how did you solve it, or clickhouse distributed how to deploy using MergeTree table | 2019-12-26T05:03:14 | 2019-12-26T05:03:14 | {} | NONE |
ClickHouse | ClickHouse | 568,516,524 | 8,354 | den-crane | https://pypi.org/project/clickhouse-cityhash/ | 2019-12-23T16:18:25 | 2019-12-23T16:18:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 568,542,244 | 8,362 | KochetovNicolai | @azat cool! If you have a reproducible example which fails at least sometimes I would be happy to see it. | 2019-12-23T18:05:37 | 2019-12-23T18:05:37 | {} | MEMBER |
ClickHouse | ClickHouse | 568,792,697 | 8,373 | yehiaelbehery | I understand, but how can I query these empty values after I ran this query that returned no errors?
If joinGet() only works with (<> ''), shouldn't be implicit or not work at all? | 2019-12-24T19:10:43 | 2019-12-24T19:17:58 | {} | NONE |
ClickHouse | ClickHouse | 568,854,770 | 8,159 | gubinjie | @weiqxu
Hello, I tried again several times, but the error log shows dfs.ha.namenodes.ns not found.
1. This is part of my hdfs-site.xml configuration file. See below configuration
2.I put this hdfs-site.xml (hdfs-client.xm) file in the / etc / clickhouse-server folder
3.I will $ LIBHDFS3_CONF = / etc / clickhouse... | 2019-12-25T07:15:31 | 2019-12-25T07:15:31 | {} | NONE |
ClickHouse | ClickHouse | 568,872,415 | 8,373 | filimonov | Maybe joinGetOrNull should be added | 2019-12-25T09:19:43 | 2019-12-25T09:19:43 | {
"+1": 3
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 568,920,905 | 8,396 | MetalRex101 | Нашел вот такие записи
```
2019.12.16 13:12:23.773606 [ 58 ] {99ae1da7-a189-4075-a316-2e2840df5e31} <Information> db.locatable_entity_states (Replicated OutputStream): Block with ID 201912_17125759411864401470_8887152332147673266 already exists; ignoring it.
```
в количестве 4 штук. Других логов нет. | 2019-12-25T18:26:18 | 2019-12-25T18:26:18 | {} | NONE |
ClickHouse | ClickHouse | 568,927,039 | 7,238 | alexey-milovidov | Not relevant after https://github.com/ClickHouse/ClickHouse/pull/8387 | 2019-12-25T20:06:15 | 2019-12-25T20:06:15 | {} | MEMBER |
ClickHouse | ClickHouse | 568,573,623 | 8,291 | vitlibar | There is no need to backport it to 19.16, 19.17, 19.18 because [19.18](https://github.com/ClickHouse/ClickHouse/tree/19.18/dbms/src/DataTypes) doesn't have the data type `DateTime64`. | 2019-12-23T20:25:54 | 2019-12-23T20:25:54 | {} | MEMBER |
ClickHouse | ClickHouse | 568,716,893 | 8,198 | hieunguyen1809 | Problem is: old data with invalid format stuck in distributed table.
Solution: truncate data from distributed table. | 2019-12-24T10:23:00 | 2019-12-24T10:23:00 | {} | NONE |
ClickHouse | ClickHouse | 568,929,757 | 8,397 | alexey-milovidov | The library TurboBase64 is at least tens percent worse than aklomp/base64 on test:
```
clickhouse-benchmark <<< "SELECT sum(length(base64Encode(URL))) FROM test.hits"
```
```
localhost:9000, queries 2319, QPS: 8.579, RPS: 76129062.788, MiB/s: 6283.060, result RPS: 8.579, result MiB/s: 0.000.
0.000% 0... | 2019-12-25T20:20:42 | 2019-12-25T20:22:02 | {} | MEMBER |
ClickHouse | ClickHouse | 568,527,536 | 8,233 | Akazz | I have identified the commit, that apparently introduced this behaviour - https://github.com/ClickHouse/ClickHouse/pull/7894. Strangely enough, registering `fileSegmentationEngineTabSeparatedImpl` engine works okay, although that may depend on our tests suit.
The commit enables code paths resulting in lockups (like ... | 2019-12-23T17:03:28 | 2019-12-23T17:03:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 568,609,401 | 8,343 | alexey-milovidov | @alesapin Tests are Ok but new config for tests requires Docker image. What is the proper sequence to merge this? | 2019-12-23T23:59:37 | 2019-12-23T23:59:37 | {} | MEMBER |
ClickHouse | ClickHouse | 568,784,355 | 8,373 | den-crane | it's expected behavior.
joinGet() returns empty values so the alter updates using this empty values.
` WHERE 1;` -- wrong
` WHERE joinGet('join_table', 'value_to_update_main_table', ID) <> '' ` -- right
| 2019-12-24T17:47:29 | 2019-12-24T17:57:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 568,903,781 | 8,334 | maxulan | @alexey-milovidov Could you please review? | 2019-12-25T14:12:40 | 2019-12-25T14:12:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,031,038 | 8,403 | alexey-milovidov | Performance test is Ok for base64 functions. | 2019-12-26T10:15:18 | 2019-12-26T10:15:18 | {} | MEMBER |
ClickHouse | ClickHouse | 571,896,031 | 8,556 | Crazylearner30 | ok got it, thanks. | 2020-01-08T05:22:50 | 2020-01-08T05:22:50 | {} | NONE |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.