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 | 589,389,228 | 9,193 | den-crane | НЕ знаю кто накосячил в docs/ru/data_types/datetime.md
```
## Примеры
**1.** Создание таблицы с столбцом типа `DateTime` и вставка данных в неё:
CREATE TABLE dt
```
не хватает ```sql перед CREATE TABLE dt
| 2020-02-20T22:43:23 | 2020-02-20T22:43:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 589,509,500 | 4,729 | filimonov | For multi-disk case target table should have super set of disks of source table, in that case that move between tables (based on hardlinks) will be possible. After part moved from one table to another you can move it to another disk by `Alter table dest move partition to volume... ` | 2020-02-21T05:51:27 | 2020-02-21T05:51:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 589,549,030 | 9,267 | dependabot-preview[bot] | OK, I won't notify you again about this release, but will get in touch when a new version is available. If you'd rather skip all updates until the next major or minor version, let me know by commenting `@dependabot ignore this major version` or `@dependabot ignore this minor version`.
If you change your mind, just re-... | 2020-02-21T08:22:08 | 2020-02-21T08:22:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 589,577,053 | 8,603 | filimonov | Related to bug #2283 | 2020-02-21T09:41:19 | 2020-02-21T09:41:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 589,738,601 | 9,219 | alexey-milovidov | ```
../contrib/FastMemcpy/FastMemcpy.h:159:29: runtime error: load of misaligned address 0x00000769aadd for type 'uint64_t' (aka 'unsigned long'), which requires 8 byte alignment
0x00000769aadd: note: pointer points here
45 52 45 20 6e 61 6d 65 20 4c 49 4b 45 20 00 52 65 63 65 69 76 65 64 20 70 61 63 6b 65 74 20... | 2020-02-21T16:52:01 | 2020-02-21T16:52:01 | {} | MEMBER |
ClickHouse | ClickHouse | 589,784,644 | 7,592 | alexey-milovidov | It's outdated after we moved to C++20 because now we can try to use `char8_t` data type for this purpose. I started to work in `char8_t` branch but didn't finish.
By the way, did you notice some performance improvements? | 2020-02-21T18:49:39 | 2020-02-21T18:49:39 | {} | MEMBER |
ClickHouse | ClickHouse | 589,809,287 | 9,156 | alexey-milovidov | Sorry, it doesn't work as expected. I'm always using static linking. And now the linker doesn't show me errors. Instead, I get errors at startup time:
```
milovidov@milovidov-Pro-P30:~/work/ClickHouse/dbms/programs/server$ ./clickhouse-server
Processing configuration file 'config.xml'.
Merging configuration file ... | 2020-02-21T19:52:14 | 2020-02-21T19:52:14 | {} | MEMBER |
ClickHouse | ClickHouse | 589,827,954 | 9,150 | abyss7 | Fixed code is not present in 20.1 | 2020-02-21T20:45:03 | 2020-02-21T20:45:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 589,979,784 | 7,208 | alexey-milovidov | To be continued. | 2020-02-22T17:35:44 | 2020-02-22T17:35:44 | {} | MEMBER |
ClickHouse | ClickHouse | 589,018,159 | 9,249 | alexey-milovidov | > Probably we need rewrite CREATE query to add columns description even for table functions
Ok. | 2020-02-20T13:18:19 | 2020-02-20T13:18:19 | {} | MEMBER |
ClickHouse | ClickHouse | 589,289,885 | 9,195 | den-crane | ZK {table_name}/blocks
```
CREATE TABLE XXX(key nt64)
ENGINE = ReplicatedReplacingMergeTree
('/clickhouse/{cluster}/tables/XXX','{replica}')
order by key
insert into XXX values (1);
SELECT name
FROM system.zookeeper
WHERE path = '/clickhouse/stage1/tables/XXX/blocks'
┌─name─────────────────────────... | 2020-02-20T20:21:12 | 2020-02-20T20:22:25 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 589,609,413 | 8,290 | zhang2014 | 👍 Great feature !!
For `compact format` it will be append mode in the future (always write to one file for multiple INSERT queries) ? | 2020-02-21T11:10:58 | 2020-02-21T11:10:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 589,815,153 | 8,784 | den-crane | @tdingus https://clickhouse.tech/docs/en/query_language/dicts/external_dicts_dict_layout/#range-hashed `arbitrary numeric type` | 2020-02-21T20:07:44 | 2020-02-21T20:07:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 589,820,916 | 8,784 | tdingus | Let me confirm that data type such as String is not supported. Is that right?
If that is true, I wonder if this new feature would be considered in very near future?
Thanks. | 2020-02-21T20:24:14 | 2020-02-21T20:24:14 | {} | NONE |
ClickHouse | ClickHouse | 589,827,923 | 8,784 | den-crane | You mean range between strings ? `WHERE 'xxxx' between 'yyyy' and 'zzzz'` ?
| 2020-02-21T20:44:54 | 2020-02-21T20:44:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 589,738,834 | 9,219 | alexey-milovidov | It looks like FastMemcpy should be compiled without UBSan. **Upd: see below.** | 2020-02-21T16:52:33 | 2020-02-22T17:39:37 | {} | MEMBER |
ClickHouse | ClickHouse | 590,016,013 | 8,594 | alexey-milovidov | @excitoon Any updates?
@nvartolomei
> Ie. will it be possible to just remove a disk and ClickHouse will run part checks/repair for replicated tables and missing parts?
If parts are not available or not readable, ClickHouse will download them from replicas. Removing a disk and removing it from the configuratio... | 2020-02-23T01:24:30 | 2020-02-23T01:24:30 | {} | MEMBER |
ClickHouse | ClickHouse | 589,021,541 | 8,775 | CurtizJ | UBSan fail is not related to this PR. See #9250 | 2020-02-20T13:26:22 | 2020-02-20T13:26:37 | {} | MEMBER |
ClickHouse | ClickHouse | 589,054,948 | 9,245 | Felixoid | @BayoNet currently, the whole page https://clickhouse.tech/docs/en/operations/system_tables/#system_tables-replicas (except replicas and mutations) is in the style of the PR. | 2020-02-20T14:04:43 | 2020-02-20T14:04:43 | {} | MEMBER |
ClickHouse | ClickHouse | 589,685,942 | 9,235 | alexey-milovidov | > No space left on device.
- CI issue. | 2020-02-21T14:51:25 | 2020-02-21T14:51:25 | {} | MEMBER |
ClickHouse | ClickHouse | 589,759,588 | 7,455 | nvartolomei | Maybe it will suffice to add ProfileEvents column to system.part_log? merges, fetches, moves are already recorded to system.part_log. | 2020-02-21T17:45:33 | 2020-02-21T17:45:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 590,009,452 | 9,304 | alexey-milovidov | Performance difference is real.
Most representative queries:
`SELECT bitCount(toUInt8(number)) FROM system.numbers`
- 25% better;
`WITH number AS x SELECT sum(x < 1 ? 1 : (x < 5 ? 2 : 3)) FROM system.numbers`
- 40% better;
`SELECT count() FROM (SELECT range(number % 100) FROM system.numbers limit 10000000... | 2020-02-22T23:43:32 | 2020-02-22T23:43:32 | {
"rocket": 1
} | MEMBER |
ClickHouse | ClickHouse | 589,719,114 | 8,429 | alexey-milovidov | > Ok, we can store them in ColumnDescription, but we may need to accept the following case, can we accept that?
Yes, we're Ok to save type aliases only in `SHOW CREATE TABLE`, `DESC TABLE` and `system.columns`, but we should return our canonical names in all formats (Native, JSON, etc...) | 2020-02-21T16:07:23 | 2020-02-21T16:07:23 | {} | MEMBER |
ClickHouse | ClickHouse | 589,747,423 | 9,162 | kreuzerkrieg | > Maybe you know some gcc hackers/experts and can invite them here?
Unfortunately not, was hoping you have some.
The precompiled headers is just not the same as it on Windows, there it is widely used and considered first class citizen compiler feature. On Linux it is some obscure, rarely used feature. I will try an... | 2020-02-21T17:13:44 | 2020-02-21T17:13:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 589,853,614 | 8,784 | tdingus | Yes. | 2020-02-21T21:55:37 | 2020-02-21T21:55:37 | {} | NONE |
ClickHouse | ClickHouse | 590,025,414 | 8,854 | amosbird | An `SQL` table funtion then, though the result schema needs to be inferred. | 2020-02-23T04:27:13 | 2020-02-23T04:27:13 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 588,978,710 | 9,195 | BernieJiangDL | > > Any solutions for this case? Many thanks~
>
> Just do not use spark jdbc to insert into CH. Generate TSV files by Spark and upload them using custom code (&jdbc) which understands CH limits and constraints.
I ran a test for the deduplication part. For the simply way, I inserted one record twice by clickhouse ... | 2020-02-20T11:41:33 | 2020-02-20T11:41:33 | {} | NONE |
ClickHouse | ClickHouse | 589,456,069 | 8,525 | aa8808021 | > It means that we cannot read data from HDFS data nodes. You need to check network connectivity - are data nodes accessible from ClickHouse server or they are closed by firewall.
hi：
In order to verify this problem, we put the server under the same computer room. This problem still occurs. I want to ask if there is ... | 2020-02-21T01:32:22 | 2020-02-21T01:32:22 | {} | NONE |
ClickHouse | ClickHouse | 589,653,975 | 158 | yehiaelbehery | Do you plan to supporting date functions for dates before 1970 for birth dates, ... and other historic dates. Can you explain the philosohy behind the decision (if not time and resources)? | 2020-02-21T13:31:48 | 2020-02-21T13:31:48 | {} | NONE |
ClickHouse | ClickHouse | 589,680,778 | 158 | den-crane | @yehiaelbehery yes. It will be implemented in DateTime64 https://github.com/ClickHouse/ClickHouse/issues/7316 | 2020-02-21T14:39:05 | 2020-02-21T14:39:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 589,728,830 | 9,075 | nikitamikhaylov | last commit is incorrect | 2020-02-21T16:29:37 | 2020-02-21T16:29:37 | {} | MEMBER |
ClickHouse | ClickHouse | 589,880,263 | 9,288 | tavplubix | The tests should be ok after docker images are updated | 2020-02-21T23:30:49 | 2020-02-21T23:30:49 | {} | MEMBER |
ClickHouse | ClickHouse | 590,015,841 | 8,836 | alexey-milovidov | ? | 2020-02-23T01:21:04 | 2020-02-23T01:21:04 | {} | MEMBER |
ClickHouse | ClickHouse | 590,016,677 | 8,775 | alexey-milovidov | Please merge with master. | 2020-02-23T01:35:37 | 2020-02-23T01:35:37 | {} | MEMBER |
ClickHouse | ClickHouse | 589,024,046 | 8,990 | abyss7 | Probably the problem is with the linker: /usr/local/Cellar/llvm@8/8.0.1_1/bin/ld.lld - you should use native linker (ld.ld64 I guess). | 2020-02-20T13:32:07 | 2020-02-20T13:32:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 589,232,782 | 8,966 | alexey-milovidov | Superseded. | 2020-02-20T18:12:18 | 2020-02-20T18:12:18 | {} | MEMBER |
ClickHouse | ClickHouse | 589,459,767 | 7,080 | cw9 | Bump, is there any update on supporting this? This would be a super useful feature! | 2020-02-21T01:48:59 | 2020-02-21T01:48:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 589,110,984 | 9,248 | Akazz | It should be noted that the destructor `~StorageMemory()` gets called consistently and all of ColumnPtr's have their respective refcounts equal to 1!
The problem seems to be related to allocator | 2020-02-20T15:19:26 | 2020-02-21T08:24:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 589,673,762 | 8,290 | alexey-milovidov | But the parts itself remain atomic and immutable. | 2020-02-21T14:22:44 | 2020-02-21T14:22:44 | {} | MEMBER |
ClickHouse | ClickHouse | 589,688,832 | 7,274 | alexey-milovidov | @clemrodriguez Sorry for the delay. Yes, it's better to open a new PR. | 2020-02-21T14:58:03 | 2020-02-21T14:58:03 | {} | MEMBER |
ClickHouse | ClickHouse | 589,739,490 | 9,066 | abyss7 | No avro in 20.1 | 2020-02-21T16:54:16 | 2020-02-21T16:54:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 590,052,838 | 9,065 | amosbird | https://github.com/ClickHouse/ClickHouse/issues/9136 seems to be doable inside this PR | 2020-02-23T10:15:39 | 2020-02-23T10:15:39 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 590,060,347 | 9,196 | alexey-milovidov | > Or else we can let format_regexp handle row delimiter as well but need to limit the maxium length.
That's interesting idea. We can apply regexp sequentially on data or even just extract all matches and skip everything else. It will require additional copying, though.
Custom row separator is more simple to imple... | 2020-02-23T11:47:19 | 2020-02-23T11:47:19 | {} | MEMBER |
ClickHouse | ClickHouse | 589,715,996 | 9,126 | abyss7 | What is this? Improvement? Feature? Bugfix? | 2020-02-21T15:59:56 | 2020-02-21T15:59:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 589,758,377 | 9,277 | den-crane | duplicate https://github.com/ClickHouse/ClickHouse/issues/7316 | 2020-02-21T17:42:20 | 2020-02-21T17:42:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 589,785,190 | 7,592 | alexey-milovidov | https://github.com/ClickHouse/ClickHouse/compare/master...char8_t
Maybe we should continue? You can work on top of my branch or I will try to find a time to continue. | 2020-02-21T18:51:07 | 2020-02-21T18:51:07 | {} | MEMBER |
ClickHouse | ClickHouse | 590,009,884 | 9,304 | alexey-milovidov | > SELECT count() FROM (SELECT range(number % 100) FROM system.numbers limit 10000000)
3 times better.
The difference is because after switching from unsigned char to char8_t, the function `FunctionRange::executeInternal` started to be vectorized.
| 2020-02-22T23:49:46 | 2020-02-22T23:49:46 | {} | MEMBER |
ClickHouse | ClickHouse | 590,028,909 | 8,429 | zhang2014 | > Yes, we're Ok to save type aliases only in SHOW CREATE TABLE, DESC TABLE and system.columns, but we should return our canonical names in all formats (Native, JSON, etc...)
Ok, I try to finish it | 2020-02-23T05:25:15 | 2020-02-23T05:25:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 589,672,787 | 9,277 | alexey-milovidov | We plan to support them via separate data types: Date32, DateTime64. | 2020-02-21T14:20:20 | 2020-02-21T14:20:20 | {
"+1": 2
} | MEMBER |
ClickHouse | ClickHouse | 589,775,476 | 9,193 | den-crane | https://github.com/DarkWanderer/ClickHouse/blob/feature/docs/add-clickhouse-client/docs/en/data_types/datetime64.md
several example outputs without microseconds `03:00:00`
but they should
```
SELECT *
FROM dt
┌───────────────timestamp─┬─event_id─┐
│ 2019-01-01 03:00:00.000 │ 1 │
│ 2019-01-01 00:... | 2020-02-21T18:24:58 | 2020-02-21T18:27:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 589,855,773 | 9,271 | den-crane | https://clickhouse.tech/docs/en/operations/settings/query_complexity/#result-overflow-mode
_What to do if the volume of the result exceeds one of the limits: 'throw' or 'break'. By default, throw. Using 'break' is similar to using LIMIT._
but it could be broken.
| 2020-02-21T22:02:09 | 2020-02-21T22:02:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 589,014,988 | 8,990 | vigneshrajarr | [CMakeError.log](https://github.com/ClickHouse/ClickHouse/files/4230842/CMakeError.log)
[CMakeOutput.log](https://github.com/ClickHouse/ClickHouse/files/4230844/CMakeOutput.log)
| 2020-02-20T13:11:27 | 2020-02-20T13:11:27 | {} | NONE |
ClickHouse | ClickHouse | 589,234,765 | 9,255 | alexey-milovidov | > Build failed without build log. It's inner CI problem. Maybe something wrong with docker or host. Just wait for build restart.
Ok. | 2020-02-20T18:16:46 | 2020-02-20T18:16:46 | {} | MEMBER |
ClickHouse | ClickHouse | 589,738,020 | 9,126 | alexey-milovidov | Improvement. | 2020-02-21T16:50:36 | 2020-02-21T16:50:36 | {} | MEMBER |
ClickHouse | ClickHouse | 589,546,189 | 8,701 | qoega | Can you add explicit section into Replicated* docs which alters are metadata only? | 2020-02-21T08:13:29 | 2020-02-21T08:13:29 | {} | MEMBER |
ClickHouse | ClickHouse | 589,562,254 | 9,260 | alesapin | Possibly fixed in one of https://github.com/ClickHouse/ClickHouse/pull/9050, https://github.com/ClickHouse/ClickHouse/pull/9106, https://github.com/ClickHouse/ClickHouse/pull/9201. | 2020-02-21T09:01:58 | 2020-02-21T09:03:11 | {} | MEMBER |
ClickHouse | ClickHouse | 589,681,325 | 9,151 | alexey-milovidov | The issue was accidentially fixed in #8929 | 2020-02-21T14:40:28 | 2020-02-21T14:40:28 | {} | MEMBER |
ClickHouse | ClickHouse | 589,719,261 | 8,429 | alexey-milovidov | No setting needed. | 2020-02-21T16:07:46 | 2020-02-21T16:07:46 | {} | MEMBER |
ClickHouse | ClickHouse | 589,786,571 | 9,196 | Avogar | @tavplubix, can you please review the changes? | 2020-02-21T18:54:55 | 2020-02-21T18:54:55 | {} | MEMBER |
ClickHouse | ClickHouse | 589,961,098 | 9,304 | alexey-milovidov | secure, shard_secure - temporarily broken in CI. | 2020-02-22T14:27:03 | 2020-02-22T14:27:03 | {} | MEMBER |
ClickHouse | ClickHouse | 589,384,270 | 4,729 | nvartolomei | https://github.com/ClickHouse/ClickHouse/pull/6168 | 2020-02-20T22:30:51 | 2020-02-20T22:30:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 590,018,095 | 9,196 | amosbird | Having another regex option would be more useful : `format_regexp_row_delimiter`. For example , vertical formats with `\n\n` as delimiter.
Or else we can let `format_regexp` handle row delimiter as well but need to limit the maxium length.
| 2020-02-23T02:02:28 | 2020-02-23T02:02:28 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 589,006,403 | 9,162 | kreuzerkrieg | > Will try locally.
Dont know why we dont see any benefit, but I have a theory, will check it later
| 2020-02-20T12:52:43 | 2020-02-20T12:52:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 589,021,999 | 8,990 | abyss7 | What version of ClickHouse do you try to build? | 2020-02-20T13:27:28 | 2020-02-20T13:27:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 589,500,750 | 8,990 | vigneshrajarr | >What version of ClickHouse do you try to build?
How can I check that ?
Latest one in CHANGELOG.md is v20.1.2.4. Will this help ?
>Probably the problem is with the linker: /usr/local/Cellar/llvm@8/8.0.1_1/bin/ld.lld - you should use native linker (ld.ld64 I guess).
Do you mean `ld64.lld` ? Removing `ld.lld` d... | 2020-02-21T05:07:51 | 2020-02-21T05:07:51 | {} | NONE |
ClickHouse | ClickHouse | 589,687,926 | 9,235 | alexey-milovidov | Thank you! | 2020-02-21T14:56:05 | 2020-02-21T14:56:05 | {} | MEMBER |
ClickHouse | ClickHouse | 589,773,682 | 9,162 | alexey-milovidov | http://llunak.blogspot.com/2019/05/why-precompiled-headers-do-not-improve.html
https://www.reddit.com/r/cpp/comments/bsdrkv/why_precompiled_headers_do_not_improve_c_compile/
https://stackoverflow.com/questions/13309228/gcc-build-time-doesnt-benefit-much-from-precompiled-headers
| 2020-02-21T18:22:53 | 2020-02-21T18:22:53 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 589,850,796 | 9,270 | den-crane | Probably it's a bug.
CH does not support nulls in dictionary.
So instead of `col6 Float64,` it should by `col6 Float64 DEFAULT 0`
`col12 String,` -> `col12 String DEFAULT '', ` | 2020-02-21T21:46:47 | 2020-02-21T21:46:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 590,010,449 | 9,304 | alexey-milovidov | For `WITH number AS x SELECT sum(x < 1 ? 1 : (x < 5 ? 2 : 3)) FROM system.numbers`
the difference in generated code looks like this:
```
│43b:┌─ mov 0x10(%r13),%rdx ... | 2020-02-22T23:58:17 | 2020-02-22T23:58:17 | {} | MEMBER |
ClickHouse | ClickHouse | 590,014,993 | 8,290 | alexey-milovidov | Now let's enable them for system logs for size less than 1 MB. | 2020-02-23T01:07:11 | 2020-02-23T01:07:11 | {} | MEMBER |
ClickHouse | ClickHouse | 590,034,246 | 9,222 | blinkov | @tac0x2a it seems like you are putting text of article inside symlink destination, instead of removing symlink (that was leading to en version) and creating a file with translation in its place. Could you fix it? | 2020-02-23T07:05:57 | 2020-02-23T07:05:57 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 590,036,437 | 9,270 | chu1070y | Thanks for your comment, I changed nullable column to default.
but, the bug still occurs. | 2020-02-23T07:42:10 | 2020-02-23T07:42:46 | {} | NONE |
ClickHouse | ClickHouse | 589,245,889 | 9,251 | alexey-milovidov | But it will be done only after we will allow moving of unqualified names and aliases that is more important for TPC-H. | 2020-02-20T18:42:27 | 2020-02-20T18:42:27 | {} | MEMBER |
ClickHouse | ClickHouse | 589,641,305 | 8,990 | abyss7 | You should use CMake option `-DLINKER_NAME=/path/to/linker` to force using native linker. | 2020-02-21T12:54:05 | 2020-02-21T12:54:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 589,783,442 | 7,455 | alexey-milovidov | @nvartolomei Yes, it will be a good solution. | 2020-02-21T18:46:28 | 2020-02-21T18:46:28 | {} | MEMBER |
ClickHouse | ClickHouse | 589,814,188 | 8,784 | tdingus | My feedback,
a. I cannot find relevant documents to explain how this range is checked, so the logic is mysterious.
b. Can range be used for different data type, for example, IP addresses, or Strings?
Thanks. | 2020-02-21T20:05:08 | 2020-02-21T20:05:08 | {} | NONE |
ClickHouse | ClickHouse | 590,060,448 | 9,196 | alexey-milovidov | @amosbird You can also help with reviewing the code if you like :) | 2020-02-23T11:48:56 | 2020-02-23T11:49:09 | {} | MEMBER |
ClickHouse | ClickHouse | 589,245,600 | 9,251 | alexey-milovidov | We discussed that the current method to extract and move conditions from WHERE to ON is strange and going to be rewritten. | 2020-02-20T18:41:44 | 2020-02-20T18:41:44 | {} | MEMBER |
ClickHouse | ClickHouse | 589,626,654 | 8,290 | CurtizJ | > For `compact format` it will be append mode in the future (always write to one file for multiple INSERT queries) ?
It will be implemented as in-memory parts with WAL, which is the append-only file in Native format.
| 2020-02-21T12:07:00 | 2020-02-21T12:07:00 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 589,798,240 | 9,290 | alexey-milovidov | Tested manually.
```
./clickhouse-server
```
- has colors;
```
./clickhouse-server 2>&1 | less
```
- no colors. | 2020-02-21T19:24:14 | 2020-02-21T19:24:14 | {} | MEMBER |
ClickHouse | ClickHouse | 589,969,916 | 9,276 | Tri0L | ```
:) KILL MUTATION WHERE database = 'db' AND table = 'table';
KILL MUTATION WHERE (database = 'db') AND (table = 'table') ASYNC
Timeout exceeded while receiving data from server. Waited for 300 seconds, timeout is 300 seconds.
Cancelling query.
``` | 2020-02-22T16:03:45 | 2020-02-22T16:03:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 590,016,050 | 8,377 | alexey-milovidov | Your own test have failed. | 2020-02-23T01:25:03 | 2020-02-23T01:25:03 | {} | MEMBER |
ClickHouse | ClickHouse | 588,994,579 | 9,162 | abyss7 | Will try locally. | 2020-02-20T12:22:31 | 2020-02-20T12:22:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 589,234,568 | 9,260 | illinar | We suspect that this might be somehow related to skipping indices making it to production in v20.1
Downgrading server to v19.17 resolves the problem.
Clarification - we don't use skipping indices. | 2020-02-20T18:16:22 | 2020-02-20T18:38:03 | {} | NONE |
ClickHouse | ClickHouse | 589,542,986 | 9,193 | qoega | @BayoNet can you check style, please? | 2020-02-21T08:03:15 | 2020-02-21T08:03:15 | {} | MEMBER |
ClickHouse | ClickHouse | 589,712,135 | 9,162 | kreuzerkrieg | > @kreuzerkrieg Removing includes from source files will make the code less readable and will compromise code navigation - not acceptable.
Thats why the commit wasnt squashed with others. Just was checking my theory and it didnt work at all. I doubt @abyss7 will see anything interesting in his local checks, but who ... | 2020-02-21T15:51:10 | 2020-02-21T15:51:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 589,737,648 | 9,162 | alexey-milovidov | Maybe you know some gcc hackers/experts and can invite them here? | 2020-02-21T16:49:47 | 2020-02-21T16:49:47 | {} | MEMBER |
ClickHouse | ClickHouse | 589,952,457 | 8,990 | fredchenbj | using the native linker `/usr/bin/ld` would be ok. @vigneshrajarr | 2020-02-22T12:40:32 | 2020-02-22T12:43:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 590,012,695 | 9,304 | alexey-milovidov | Congratulations! It is a clear win! | 2020-02-23T00:26:14 | 2020-02-23T00:26:14 | {} | MEMBER |
ClickHouse | ClickHouse | 590,015,829 | 8,840 | alexey-milovidov | ? | 2020-02-23T01:20:48 | 2020-02-23T01:20:48 | {} | MEMBER |
ClickHouse | ClickHouse | 590,058,572 | 9,316 | alexey-milovidov | Ok. Only minor changes required... | 2020-02-23T11:25:09 | 2020-02-23T11:25:09 | {} | MEMBER |
ClickHouse | ClickHouse | 589,011,644 | 8,990 | abyss7 | Please, provide the contents of CMakeFiles/CMakeError.log | 2020-02-20T13:04:11 | 2020-02-20T13:04:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 589,018,564 | 8,990 | alexey-milovidov | > We don't support build with Apple Clang.
Let's add a check in CMake with a proper error message? | 2020-02-20T13:19:12 | 2020-02-20T13:19:12 | {} | MEMBER |
ClickHouse | ClickHouse | 589,180,331 | 9,082 | alexey-milovidov | > Functional stateless tests (thread) — fail:21, passed:1216, skipped:3
Looks like OOM in CI. | 2020-02-20T16:52:08 | 2020-02-20T16:52:08 | {} | MEMBER |
ClickHouse | ClickHouse | 589,708,731 | 9,162 | alexey-milovidov | @kreuzerkrieg Removing includes from source files will make the code less readable and will compromise code navigation - not acceptable. | 2020-02-21T15:43:27 | 2020-02-21T15:43:27 | {} | MEMBER |
ClickHouse | ClickHouse | 589,719,410 | 8,429 | alexey-milovidov | Let's continue :) | 2020-02-21T16:08:06 | 2020-02-21T16:08:06 | {} | MEMBER |
ClickHouse | ClickHouse | 589,841,409 | 9,288 | alexey-milovidov | Unbundled: https://a.yandex-team.ru/review/1145630/details | 2020-02-21T21:25:04 | 2020-02-21T21:25:35 | {} | MEMBER |
ClickHouse | ClickHouse | 589,980,150 | 9,219 | alexey-milovidov | You forgot to compare the MemCpy source code for patches. | 2020-02-22T17:39:10 | 2020-02-22T17:39:10 | {} | MEMBER |
ClickHouse | ClickHouse | 590,015,848 | 8,597 | alexey-milovidov | ? | 2020-02-23T01:21:13 | 2020-02-23T01:21:13 | {} | MEMBER |
ClickHouse | ClickHouse | 600,618,509 | 9,715 | alexey-milovidov | Don't understand why it's better than the 2nd.
Example:
`SELECT x + 1 AS x, x FROM (SELECT 1 AS x) t`
In current version the query will return:
`2, 2`
If we implement the first variant, the query will return:
`2, 2`
If we implement the second variant, the query will return:
`2, 1`
Another example:
... | 2020-03-18T13:20:34 | 2020-03-18T13:20:34 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.