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 | 535,508,354 | 6,781 | abyss7 | Please provide description in some way about the feature in this PR | 2019-09-26T13:44:23 | 2019-09-26T13:44:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 535,605,673 | 7,113 | alesapin | You can send data and queries to clickhouse-feedback@yandex-team.ru. | 2019-09-26T17:26:19 | 2019-09-26T17:26:19 | {} | MEMBER |
ClickHouse | ClickHouse | 535,802,934 | 7,092 | YiuRULE | Done ! :)
I modified one integration test where the http server do a simili-authentication system | 2019-09-27T06:12:49 | 2019-09-27T06:12:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 536,051,906 | 7,007 | nicelulu | For ```Functional stateless tests``` "00017_in_subquery_with_empty_result"
This PR result :
```
SET output_format_write_statistics = 0
Ok.
0 rows in set. Elapsed: 0.004 sec.
```
```
SELECT count()
FROM
(
SELECT *
FROM system.numbers
LIMIT 1000
)
WHERE 1 IN
(
SELECT 0
WHERE 0
... | 2019-09-27T18:34:53 | 2019-09-27T18:35:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 536,073,058 | 7,007 | alexey-milovidov | > Maybe it's normal?
Yes, it's perfect. | 2019-09-27T19:45:09 | 2019-09-27T19:45:09 | {} | MEMBER |
ClickHouse | ClickHouse | 536,326,440 | 7,120 | amosbird | Is it for packaging without docker? What's the gain of this instead of polishing the output of dpkg alien? | 2019-09-29T18:09:26 | 2019-09-29T18:09:26 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 536,487,607 | 7,151 | alexey-milovidov | You can send SIGHUP to clickhouse-server to create next log file. | 2019-09-30T09:46:07 | 2019-09-30T09:46:07 | {} | MEMBER |
ClickHouse | ClickHouse | 534,991,713 | 7,095 | aaa-user | Version of ClickHouse 19.9.2.4 | 2019-09-25T12:11:30 | 2019-09-25T12:11:30 | {} | NONE |
ClickHouse | ClickHouse | 535,031,308 | 7,094 | alexey-milovidov | That's reasonable.
Adding a matcher to match any event except one should be very easy. | 2019-09-25T13:47:04 | 2019-09-25T13:47:04 | {} | MEMBER |
ClickHouse | ClickHouse | 535,154,070 | 7,093 | alexey-milovidov | Need to add a test that the value of this ProfileEvent is greater than zero after merge. | 2019-09-25T18:33:44 | 2019-09-25T18:33:44 | {} | MEMBER |
ClickHouse | ClickHouse | 535,338,628 | 4,625 | andrewrbe | It's prestable version. The issue was with table which i don't use any more | 2019-09-26T05:09:15 | 2019-09-26T05:09:15 | {} | NONE |
ClickHouse | ClickHouse | 535,367,547 | 7,092 | alexey-milovidov | Ok. Now it's almost ready!
We also need to update integration test to ensure that this feature continue to work.
Look at `tests/integration//test_redirect_url_storage`
or
`tests/queries/0_stateless/00646_url_engine.python`
| 2019-09-26T06:59:03 | 2019-09-26T06:59:03 | {} | MEMBER |
ClickHouse | ClickHouse | 536,060,146 | 7,117 | alexey-milovidov | Maybe it's an issue with `unixodbc` library. It needs further investigation.
We need to check (with debug prints or whatever):
- how the connection string is being passed to unixodbc;
- what DSN is searched inside `unixodbc`.
Does the issue get "fixed" if you remove backslash from DSN? | 2019-09-27T19:00:18 | 2019-09-27T19:35:38 | {} | MEMBER |
ClickHouse | ClickHouse | 536,404,533 | 6,460 | amosbird | Without shipped ICU, it breaks gentoo clang build too (with contrib libc++).
```
ld.lld: error: undefined symbol: std::condition_variable::condition_variable()
>>> referenced by umutex.ao:(icu::initCondition()) in archive /home/amos/gentoo/usr/lib/libicuuc.a
>>> referenced by umutex.ao:(icu::UConditionVar::UCondi... | 2019-09-30T05:16:27 | 2019-09-30T05:16:27 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 535,014,247 | 7,088 | alexey-milovidov | If it will have no performance impact (that's unlikely), we can even get rid of a huge chunk of code, that it very optimistic. | 2019-09-25T13:09:34 | 2019-09-25T13:09:34 | {} | MEMBER |
ClickHouse | ClickHouse | 535,017,347 | 7,091 | silviucpp | Thanks. good to know. I'll look into `cancel_http_readonly_queries_on_client_close` and debug a bit to see if I can fix the problem.
Anyway this won't fix the `max_execution_time` so I will let this open. | 2019-09-25T13:17:13 | 2019-09-25T13:17:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 535,075,099 | 6,107 | filimonov | Workaround:
add a line `<allow_experimental_data_skipping_indices>1</allow_experimental_data_skipping_indices>` into users.xml in profiles / default subsection. | 2019-09-25T15:22:40 | 2019-09-25T15:22:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 535,321,623 | 7,092 | YiuRULE | Ok done :) | 2019-09-26T03:53:57 | 2019-09-26T03:53:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 535,343,073 | 4,625 | filimonov | Prestable version should not segfault. That case should be checked | 2019-09-26T05:29:07 | 2019-09-26T05:29:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 535,701,926 | 7,121 | den-crane | http://localhost:8123/?format_csv_delimiter=|&query
`echo "1;2"|curl 'http://localhost:8123/?format_csv_delimiter=;query=INSERT%20INTO%20test.csv_stream%20format%20CSV' --data-binary @-` | 2019-09-26T21:58:47 | 2019-09-26T21:59:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 536,062,350 | 7,117 | alexey-milovidov | BTW, do you literally have `ServerName\InstanceName` in your odbc.ini or the backslash has some special meaning? | 2019-09-27T19:06:45 | 2019-09-27T19:06:45 | {} | MEMBER |
ClickHouse | ClickHouse | 536,379,697 | 7,120 | amosbird | > for whatever reason.
Yeah, I asked because I'm also doing some package chores, and am afraid of the pitfalls that dpkg alien might entail. | 2019-09-30T02:50:47 | 2019-09-30T02:50:47 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 536,457,755 | 7,117 | nicolasblaye | Sorry for the delay. It doesn't work with the data source name not found error.
EDIT: The curly braces are valid though, if I try with {ServerName} it's ok. I guess we will need to use an alias but I find it weird that we can't select an instance from a server | 2019-09-30T08:25:30 | 2019-09-30T08:28:17 | {} | NONE |
ClickHouse | ClickHouse | 536,492,107 | 7,141 | alexey-milovidov | If you want to create Distributed tables that will look at multiple clusters and unify them as shards, you can specify a cluster that will contain all of the shards of all clusters and create Distributed table with this unified cluster. | 2019-09-30T09:58:19 | 2019-09-30T09:58:19 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 534,967,383 | 7,091 | filimonov | Can comment on that part:
> no matter if you close the client connection.
By default HTTP doesn't kill the queries after disconnect, there is `cancel_http_readonly_queries_on_client_close` setting (which is broken in last releases if that report is correct: https://github.com/ClickHouse/ClickHouse/issues/7019 ) | 2019-09-25T10:55:45 | 2019-09-25T10:56:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 535,097,336 | 7,073 | filimonov | Before:
```
➜ sudo yum install *.rpm
Loaded plugins: fastestmirror, langpacks, priorities
Examining clickhouse-client-19.14.6.12-2.noarch.rpm: clickhouse-client-19.14.6.12-2.noarch
Marking clickhouse-client-19.14.6.12-2.noarch.rpm as an update to clickhouse-client-19.4.4.33-1.el7.x86_64
Examining clickhouse-co... | 2019-09-25T16:13:37 | 2019-09-25T16:14:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 535,625,611 | 3,268 | cw9 | Hi, I'm hitting this issue as well, any plan to get this fixed? | 2019-09-26T18:17:30 | 2019-09-26T18:17:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 535,791,133 | 7,109 | yuzhichang | @alexey-milovidov Renamed per your suggestion. | 2019-09-27T05:24:10 | 2019-09-27T05:24:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 535,967,632 | 7,126 | den-crane | Deduplication works for blocks.
So if your blocks / batches completely the same (though it's super unusual) you can add some fake column to your table (batchID UUID) to break deduplication. And fill this column with uuid (the same uuid value for whole insert) at the client side.
Or you can disable deduplication... | 2019-09-27T14:38:46 | 2019-09-27T14:42:18 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 536,066,270 | 7,132 | alexey-milovidov | Unit tests — fail: 1 - flap. | 2019-09-27T19:20:29 | 2019-09-27T19:20:29 | {} | MEMBER |
ClickHouse | ClickHouse | 536,072,949 | 7,117 | nicolasblaye | Will try that tomorrow :) | 2019-09-27T19:44:48 | 2019-09-27T19:44:48 | {} | NONE |
ClickHouse | ClickHouse | 535,325,661 | 4,625 | filimonov | @andrewrbe can you create a separate issue? | 2019-09-26T04:14:13 | 2019-09-26T04:14:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 535,429,903 | 6,955 | yingfeng | Implemented by PR [7109](https://github.com/ClickHouse/ClickHouse/pull/7109) | 2019-09-26T09:52:39 | 2019-09-26T09:52:39 | {} | NONE |
ClickHouse | ClickHouse | 535,550,721 | 7,104 | zhang2014 | Can you put this pr to #7106 and close this pr? Because they are all modifying a file.
| 2019-09-26T15:12:05 | 2019-09-26T15:12:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 535,976,043 | 6,661 | arruzk | @romankos yes, I truncated table before running a new test.
Maybe we are talking about slightly different problems. I'm inserting data (of course different than in Kafka topic) to the table with `Kafka` engine directly (https://github.com/ClickHouse/ClickHouse/issues/5834 it was added in version 19.14.3.3). Consumers ... | 2019-09-27T14:59:15 | 2019-09-27T14:59:15 | {} | NONE |
ClickHouse | ClickHouse | 536,463,326 | 7,141 | filimonov | AFAIK that story is in the backlog, but I'm not sure about terms and schedule.
@alexey-milovidov can you clarify that? | 2019-09-30T08:41:08 | 2019-09-30T08:41:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 535,039,516 | 7,096 | alexey-milovidov | That's exactly what we need. | 2019-09-25T14:04:56 | 2019-09-25T14:04:56 | {} | MEMBER |
ClickHouse | ClickHouse | 535,393,979 | 6,459 | yathatguy | > You can get this branch `remotes/yandex/19.8`
It seems like it still connot be build from this branch
```
-- The following OPTIONAL packages have not been found:
* Threads
-- Using OpenSSL from /Users/thatguy/Downloads/ClickHouse/contrib/ssl
CMake Error at /usr/local/Cellar/cmake/3.15.3/share/cmake/Mod... | 2019-09-26T08:16:12 | 2019-09-26T08:16:12 | {
"+1": 2
} | NONE |
ClickHouse | ClickHouse | 535,627,922 | 7,070 | alexey-milovidov | optimisation -> optimization | 2019-09-26T18:23:43 | 2019-09-26T18:23:43 | {} | MEMBER |
ClickHouse | ClickHouse | 536,463,813 | 7,147 | Tasselmi | > > BUT what -1 mean ?
>
> `-1` means that last element of array (`2`) minus previous one (`3`) equals `2 - 3 = -1`
>
> Isn't it the expected behaviour?
>
> > array(T) takes T type, but a function can not support String type......
>
> Does strings support substraction? What is the result of 'Hello' minus 'w... | 2019-09-30T08:42:31 | 2019-09-30T08:42:31 | {} | NONE |
ClickHouse | ClickHouse | 535,041,821 | 7,095 | alexey-milovidov | Current behaviour of function count is:
- accept any number of arguments and count how many times no arguments were NULL.
I checked that MySQL and PostgreSQL allows only single argument for function count:
http://sqlfiddle.com/#!17/0a28f/264
So, we can also allow only single argument or zero number of arguments... | 2019-09-25T14:10:09 | 2019-09-25T14:10:09 | {} | MEMBER |
ClickHouse | ClickHouse | 535,535,675 | 7,111 | alesapin | If one run of integration tests passed, so it can be merged. | 2019-09-26T14:41:13 | 2019-09-26T16:13:38 | {} | MEMBER |
ClickHouse | ClickHouse | 535,824,429 | 7,113 | valenciaj | Send it! | 2019-09-27T07:24:54 | 2019-09-27T07:24:54 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 535,936,415 | 6,227 | orloffv | I would also like to be able to reset to default | 2019-09-27T13:21:53 | 2019-09-27T13:21:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 535,971,107 | 7,125 | den-crane | Enable directIO for all queries
min_bytes_to_use_direct_io=1 | 2019-09-27T14:47:08 | 2019-09-27T14:47:08 | {
"+1": 6
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 536,067,457 | 7,117 | nicolasblaye | Yes we have \ in the name. It works when we connect via dbeaver or ssm.
I need to try without backlash but I doubt it will work | 2019-09-27T19:24:34 | 2019-09-27T19:24:34 | {} | NONE |
ClickHouse | ClickHouse | 536,286,020 | 7,054 | yingfeng | Caused by issue [6956](https://github.com/ClickHouse/ClickHouse/issues/6956), the query analysis has taken so long a time that it's a misunderstanding for parallel execution. After the query has been finished prepared, the query pipeline is executed in parallel. | 2019-09-29T11:17:00 | 2019-09-29T11:17:00 | {} | NONE |
ClickHouse | ClickHouse | 536,463,800 | 7,141 | filimonov | See also: #1726 | 2019-09-30T08:42:28 | 2019-09-30T08:42:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 535,021,279 | 7,096 | alexey-milovidov | > 1. Have a logging function that only takes a message
Ok.
> 2. Switch to a printf-like format strings.
Ok. The main reason is log compression proposed by @4ertus2 (exactly what you described).
Not with printf-like format strings but with python/boost::format/fmt-like format strings.
Formatting can continue ... | 2019-09-25T13:26:09 | 2019-09-25T13:26:09 | {} | MEMBER |
ClickHouse | ClickHouse | 535,106,750 | 7,052 | filimonov | There is workaround which will give the results you expect without data copy.
Let's say you have a table `test`. You can rename it to the `test_old`, create table `test_new` with proper keys, create table `test` with Engine=Merge() above those 2 tables, and switch inserts to test_new. Selects will see data from both... | 2019-09-25T16:37:03 | 2019-09-25T16:37:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 535,154,201 | 7,093 | alexey-milovidov | You should fill the description for changelog. | 2019-09-25T18:34:03 | 2019-09-25T18:34:03 | {} | MEMBER |
ClickHouse | ClickHouse | 535,555,359 | 6,984 | zhang2014 | @alesapin @alexey-milovidov
I'm not sure if this PR needs some other modifications, but it has almost solved all my doubts.
| 2019-09-26T15:22:12 | 2019-09-26T15:22:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 536,445,053 | 7,146 | pyscala | got | 2019-09-30T07:47:32 | 2019-09-30T07:47:32 | {} | NONE |
ClickHouse | ClickHouse | 536,482,511 | 7,120 | filimonov | My goals are:
1) confirm that ClickHouse can be build properly on centos (and potentially some other rpm based distros) using avaliable there packages / compilers (lot of corp. users use centos, and prefer to make their own builds, so it's good to guarantee that it works 'out of the box').
2) provide a proper spec f... | 2019-09-30T09:31:39 | 2019-09-30T09:56:17 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 534,985,652 | 7,094 | killwort | **Что**
Иногда при анализе цепочек событий хочется добавить условие отсутсвия какого-либо события. Например у нас есть табличка действий людей со строками такого вида: `(person_id String, deed String, timestamp DateTime)`.
Пусть в ней будут записи типа `('person1','born','2000-01-01 00:00:00')`, `('person1','graduate... | 2019-09-25T11:53:45 | 2019-09-25T11:53:45 | {} | NONE |
ClickHouse | ClickHouse | 535,084,402 | 6,948 | imiskolee | @filimonov I using a string timezone field like "HongKong/Asia". but we also can be try using offset,but need migrate. thanks! | 2019-09-25T15:43:22 | 2019-09-25T15:43:22 | {} | NONE |
ClickHouse | ClickHouse | 535,165,042 | 7,087 | alexey-milovidov | It's not added to ToC: https://clickhouse.yandex/docs/en/development/build_cross/ | 2019-09-25T19:00:59 | 2019-09-25T19:00:59 | {} | MEMBER |
ClickHouse | ClickHouse | 535,165,894 | 7,087 | alexey-milovidov | It isn't added to zh, fa: https://github.com/ClickHouse/ClickHouse/tree/master/docs/zh/development | 2019-09-25T19:03:07 | 2019-09-25T19:03:07 | {} | MEMBER |
ClickHouse | ClickHouse | 535,408,314 | 7,105 | lvxin1986 | I hereby agree to the terms of the CLA available at: https://yandex.ru/legal/cla/?lang=en
For changelog. Remove if this is non-significant change.
Category :
Other | 2019-09-26T08:54:45 | 2019-09-26T08:54:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 535,436,801 | 6,880 | yingfeng | Update
PR [7109](https://github.com/ClickHouse/ClickHouse/pull/7109) has been submitted, which introduces aggregation families to bitmap functions. In cases that aggregate bitmap functions could replace non-aggregate bitmap ones, the performance could be improved remarkably, because of the overhead of serialization ... | 2019-09-26T10:11:25 | 2019-09-26T10:46:39 | {} | NONE |
ClickHouse | ClickHouse | 535,841,111 | 5,491 | thyn | any updates here? | 2019-09-27T08:15:32 | 2019-09-27T08:15:32 | {} | NONE |
ClickHouse | ClickHouse | 535,901,047 | 6,661 | romankos | Temporary solution for preventing duplicates - to use only 1 consumer:
kafka_num_consumers = 1
But this strange situation with duplicates must be solved on ClickHouse side. | 2019-09-27T11:27:37 | 2019-09-27T11:27:37 | {} | NONE |
ClickHouse | ClickHouse | 536,441,533 | 7,113 | valenciaj | Need more data? | 2019-09-30T07:36:49 | 2019-09-30T07:36:49 | {} | NONE |
ClickHouse | ClickHouse | 536,479,311 | 7,106 | blinkov | @zhang2014 so what do you think about these 3 PRs? | 2019-09-30T09:22:46 | 2019-09-30T09:22:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 536,491,145 | 7,141 | alexey-milovidov | Distributed table on Distributed table is supported if one of them contains no more than one shard.
For example, you can point to distributed table on one cluster from another cluster; you can even specify more than one replica for load balancing. | 2019-09-30T09:55:40 | 2019-09-30T09:55:40 | {} | MEMBER |
ClickHouse | ClickHouse | 534,832,744 | 7,077 | yingfeng | @alexey-milovidov
Thanks for your advice. We will add performance test case to CI later after we had got familiar with corresponding rules.
Regarding to escaping of zero bytes by `\0`, it seems not that necessary due to the PR [6773](https://github.com/ClickHouse/ClickHouse/pull/6773),
```
@@ -218,9 +218,11 ... | 2019-09-25T03:14:44 | 2019-09-25T03:18:24 | {} | NONE |
ClickHouse | ClickHouse | 535,365,160 | 6,984 | ghost | @zhang2014
Please tell me about anything what to do for merging this PR~ :) | 2019-09-26T06:50:47 | 2019-09-26T06:50:47 | {} | NONE |
ClickHouse | ClickHouse | 535,552,260 | 7,108 | zhang2014 | LGTM | 2019-09-26T15:15:27 | 2019-09-26T15:15:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 536,321,933 | 7,019 | geldot | Looks like one of these two Poco updates (#6386 and #6466) could be responsible. Didn't have a chance to build and test yet | 2019-09-29T17:15:51 | 2019-09-29T17:15:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 536,341,290 | 7,120 | alexey-milovidov | @amosbird There are no benefits of native centos build for production usage in comparison to our existing rpm packages. It isn't guaranteed that this build has passed all the tests in our CI.
It can be used if the customer need to use native centos build for whatever reason. | 2019-09-29T21:08:17 | 2019-09-29T21:08:17 | {} | MEMBER |
ClickHouse | ClickHouse | 535,503,969 | 7,100 | den-crane | for search:
Code: 171. DB::Exception: Received from localhost:9000. DB::Exception: Cannot convert column ... because it is constant but values of constants are different in source and result.
workaround for `now()` -> `materialize(now())`
Issues persists at least in 19.14.6.12 stable.
| 2019-09-26T13:33:40 | 2019-09-26T13:33:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 535,506,418 | 7,113 | akuzm | Looks like a bug, it would help if we could reproduce it locally. Is it possible to share the table definition and a sample of data for which the restart reproduces? | 2019-09-26T13:39:45 | 2019-09-26T13:39:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 535,640,217 | 7,109 | alexey-milovidov | Let's make the names consistent with our existing groupBitAnd, groupBitOr, groupBitXor, groupArray functions (use "group" instead of "agg"):
groupBitmapAnd, groupBitmapOr, groupBitmapXor. | 2019-09-26T18:56:32 | 2019-09-26T18:56:32 | {} | MEMBER |
ClickHouse | ClickHouse | 535,691,828 | 7,120 | filimonov | >
> I have a fork of https://github.com/Altinity/clickhouse-rpm somewhere, I just need to push my latest version and will add a link in the comments
sure. That's one is also based on Altinity (and i'm actually from Altinity :) ) | 2019-09-26T21:24:07 | 2019-09-26T21:24:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 535,922,106 | 6,661 | arruzk | @romankos it doesn't work(
Still the same ~50% of duplicates (`41549936`) | 2019-09-27T12:40:07 | 2019-09-27T12:40:07 | {} | NONE |
ClickHouse | ClickHouse | 535,970,439 | 6,661 | romankos | > @romankos it doesn't work(
> Still the same ~50% of duplicates (`41549936`)
This solution prevent only new duplicates.
For already exists duplicates you can use DISTINCT in queries or remove duplicated records via "alter table delete" | 2019-09-27T14:45:41 | 2019-09-27T14:45:41 | {} | NONE |
ClickHouse | ClickHouse | 536,031,419 | 7,111 | alesapin | mutations tests flap :( | 2019-09-27T17:34:29 | 2019-09-27T17:34:29 | {} | MEMBER |
ClickHouse | ClickHouse | 536,462,869 | 7,147 | Tasselmi | > > BUT what -1 mean ?
>
> `-1` means that last element of array (`2`) minus previous one (`3`) equals `2 - 3 = -1`
>
> Isn't it the expected behaviour?
>
> > array(T) takes T type, but a function can not support String type......
>
> Does strings support substraction? What is the result of 'Hello' minus 'w... | 2019-09-30T08:39:56 | 2019-09-30T08:39:56 | {} | NONE |
ClickHouse | ClickHouse | 535,636,162 | 3,834 | alexey-milovidov | Now works as expected. | 2019-09-26T18:45:53 | 2019-09-26T18:45:53 | {} | MEMBER |
ClickHouse | ClickHouse | 535,895,912 | 6,661 | arruzk | It's still reproducing in version `19.14.6.12`. I inserted data directly to the table with `Kafka` engine and got almost 50% of duplicates
```
select count(*), countDistinct(*)
41549936 19716621
``` | 2019-09-27T11:08:35 | 2019-09-27T11:08:35 | {} | NONE |
ClickHouse | ClickHouse | 536,272,549 | 6,880 | andyyzh | nice! | 2019-09-29T09:07:33 | 2019-09-29T09:07:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 536,465,499 | 7,120 | alexey-milovidov | We are not using `alien` directly, we also apply some modifications to the result and repack it. | 2019-09-30T08:47:12 | 2019-09-30T08:47:12 | {} | MEMBER |
ClickHouse | ClickHouse | 534,927,121 | 6,938 | yuzhichang | `1602daba5937311a5d70a45dbae78435da9a8200` back-ported #6911 to 19.14. | 2019-09-25T09:05:20 | 2019-09-25T09:05:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 535,019,516 | 7,096 | abyss7 | I insist on supporting C++ streaming (maybe alongside with formatted strings), since it's much simpler to implement global `operator<<` for any custom structures and print them in a simple manner. | 2019-09-25T13:22:21 | 2019-09-25T13:22:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 535,765,182 | 7,121 | ashrith | @den-crane thank you! Could this be a part of the documentation so that others who might be seeking this information is easily available? | 2019-09-27T02:56:47 | 2019-09-27T02:56:47 | {} | NONE |
ClickHouse | ClickHouse | 536,461,776 | 7,147 | filimonov | > BUT what -1 mean ?
`-1` means that last element of array (`2`) minus previous one (`3`) equals `2 - 3 = -1`
Isn't it the expected behaviour?
> array(T) takes T type, but a function can not support String type......
Does strings support substraction? What is the result of 'Hello' minus 'world!' ? | 2019-09-30T08:36:44 | 2019-09-30T08:36:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 534,980,363 | 7,038 | 4ertus2 | It's regression but not a bug. USING syntax is not expected to support qualified names. The same behaviour is e.g. in MySQL:
```
mysql> SELECT max(a) FROM (SELECT 1 AS a) AS x JOIN (SELECT 1 AS a) AS y USING(a);
+--------+
| max(a) |
+--------+
| 1 |
+--------+
1 row in set (0.00 sec)
mysql> SELECT ma... | 2019-09-25T11:36:10 | 2019-09-25T11:36:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 535,022,260 | 7,096 | akuzm | > I insist on supporting C++ streaming (maybe alongside with formatted strings), since it's much simpler to implement global `operator<<` for any custom structures and print them in a simple manner.
Printing custom structures with printf is not different, only you implement toString instead of operator<<. Performanc... | 2019-09-25T13:28:13 | 2019-09-25T13:28:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 535,093,707 | 7,073 | filimonov | More: https://en.opensuse.org/openSUSE:Upgrade_dependencies_explanation | 2019-09-25T16:04:39 | 2019-09-25T16:04:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 535,159,496 | 7,073 | alexey-milovidov | > While config is rewritten with default one...
Not good but acceptable because configuration changes made via config.d remain. | 2019-09-25T18:47:04 | 2019-09-25T18:47:04 | {} | MEMBER |
ClickHouse | ClickHouse | 535,570,245 | 7,113 | valenciaj | Yes of course, tell me where I sent the information in a private way. | 2019-09-26T15:56:36 | 2019-09-26T15:56:36 | {} | NONE |
ClickHouse | ClickHouse | 535,702,782 | 3,268 | den-crane | BTW I always used a workaround.
One more cluster with all nodes as dedicated shards (internal_replication=false).
So for this cluster any DDL just works at all nodes. | 2019-09-26T22:01:52 | 2019-09-26T22:06:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 536,070,366 | 7,117 | alexey-milovidov | Would it help if you write
`DSN={ServerName\InstanceName};UID=username;PWD=pwd`
(name in curly braces)? | 2019-09-27T19:35:08 | 2019-09-27T19:35:17 | {} | MEMBER |
ClickHouse | ClickHouse | 534,919,778 | 4,625 | andrewrbe | My logs:
https://gist.github.com/andrewrbe/42bbe8792b5450957456a6aa486d33c5 | 2019-09-25T08:46:57 | 2019-09-25T08:46:57 | {} | NONE |
ClickHouse | ClickHouse | 535,030,860 | 7,096 | abyss7 | I already made a prototype one day, but removed it later.
The API I suggest is as follows.
Implement some base tempate-class that provides all logging facilities and derive all other classes from it.
The syntax should be like:
```
LOG(WARN) << "some warning: " << additional_info;
LOG(INFO) << "useful form... | 2019-09-25T13:46:03 | 2019-09-25T13:46:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 535,035,332 | 7,096 | alexey-milovidov | > The syntax should be like
It doesn't support lazy evaluation of arguments (no evaluation when logging is disabled). | 2019-09-25T13:55:48 | 2019-09-25T13:55:48 | {} | MEMBER |
ClickHouse | ClickHouse | 535,036,968 | 7,096 | amosbird | P.S. C++20 will have std::format https://fmt.dev/Text%20Formatting.html | 2019-09-25T13:59:21 | 2019-09-25T13:59:21 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 534,967,998 | 7,088 | filimonov | Also it has a lower memory usage than hashed/complex_key_hashed, without significant performance impact. | 2019-09-25T10:57:40 | 2019-09-26T04:18:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 535,551,055 | 7,105 | zhang2014 | Can you put this pr to #7106 and close this pr? Because they are all modifying a same file.
| 2019-09-26T15:12:50 | 2019-09-26T15:12:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 551,027,335 | 7,667 | Slach | Please read https://clickhouse.yandex/docs/en/operations/table_engines/replication/#creating-replicated-tables
you shall manage databases and tables schema separately on each ClickHouse server
you can use ON CLUSTER statement described here https://clickhouse.yandex/docs/en/query_language/create/#distributed-ddl-... | 2019-11-07T10:50:06 | 2019-11-07T10:50:06 | {} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.