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 | 336,249,125 | 1,312 | alexey-milovidov | I still have some doubts about this code. | 2017-10-12T19:59:58 | 2017-10-12T19:59:58 | {} | MEMBER |
ClickHouse | ClickHouse | 336,252,673 | 1,314 | alexey-milovidov | Cool! Now we support multidimensional arrays in MergeTree. | 2017-10-12T20:13:53 | 2017-10-12T20:13:53 | {} | MEMBER |
ClickHouse | ClickHouse | 336,436,606 | 1,344 | KochetovNicolai | To improve performance, you can add limits on primary key. For example, use
`SELECT * from logs where date = today()`
or
`SELECT * from logs where dateTime > now() - 60 * 60` | 2017-10-13T12:17:31 | 2017-10-13T12:17:31 | {} | MEMBER |
ClickHouse | ClickHouse | 336,565,362 | 1,331 | vavrusa | examples should be disabled here https://github.com/yandex/ClickHouse/pull/1331/files#diff-0dd08584c0e6ad192dd81ad6ee8f62bfR65
@alexey-milovidov Should I open a new PR to address comments/improvements? | 2017-10-13T20:56:48 | 2017-10-13T20:56:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 334,776,801 | 1,094 | fleapapa | here you go.
```
$ make VERBOSE=1
/usr/bin/cmake -H/mnt/ClickHouse -B/mnt/ClickHouse/build --check-build-system CMakeFiles/Makefile.cmake 0
/usr/bin/cmake -E cmake_progress_start /mnt/ClickHouse/build/CMakeFiles /mnt/ClickHouse/build/CMakeFiles/progress.marks
make -f CMakeFiles/Makefile2 all
make[1]: Entering d... | 2017-10-06T14:48:34 | 2017-10-06T14:48:34 | {} | NONE |
ClickHouse | ClickHouse | 335,765,924 | 1,327 | ludv1x | autotests | 2017-10-11T10:24:21 | 2017-10-11T10:24:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 336,383,116 | 1,344 | gsmetal | If it's possible in clichouse internals to just get last N elements without full scan, maybe there shoud be special construction like `LIMIT -10` or something like that? It always possible to make count first and then do `LIMIT (COUNT - N), N`, but this realisation seems a little weird. | 2017-10-13T08:15:08 | 2017-10-13T08:15:08 | {} | NONE |
ClickHouse | ClickHouse | 336,588,183 | 1,263 | alexey-milovidov | Hope this will help: https://github.com/yandex/ClickHouse/commit/10392d54523b8eb2f8892f52220f350fb1f61096 | 2017-10-13T23:13:08 | 2017-10-13T23:13:08 | {} | MEMBER |
ClickHouse | ClickHouse | 336,605,406 | 1,356 | alexey-milovidov | It's intended that ClickHouse is build with latest release version of clang (currently it is 5) or even trunk version. 3.8 is outdated, it produce suboptimal code, and I think it is ok to obsolete it. | 2017-10-14T03:04:03 | 2017-10-14T03:04:03 | {} | MEMBER |
ClickHouse | ClickHouse | 336,606,104 | 1,356 | vavrusa | The latest clang shipped with Xcode is 4.2.1 Compatible Apple LLVM 9.0.0 (clang-900.0.37). I use it just for development, the production is built with gcc, so it's okay if you want to drop it (it'd probably fix the `std::experimental::option` only though. | 2017-10-14T03:20:51 | 2017-10-14T03:20:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 336,610,714 | 1,357 | vavrusa | Ugh, you're right. Not sure how to enable `TypeName<T>` specializations only if the types are different though 😦 | 2017-10-14T05:08:06 | 2017-10-14T05:08:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 336,797,083 | 777 | Slach | I checked latest version Docker image
all working fine
when you can use latest zetcd and latest clickhouse together ;) thank you guys for your great job | 2017-10-16T06:59:45 | 2017-10-16T06:59:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 337,128,705 | 1,361 | jaykelin | CLICKHOUSE VERSION:
>┌─version()─┐
>│ 1.1.54245 │
>└───────────┘
CREATE TABLE:
```sql
create table dal_oth_local.oa_copyright_music_scid_y on cluster kg_bi_cluster (dt_year Date, scid String, plat_state AggregateFunction(groupArray, String), user_type_state AggregateFunction(groupArray, String), is_songpay_stat... | 2017-10-17T06:14:35 | 2017-10-17T07:05:43 | {} | NONE |
ClickHouse | ClickHouse | 334,779,784 | 1,094 | proller | Looks like you have problem in your /usr/lib/x86_64-linux-gnu/libxml2.so.2
try reinstall libxml2 and cmake | 2017-10-06T14:58:35 | 2017-10-06T14:58:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 336,539,784 | 1,263 | alexey-milovidov | contrib/libpoco was not deleted. | 2017-10-13T19:00:00 | 2017-10-13T19:00:00 | {} | MEMBER |
ClickHouse | ClickHouse | 337,025,234 | 1,362 | silviucpp | From documentation doesn't seems to be a bug but it's not a consistent behavior.
What I mean is the default for Nullable(Float64) fields is NULL, for Float64 is 0 and for Array(Nullable(Float64)) it's also 0 not NULL as you are expecting.
| 2017-10-16T20:16:32 | 2017-10-16T20:16:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 337,402,401 | 1,330 | vavrusa | @ludv1x I ran the mentioned test and fixes several issues here https://github.com/yandex/ClickHouse/pull/1367 it'd be great to be able to run these via CI or something automatically, are you working on that? | 2017-10-17T23:04:01 | 2017-10-17T23:04:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 337,535,833 | 1,342 | Ultraman95 | +1 I hope Support Parquet | 2017-10-18T10:05:21 | 2017-10-18T10:05:21 | {} | NONE |
ClickHouse | ClickHouse | 333,853,227 | 1,313 | alexey-milovidov | The fix is unclear. | 2017-10-03T14:07:55 | 2017-10-03T14:07:55 | {} | MEMBER |
ClickHouse | ClickHouse | 334,732,474 | 1,094 | proller | please show cmake console output | 2017-10-06T11:45:41 | 2017-10-06T11:45:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 335,340,578 | 1,272 | jackpgao | ### As ClickHouse is multer-master replication, the best arch should be as below:
 | 2017-10-10T02:24:23 | 2017-10-10T02:24:23 | {
"+1": 8
} | NONE |
ClickHouse | ClickHouse | 336,231,382 | 1,344 | KochetovNicolai | Hello.
ClickHouse don't have optimization which allows skipping rows while reading first (or last) N rows ordered by primary key. So, results you see are expected.
Even when you need to read first N rows, ordered by primary key, it's necessary to merge data from different parts, which is full scan if your query d... | 2017-10-12T18:50:34 | 2017-10-12T18:50:34 | {} | MEMBER |
ClickHouse | ClickHouse | 336,256,649 | 1,344 | dmgtn | Hi!
If you store (structured) logs, it's a frequent need to see "last logs". In this scenario, table primary key is timestamp (or, timestamp + nanoseconds, as shown above).
Time-to-time it could be "last logs" without any filter, or filtering by couple columns. But it always *last logs*. By last I mean last 500-... | 2017-10-12T20:29:22 | 2017-10-12T20:29:22 | {
"+1": 2
} | NONE |
ClickHouse | ClickHouse | 336,441,825 | 1,335 | alexey-milovidov | Thanks! | 2017-10-13T12:42:43 | 2017-10-13T12:42:43 | {} | MEMBER |
ClickHouse | ClickHouse | 337,133,096 | 1,361 | jaykelin | clickhouse cluster: 6hosts
[clickhouse_settings.txt](https://github.com/yandex/ClickHouse/files/1390014/clickhouse_settings.txt)
[config.txt](https://github.com/yandex/ClickHouse/files/1390021/config.txt)
[user.txt](https://github.com/yandex/ClickHouse/files/1390022/user.txt)
| 2017-10-17T06:39:48 | 2017-10-17T06:39:48 | {} | NONE |
ClickHouse | ClickHouse | 337,145,252 | 1,291 | ml31415 | Oh, I guess that fixes it. Thanks! | 2017-10-17T07:39:12 | 2017-10-17T07:39:12 | {} | NONE |
ClickHouse | ClickHouse | 333,946,430 | 1,317 | alexey-milovidov | Thanks! | 2017-10-03T19:06:43 | 2017-10-03T19:06:43 | {} | MEMBER |
ClickHouse | ClickHouse | 334,801,658 | 1,094 | fleapapa | Quote:
> btw, why didn't previous version of CH code base have the issue?
I got the reason. It's because before this time i've been using the following command to build CH:
```
rm -f ../clickhouse*.deb
./release
```
That command still works without any problem.
However, the following commands break ... | 2017-10-06T16:16:39 | 2017-10-06T16:16:39 | {} | NONE |
ClickHouse | ClickHouse | 335,040,657 | 1,326 | alexey-milovidov | Ok, but need to change name of function, update copyright and comment. | 2017-10-08T21:40:30 | 2017-10-08T21:40:30 | {} | MEMBER |
ClickHouse | ClickHouse | 336,227,134 | 393 | alexey-milovidov | This is a bug. Fixed in:
https://github.com/yandex/ClickHouse/commit/d468473d624f7d9f25d23c98bd3ceb9ca3a4cffd | 2017-10-12T18:34:56 | 2017-10-12T18:34:56 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 336,248,470 | 1,302 | alexey-milovidov | https://github.com/yandex/ClickHouse/commit/e57e79a5ffde8f5f4c89778f36fa886d9e0579c4 | 2017-10-12T19:57:24 | 2017-10-12T19:57:24 | {} | MEMBER |
ClickHouse | ClickHouse | 336,452,732 | 1,347 | KochetovNicolai | Try to increase `receive_timeout` and `send_timeout` options. By default it's 300 sec. | 2017-10-13T13:29:53 | 2017-10-13T13:29:53 | {} | MEMBER |
ClickHouse | ClickHouse | 336,568,598 | 1,263 | vavrusa | The vendored version breaks build with the latest OS X clang:
```
clickhouse/libs/libcommon/src/JSON.cpp:459:52: error: call to 'format' is ambiguous
throw JSONException("JSON: array index " + Poco::NumberFormatter::format(n) + " out of bounds.");
^~~~~~~... | 2017-10-13T21:11:45 | 2017-10-13T21:26:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 336,589,152 | 1,331 | vavrusa | It shouldn't be so difficult, I'll try to add some in the next PR. | 2017-10-13T23:21:08 | 2017-10-13T23:21:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 336,965,036 | 1,291 | jotes | @ml31415 did you see https://github.com/Infinidat/infi.clickhouse_fdw/ ? | 2017-10-16T17:40:06 | 2017-10-16T17:40:06 | {} | NONE |
ClickHouse | ClickHouse | 336,995,977 | 1,362 | silviucpp | @proller very strange. I'm on the testing branch:
deb http://repo.yandex.ru/clickhouse/trusty testing main
And my server doesn't show any update available. My server shows 1.1.54298 which is not last release (there are 2 others). I thought commits in master automatically generates updates available on this test... | 2017-10-16T18:57:53 | 2017-10-16T18:57:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 337,230,467 | 1,362 | proller | Yes look as bug.
Will be fixed.
in current master just fixed table corruption "bad size of marks file"
| 2017-10-17T13:25:57 | 2017-10-17T13:25:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 334,019,160 | 1,299 | vavrusa | @alexey-milovidov could you take a look? I have some other changes, but don't want to start separate PRs depending on this. | 2017-10-04T00:42:01 | 2017-10-04T00:42:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 335,087,700 | 1,283 | hamsterready | Thanks for this guys, I will upgrade as soon as I can.
Cheers
| 2017-10-09T07:57:27 | 2017-10-09T07:57:27 | {} | NONE |
ClickHouse | ClickHouse | 336,599,142 | 1,263 | vavrusa | https://github.com/yandex/ClickHouse/pull/1356 | 2017-10-14T01:11:19 | 2017-10-14T01:11:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 336,610,217 | 1,357 | alexey-milovidov | This is totally incorrect (and still don't compile).
Both `long` and `long long` are 64 bit on both Linux and Mac OS X,
so `POCO_LONG_IS_64_BIT` is true.
The difference is only when `int64_t` is `long` (Linux) or `long long` (Mac OS X). | 2017-10-14T04:57:15 | 2017-10-14T04:57:15 | {} | MEMBER |
ClickHouse | ClickHouse | 337,313,330 | 1,330 | vavrusa | I'll take a look! | 2017-10-17T17:52:19 | 2017-10-17T17:52:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 334,781,474 | 1,094 | fleapapa | i did, but it shows:
```
libxml2 is already the newest version (2.9.3+dfsg1-1ubuntu0.3).
libxml2-dev is already the newest version (2.9.3+dfsg1-1ubuntu0.3).
``` | 2017-10-06T15:04:25 | 2017-10-06T15:04:25 | {} | NONE |
ClickHouse | ClickHouse | 336,245,678 | 1,302 | alexey-milovidov | Test is Ok, but the fix is not. I will fix by myself. | 2017-10-12T19:46:32 | 2017-10-12T19:46:32 | {} | MEMBER |
ClickHouse | ClickHouse | 336,158,829 | 1,343 | KochetovNicolai | 1. Yes. Data blocks are written in /var/lib/clickhouse/data/database/table/ folder. Special thread checks directory periodically and tries to send data. If it can't, it will try next time.
2. Yes.
3. No. Data is stored in /var/lib/clickhouse/data/database/table/ until it won't be send.
Also, if you use ReplicatedM... | 2017-10-12T14:42:19 | 2017-10-12T14:42:19 | {} | MEMBER |
ClickHouse | ClickHouse | 336,373,306 | 1,342 | amosbird | +1. This would be really useful to migrate data from hadoop ecosystem to ch. | 2017-10-13T07:28:39 | 2017-10-13T07:28:39 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 336,564,930 | 1,331 | alexey-milovidov | examples in rdkafka library doesn't compile. | 2017-10-13T20:54:54 | 2017-10-13T20:54:54 | {} | MEMBER |
ClickHouse | ClickHouse | 336,572,924 | 1,331 | alexey-milovidov | > examples should be disabled here https://github.com/yandex/ClickHouse/pull/1331/files#diff-0dd08584c0e6ad192dd81ad6ee8f62bfR65
Actually, they are not disabled... | 2017-10-13T21:33:53 | 2017-10-13T21:33:53 | {} | MEMBER |
ClickHouse | ClickHouse | 336,619,914 | 29 | silviucpp | This task should be closed | 2017-10-14T08:28:24 | 2017-10-14T08:28:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 336,986,250 | 1,362 | proller | Fixed in master.
this tests
```
DROP TABLE IF EXISTS test.table_array4;
CREATE TABLE test.table_array4(`array_empty` Array(Nullable(Int64)), `partition` Date, `timestamp` DateTime) ENGINE = ReplicatedMergeTree('/var/lib/clickhouse/tables/{shard}/table_array4','{replica}',`partition`,`timestamp`,8192);
INSERT INTO ... | 2017-10-16T18:31:51 | 2017-10-16T18:31:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 337,631,841 | 812 | gnauq3mg6m64rs2n | 2017.10.18 18:26:27.964468 [ 5 ] <Error> void DB::BackgroundProcessingPool::threadFunction(): Code: 173, e.displayText() = DB::Exception: Allocator: Cannot mmap., errno: 12, strerror: Cannot allocate memory: (while reading column _value): (while reading from part /data/clickhouse//data/default/event_detect/20170207_201... | 2017-10-18T15:33:41 | 2017-10-18T15:33:41 | {
"-1": 1
} | NONE |
ClickHouse | ClickHouse | 333,946,549 | 1,316 | alexey-milovidov | :+1: | 2017-10-03T19:07:13 | 2017-10-03T19:07:13 | {} | MEMBER |
ClickHouse | ClickHouse | 335,527,619 | 1,331 | vavrusa | @filimonov That's sort of what we do right now, using https://github.com/kshvakov/clickhouse. The main problem with bridges is that it's another piece of architecture to write, maintain, monitor, and operate. It's also not trivial with growing throughput (and number of different bridges). Our current implementation has... | 2017-10-10T16:18:13 | 2017-10-10T16:18:13 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 336,589,662 | 1,263 | vavrusa | @alexey-milovidov It affects more places where `UInt64` and `size_t` were used interchangeably. Like:
```
In file included from /clickhouse/dbms/src/Parsers/IAST.cpp:4:
In file included from /clickhouse/dbms/src/IO/Operators.h:5:
/clickhouse/dbms/src/IO/ReadHelpers.h:112:5: error: no matching function for call to... | 2017-10-13T23:26:06 | 2017-10-13T23:26:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 336,799,226 | 777 | Slach | =(( oops sorry
when i first run test queries it was OK
but second vagrant provision failed with error
```
==> clickhouse_zetcd: INSERT INTO zetcd_test.test_replicated (timestamp, trackerId, userId ) VALUES (1, 'test1','test1')
==> clickhouse_zetcd: Code: 32. DB::Exception: Attempt to read after eof: while receivi... | 2017-10-16T07:11:02 | 2017-10-16T07:11:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 337,291,800 | 1,364 | proller | todo: look to this lib:
https://github.com/vstakhov/rmilter/blob/master/contrib/lc-btrie/btrie.c | 2017-10-17T16:40:36 | 2017-10-17T16:40:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 335,040,749 | 1,299 | alexey-milovidov | Sorry for the delay. Looking right now... | 2017-10-08T21:42:12 | 2017-10-08T21:42:12 | {} | MEMBER |
ClickHouse | ClickHouse | 335,058,637 | 1,311 | wangyinsheng | `SELECT count(*) FROM table group by dt` , dt is the date field, there is two date, the count of first date dosen't change | 2017-10-09T02:55:44 | 2017-10-09T02:55:44 | {} | NONE |
ClickHouse | ClickHouse | 335,360,706 | 1,330 | vavrusa | Current implementation when just summing numeric column:
```
$ ./dbms/src/DataStreams/tests/aggregating_stream 500000 sum
Elapsed 0.02 sec., 26839873.32 rows/sec.
SummingSorted
One × 2
499996 abc abc
499997 def abcd
499998 abcd def
499999 defg abc
```
Current implementation when adding a nested colum... | 2017-10-10T04:55:23 | 2017-10-12T04:21:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 336,619,826 | 23 | silviucpp | Maybe this task should be closed | 2017-10-14T08:26:58 | 2017-10-14T08:26:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 334,643,346 | 1,094 | fleapapa | I'm also experiencing this same issue. Is there any fix? | 2017-10-06T02:36:23 | 2017-10-06T02:36:23 | {} | NONE |
ClickHouse | ClickHouse | 334,775,620 | 1,094 | proller | and fail message with last command from
make VERBOSE=1
please | 2017-10-06T14:44:26 | 2017-10-06T14:44:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 335,937,714 | 1,340 | alexey-milovidov | Thanks! | 2017-10-11T20:28:06 | 2017-10-11T20:28:06 | {} | MEMBER |
ClickHouse | ClickHouse | 336,265,527 | 1,330 | vavrusa | I should proofread my comments, sorry 🙈 | 2017-10-12T21:00:17 | 2017-10-12T21:00:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 336,536,701 | 1,351 | alexey-milovidov | Ok. | 2017-10-13T18:47:08 | 2017-10-13T18:47:08 | {} | MEMBER |
ClickHouse | ClickHouse | 336,540,847 | 1,263 | alexey-milovidov | CMAKE_BUILD_TYPE was suddenly changed to Release.
```CMAKE_BUILD_TYPE:STRING=Release```
We have RelWithDebInfo as default.
Upd. After cleaning cache, it get fixed. | 2017-10-13T19:04:24 | 2017-10-13T19:11:15 | {} | MEMBER |
ClickHouse | ClickHouse | 336,976,523 | 1,361 | proller | Can you provide failing query, SHOW CREATE TABLE your_table and version of clickhouse? | 2017-10-16T18:07:22 | 2017-10-16T18:07:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 337,231,007 | 1,362 | silviucpp | @proller do you want to open a separate ticket for this bug with inconsistent default values ?
I see the one "bad size of marks file" it's already reported and tracked here : https://github.com/yandex/ClickHouse/issues/1282
I would like to close mine | 2017-10-17T13:27:45 | 2017-10-17T13:27:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 337,458,363 | 1,300 | filimonov | Please check also #1297 #703 (duplicates?) | 2017-10-18T04:38:50 | 2017-10-18T04:38:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 335,860,776 | 1,285 | silviucpp | Hello,
Any idea if this bug is duplicated with https://github.com/yandex/ClickHouse/issues/1322 ? | 2017-10-11T16:04:16 | 2017-10-11T16:04:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 335,882,201 | 1,336 | KochetovNicolai | Merge таблица делает одновременно один и тот же запрос к нескольким таблицам, а затем склеивает результат. При этом, например, `group by N1` будет работать по-разному. Если такое поведение не является проблемой, то merge таблица должна работать быстрее по крайней мере для запросов с ограничением на host. | 2017-10-11T17:16:34 | 2017-10-11T17:16:34 | {} | MEMBER |
ClickHouse | ClickHouse | 336,091,582 | 393 | gudmojo | I'm experiencing the same issue on 1.1.54245
Sometimes it results in partial data loss in the backup as I described here https://serverfault.com/questions/876648/clickhouse-inconsistent-backup-restore-for-latest-partition-when-using-alter-ta | 2017-10-12T10:44:51 | 2017-10-12T10:44:51 | {} | NONE |
ClickHouse | ClickHouse | 336,233,667 | 1,285 | Lakret | The same behaviour happens when trying to concat results of `toString` operations on nullable `Int32` columns. For table with the following structure:
```
CREATE TABLE test.save_19220892784955 ( Year Nullable(Int32), Month Nullable(Int32), DayofMonth Nullable(Int32), DayOfWeek Nullable(Int32), DepTime Nullable(... | 2017-10-12T18:59:10 | 2017-10-12T18:59:10 | {} | NONE |
ClickHouse | ClickHouse | 336,587,693 | 1,331 | alexey-milovidov | How difficult is to install Kafka for integration tests?
We have `tests/integration`. If it is possible and not over-complicated, better to add a test. | 2017-10-13T23:08:45 | 2017-10-13T23:08:45 | {} | MEMBER |
ClickHouse | ClickHouse | 336,609,890 | 1,356 | vavrusa | Ok, I tried to write conditional support for `FieldVisitorConvertToNumber` instead and fix `readVarUInt`: https://github.com/yandex/ClickHouse/pull/1357 | 2017-10-14T04:48:57 | 2017-10-14T04:48:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 337,019,528 | 1,362 | silviucpp | @proller any idea if the fact that for a column of type : Array(Nullable(Float64)) the default value is considered 0 and not NULL it's expected behavior ?
For example if you do:
```sql
SELECT array_field[4] from table
```
where that element on that index is not existing you receive 0 instead NULL. Is this co... | 2017-10-16T20:00:33 | 2017-10-16T20:00:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 337,191,758 | 822 | dorantor | BTW, what would be an optimal way to convert/alter table from UUID field(FixedString(16)) as sampling field, to something like `sipHash64(cid)`(or maybe there is a better option?), where `cid` defined as `cid FixedString(16)` and contains UUIDs in binary form from `UUIDStringToNum()`.
I mean, data in table I would lik... | 2017-10-17T10:41:52 | 2017-10-17T10:41:52 | {} | NONE |
ClickHouse | ClickHouse | 334,796,467 | 1,094 | fleapapa | Thanks. Ran the command and cmake again, but make still has the same error. | 2017-10-06T15:57:02 | 2017-10-06T15:57:02 | {} | NONE |
ClickHouse | ClickHouse | 335,886,498 | 1,336 | slw | а в чем будет выражаться разная работа group by N1?
теоретически такие запросы могут быть, но при этом (на данный момент) либо данные N1 будут уникальны промеж разных хостов, либо либо все равно будет подразумеваться их уникальность, например через объединение с хостом. в любом случае такие значения будут включены в и... | 2017-10-11T17:30:47 | 2017-10-11T17:30:47 | {} | NONE |
ClickHouse | ClickHouse | 336,275,964 | 1,335 | alexey-milovidov | Ok. Does it really works when name of Nested column is specified?
You can add a test (look at `dbms/tests`, `dbms/tests/queries` directory). | 2017-10-12T21:25:00 | 2017-10-12T21:25:00 | {} | MEMBER |
ClickHouse | ClickHouse | 336,371,998 | 1,335 | amosbird | Yeah, it works. I've added some tests :) | 2017-10-13T07:22:04 | 2017-10-13T07:22:04 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 336,383,902 | 1,344 | gsmetal | Just tried, this limit-with-offset solution also does not work, it does full scan:
```
:) select count() from logs;
SELECT count()
FROM logs
┌────count()─┐
│ 4277567828 │
└────────────┘
1 rows in set. Elapsed: 1.362 sec. Processed 4.28 billion rows, 17.11 GB (3.14 billion rows/s., 12.56 GB/s.)
:) sel... | 2017-10-13T08:18:38 | 2017-10-13T08:18:38 | {} | NONE |
ClickHouse | ClickHouse | 336,614,877 | 1,357 | vavrusa | I couldn't figure out how to do this nicely with SFINAE, so I partially reverted this and used supported fixed-size types instead. I changed the return values as well to address https://github.com/yandex/ClickHouse/pull/1356#discussion_r144685939 | 2017-10-14T06:44:01 | 2017-10-14T06:44:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 337,003,147 | 1,362 | proller | This fix not released.
You can build unreleased code manually, but at current moment latests master have some bugs. | 2017-10-16T19:16:08 | 2017-10-16T19:16:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 337,223,160 | 1,291 | ml31415 | Being based on multicorn, unfortunately it doesn't seem to be a speedy option :/ | 2017-10-17T12:58:48 | 2017-10-17T12:58:48 | {} | NONE |
ClickHouse | ClickHouse | 337,237,486 | 1,330 | ludv1x | @vavrusa It looks like that your changes broke the following tests:
```
00148_summing_merge_tree_nested_map_multiple_values: [ FAIL ] - return code 9
Received exception from server:
Code: 9. DB::Exception: Received from localhost:9000, ::1. DB::Exception: Sizes of columns doesn't match: k: 4, payl... | 2017-10-17T13:49:31 | 2017-10-17T13:49:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 333,368,196 | 1,310 | alexey-milovidov | On newer OS, `cmake3` package is named as just `cmake`. You can correct the script.
Clang is optional and used only when you enable `compile` setting. Look at https://clickhouse.yandex/docs/en/single/#compile | 2017-10-01T10:46:28 | 2017-10-01T10:46:28 | {} | MEMBER |
ClickHouse | ClickHouse | 334,648,124 | 1,094 | fleapapa | This issue seems from incomplete build of libz.a, which misses gzdxxxx (xxx = open/close/...).
Accidentally found that somehow, on line #511 of file contrib/libzlib-ng/CMakeLists.txt, by changing the tailing 'z)' to anything else such like 'xyz)' will bypass the linker error. Frankly, i don't know the reason why. I ... | 2017-10-06T03:20:02 | 2017-10-06T03:23:18 | {} | NONE |
ClickHouse | ClickHouse | 334,768,143 | 1,094 | fleapapa | here you go.
```
$ cmake ..
-- CMAKE_BUILD_TYPE is not set, set to default = RELWITHDEBINFO
-- CMAKE_BUILD_TYPE: RELWITHDEBINFO
-- Building for: Linux-4.10.0-35-generic x86_64 x86_64-linux-gnu ; USE_STATIC_LIBRARIES=ON MAKE_STATIC_LIBRARIES=ON UNBUNDLED=OFF CCACHE=/usr/bin/ccache 3.2.4
-- Using openssl=TRUE: /u... | 2017-10-06T14:18:10 | 2017-10-06T14:18:10 | {} | NONE |
ClickHouse | ClickHouse | 334,781,867 | 1,094 | fleapapa | btw, why didn't previous version of CH code base have the issue? | 2017-10-06T15:05:45 | 2017-10-06T15:05:45 | {} | NONE |
ClickHouse | ClickHouse | 335,077,279 | 1,290 | kislenok-roman | Just updated database and recheck -- everything seems fine. | 2017-10-09T06:51:14 | 2017-10-09T06:51:14 | {} | NONE |
ClickHouse | ClickHouse | 335,395,616 | 1,331 | filimonov | Sounds interesting. But why engine, not standalone ETL-like configurable brigde kafka->clickhouse?
Something that will extract data from Kafka (with librdkafka for example), transform it as needed (apply schema, deserialize etc), and load it into clickhouse (using clickhouse-cpp for ex.)? | 2017-10-10T08:12:02 | 2017-10-10T08:12:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 336,017,331 | 1,330 | vavrusa | Implemented more `sumMap` a bit more efficiently, and added support for single key and multiple arrays to sum. New implementation:
```
$ ./dbms/src/DataStreams/tests/aggregating_stream 500000 sum nested
Elapsed 0.08 sec., 6263231.08 rows/sec.
SummingSorted
One × 2
499996 abc abc [0,1,2,3,4] [75001,74999,75... | 2017-10-12T04:27:01 | 2017-10-12T04:27:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 336,572,845 | 1,331 | alexey-milovidov | > @alexey-milovidov Should I open a new PR to address comments/improvements?
Yes. | 2017-10-13T21:33:25 | 2017-10-13T21:33:25 | {} | MEMBER |
ClickHouse | ClickHouse | 336,885,736 | 1,360 | proller | show create table default.net;
and maybe 1-10-... lines of sample data | 2017-10-16T13:25:53 | 2017-10-16T13:25:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 337,478,475 | 822 | dorantor | In case somebody else will face same problem sharing my solution. There is no direct way to change MergeTree() options. You have to CREATE new table, INSERT SELECT data from old table, DROP old one and RENAME newly created.
I was lucky, I had only 800mil records and I could stop data flow into my Clickhouse for about... | 2017-10-18T06:59:10 | 2017-10-18T06:59:38 | {} | NONE |
ClickHouse | ClickHouse | 333,377,844 | 1,310 | fleapapa | Hi @alexey-milovidov. Thanks for the pointer. | 2017-10-01T13:48:30 | 2017-10-01T13:48:30 | {} | NONE |
ClickHouse | ClickHouse | 334,155,180 | 1,319 | alexey-milovidov | Great, thanks! | 2017-10-04T13:29:26 | 2017-10-04T13:29:26 | {} | MEMBER |
ClickHouse | ClickHouse | 334,790,877 | 1,094 | proller | `sudo apt install --reinstall libxml2 libxml2-dev cmake` | 2017-10-06T15:36:53 | 2017-10-06T15:37:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 335,895,581 | 1,336 | KochetovNicolai | Я говорил о ситуации, когда, например, одно и то же значение N1 встретилось в разных таблицах, а в запросе group by только по N1. Тогда результат Merge вернет несколько строчек с одинаковыми N1, а общая таблица - только одну. Кажется, что в вашем случае это не проблема.
Также, могут возникнуть проблемы, если создать... | 2017-10-11T17:57:36 | 2017-10-11T17:57:36 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.