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 | 331,137,402 | 1,270 | filimonov | Views - are acceptable, but custom query still seems to be better solution.
When using views:
* it breaks single responsibility principle: you need to configure same stuff in 2 places - in view definition, and in dict config file
* more bureaucracy: you need to change DB schema - which can be under version contro... | 2017-09-21T12:09:48 | 2017-09-21T12:09:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 331,706,462 | 1,280 | bookin | I think it is problem for Ukrainian. I will try to use other vpn and say about result | 2017-09-24T12:20:06 | 2017-09-24T12:20:06 | {} | NONE |
ClickHouse | ClickHouse | 332,539,585 | 1,297 | filimonov | Sorry, duplicate https://github.com/yandex/ClickHouse/issues/703
| 2017-09-27T14:28:25 | 2017-09-27T14:28:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 331,248,311 | 1,257 | vavrusa | We just had a similar issue (not sure it's the same thing though), several hosts segfaulted at the same time. Lots of:
```
<Error> BaseDaemon: (from thread 16) Received signal Segmentation fault (11).
<Error> BaseDaemon: Address: NULL pointer.
<Error> BaseDaemon: 0. /usr/bin/clickhouse-server(tcmalloc::ThreadCach... | 2017-09-21T18:52:04 | 2017-09-21T18:52:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 331,706,615 | 1,280 | alexey-milovidov | For documentation, you may use this mirror: http://clickhouse-docs.readthedocs.io/en/latest/index.html
(probably outdated)
Or read documentation directly by browsing GitHub:
https://github.com/yandex/ClickHouse/tree/master/docs/ru | 2017-09-24T12:22:07 | 2017-09-24T12:22:07 | {} | MEMBER |
ClickHouse | ClickHouse | 331,809,235 | 1,261 | wangyinsheng | thanks a lot!
well, i tried to execute `resharding` to move some data to the new servers , then there is an error `This host is not allowed to subscribe to coordinator......` , it something wrong? | 2017-09-25T08:16:39 | 2017-09-25T08:16:39 | {} | NONE |
ClickHouse | ClickHouse | 331,984,467 | 1,225 | filimonov | In your sample : try this (if your default tz is UTC)
```
select count(*) as count, toRelativeDayNum(date+3*3600) - toRelativeDayNum(X+3*3600) from table
when date >= X and date =< Y
``` | 2017-09-25T19:18:46 | 2017-09-25T19:20:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 332,538,311 | 1,297 | filimonov | Works the same on 1.1.54289. (Centos, @Altinity build) | 2017-09-27T14:24:31 | 2017-09-27T14:24:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 330,769,572 | 198 | remenska | +1, getting the same error | 2017-09-20T07:32:28 | 2017-09-20T07:32:28 | {} | NONE |
ClickHouse | ClickHouse | 330,830,071 | 1,265 | ludv1x | Hi!
You could use lambdas:
```
SELECT arrayMap((a, b) -> (a - b), [5, 6, 7], [1, 2, 3])
┌─arrayMap(lambda(tuple(a, b), minus(a, b)), [5, 6, 7], [1, 2, 3])─┐
│ [4,4,4] │
└──────────────────────────────────────────────────────────────────┘
1 rows in set.... | 2017-09-20T12:00:46 | 2017-09-20T12:00:46 | {
"+1": 3
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 331,939,476 | 1,225 | silviucpp | I think my observations or mostly for toRelative[*]Num functions where your example is no longer working. | 2017-09-25T16:38:08 | 2017-09-25T16:38:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 331,986,090 | 1,225 | silviucpp | @filimonov yes very interesting your suggestion :) this way you can bypass also the issue with fractional tz that's missing. do I miss something ? | 2017-09-25T19:25:32 | 2017-09-25T19:25:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 332,003,032 | 1,283 | vavrusa | Does it work on regular columns (not nullable)? | 2017-09-25T20:31:17 | 2017-09-25T20:31:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 332,114,047 | 1,283 | hamsterready | ```
:) select version()
SELECT version()
┌─version()─┐
│ 1.1.54284 │
└───────────┘
1 rows in set. Elapsed: 0.019 sec.
```
| 2017-09-26T07:43:02 | 2017-09-26T07:43:02 | {} | NONE |
ClickHouse | ClickHouse | 332,572,021 | 622 | filimonov | Issue here: after droping column from partition system.columns data about colums size is not refreshed.
And even if you will drop whole partition later sizes of column droped this way don't get zeros. | 2017-09-27T16:05:23 | 2017-09-27T16:08:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 332,801,926 | 127 | VKoptev | +1 | 2017-09-28T10:57:58 | 2017-09-28T10:57:58 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 332,948,839 | 1,303 | KochetovNicolai | Попытался ответить в [гугл-группе](https://groups.google.com/forum/#!topic/clickhouse/cBgJgFrw4cg). | 2017-09-28T20:05:44 | 2017-09-28T20:05:44 | {} | MEMBER |
ClickHouse | ClickHouse | 333,199,911 | 1,307 | fleapapa | Hi @ztlpn,
Sure. Here you go.
```
create table trip (
_cx Date,
_c0 String,
_c1 String,
_c2 String,
_c3 String,
_c4 String,
_c5 DateTime,
_c6 DateTime,
_c7 Int8,
_c8 UInt32,
_c9 ... | 2017-09-29T18:15:54 | 2017-09-29T18:15:54 | {} | NONE |
ClickHouse | ClickHouse | 330,787,290 | 198 | DeamonMV | As i remember Alexey have wrote that it maybe will be resolved when team end development of different partition - like day, hour, week, instead only month - and after this maybe we get resharding | 2017-09-20T08:47:18 | 2017-09-20T08:47:18 | {} | NONE |
ClickHouse | ClickHouse | 330,916,156 | 198 | alexey-milovidov | Right now, resharding is frozen and abandoned. We have a chance to continue development after a while. | 2017-09-20T17:01:39 | 2017-09-20T17:01:39 | {} | MEMBER |
ClickHouse | ClickHouse | 331,706,656 | 1,280 | bookin | Thanks for help, it will be better | 2017-09-24T12:22:47 | 2017-09-24T12:22:47 | {} | NONE |
ClickHouse | ClickHouse | 331,949,245 | 1,281 | KochetovNicolai | Now I can't reproduce the problem:
```
ClickHouse client version 1.1.54292.
Connecting to localhost:9004.
Connected to ClickHouse server version 1.1.54292.
:) SELECT floor(NULL) format JSONEachRow;
SELECT floor(NULL)
FORMAT JSONEachRow
{"floor(NULL)":null}
1 rows in set. Elapsed: 0.004 sec.
:)
```... | 2017-09-25T17:13:25 | 2017-09-25T17:13:25 | {} | MEMBER |
ClickHouse | ClickHouse | 332,187,390 | 1,290 | KochetovNicolai | Hi,
Which ClickHouse version do you use?
Can't reproduce for 1.1.54289 and 1.1.54292 | 2017-09-26T12:49:22 | 2017-09-26T12:49:22 | {} | MEMBER |
ClickHouse | ClickHouse | 332,540,325 | 703 | filimonov | Still here, see https://github.com/yandex/ClickHouse/issues/1297 | 2017-09-27T14:30:36 | 2017-09-27T14:30:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 333,304,186 | 1,311 | ztlpn | H! How exactly are you counting rows? With `SELECT count(*) FROM table`? | 2017-09-30T12:04:28 | 2017-09-30T12:04:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 330,916,641 | 1,231 | alexey-milovidov | Yes, the function `greatCircleDistance` is in production state. It is still missing in official documentation. | 2017-09-20T17:03:35 | 2017-09-20T17:03:35 | {} | MEMBER |
ClickHouse | ClickHouse | 331,580,707 | 1,277 | ztlpn | The directory name for temporary files is controlled by the `tmp_path` config parameter. It is `/var/lib/clickhouse/tmp/` by default which is inside the data directory. | 2017-09-22T23:12:02 | 2017-09-22T23:12:02 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 332,277,242 | 1,288 | alexey-milovidov | Where and why the exception is thrown inside the signal handler? | 2017-09-26T17:39:45 | 2017-09-26T17:39:45 | {} | MEMBER |
ClickHouse | ClickHouse | 332,487,350 | 1,283 | KochetovNicolai | What type does `ClientIpAddress` has and what type `topK(1)`does return?
```
select toTypeName(ClientIpAddress) from pageviews_dist limit 1;
select toTypeName(topK(1)(ClientIpAddress)) from pageviews_dist limit 1;
``` | 2017-09-27T11:08:00 | 2017-09-27T11:08:00 | {} | MEMBER |
ClickHouse | ClickHouse | 332,749,362 | 1,299 | vavrusa | ```sql
:) SELECT * FROM system.clusters
┌─cluster─┬─shard_num─┬─shard_weight─┬─replica_num─┬─host_name─┬─host_address─┬─port─┬─is_local─┬─user────┬─default_database─┐
│ logs │ 1 │ 1 │ 1 │ localhost │ ::1 │ 9000 │ 1 │ default │ │
│ logs │ ... | 2017-09-28T07:15:06 | 2017-09-28T07:16:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 333,209,635 | 1,307 | ztlpn | Well, that's not exactly small :) I'll try to reproduce, but maybe the problem manifests itself even with a few rows? Also, what are the `setc5` and `trip2` tables?
> A quick question. Does this type of table allow multiple rows with duplicate primary key as defined above?
MergeTree allows rows with the same prim... | 2017-09-29T18:55:18 | 2017-09-29T18:56:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 333,303,161 | 1,307 | ztlpn | If you haven't yet deleted the table, can you check the `<clickhouse data directory>/data/default/trip/detached/` directory? Are there any subdirectories there?
> Afterward i changed _c8 to Int32 and _c13 to nullable to ignore the errors.
How exactly did you change it? By creating a new table or using ALTER? | 2017-09-30T11:42:30 | 2017-09-30T12:09:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 331,966,963 | 1,287 | ludv1x | Hi!
Indeed, for the simplicity of the first distributed DDL version, I intentionally skipped implementation of the case when a server has several shards in different databases. I thought it is not the widespread case.
It is not complicated to add support of this case. | 2017-09-25T18:14:14 | 2017-09-25T18:14:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 332,291,715 | 197 | f1yegor | can we close it?) | 2017-09-26T18:28:21 | 2017-09-26T18:28:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 333,308,692 | 1,307 | fleapapa | Unfortunately i have deleted the table and reloaded it. This time i changed _c8 type to Int32 so the block insert API used in my python script wouldn't fail. For _c13 i simply made its value 0 if the source is an empty string to bypass python float() conversion error. Please see [the script](https://github.com/fleapapa... | 2017-09-30T13:30:00 | 2017-09-30T13:30:00 | {} | NONE |
ClickHouse | ClickHouse | 331,676,529 | 1,259 | silviucpp | Hello @alexey-milovidov ,
Any idea about this problem ?
| 2017-09-23T23:36:17 | 2017-09-23T23:36:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 331,840,507 | 1,278 | KochetovNicolai | Could you check that ClickHouse version is the same on remote and local server? | 2017-09-25T10:25:19 | 2017-09-25T10:25:19 | {} | MEMBER |
ClickHouse | ClickHouse | 331,944,780 | 1,283 | vavrusa | Are other distributed queries working (e.g. GROUP BY)? Does it work on regular columns (not nullable)? | 2017-09-25T16:57:02 | 2017-09-25T16:57:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 331,951,179 | 1,281 | silviucpp | I found the problem :)
Try using HTTP interface using POST (`SELECT floor(NULL) format JSONEachRow;`)
Server is crashing with above stack. If you use the clickhouse-client works nice
Silviu | 2017-09-25T17:20:40 | 2017-09-25T17:20:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 331,959,410 | 1,225 | filimonov | Sorry, i don't get it. Can you give some example? | 2017-09-25T17:47:55 | 2017-09-25T17:47:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 331,964,937 | 1,225 | silviucpp | Basically your example just convert a date into another timezone. This is simple to achieve. But in case you want to group your results let's say daily you have to use toRelativeDayNum(date).
In this case if you also want to take the timezone into account you can use toRelativeMonthNum(date, timezone) (A day into a re... | 2017-09-25T18:06:54 | 2017-09-25T18:06:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 331,981,683 | 1,283 | hamsterready | I can run all kids of queries:
```
ubuntu@backend-03:~$ clickhouse-client
ClickHouse client version 1.1.54284.
Connecting to localhost:9000.
Connected to ClickHouse server version 1.1.54284.
:) select count(1) from pageviews_dist where SessionId = UUIDStringToNum('6af5e8fa-2d9a-407c-b37f-fe562355077e')
SE... | 2017-09-25T19:07:26 | 2017-09-25T19:08:36 | {} | NONE |
ClickHouse | ClickHouse | 332,002,849 | 1,287 | vavrusa | Ah I see, supporting this use case would be wonderful 👌 | 2017-09-25T20:30:39 | 2017-09-25T20:30:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 332,807,733 | 1,301 | proller | Already fixed - https://github.com/yandex/ClickHouse/commit/c46ce859a183bcd89441c1a16462ada150f6ae8d | 2017-09-28T11:24:34 | 2017-09-28T11:24:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 332,809,668 | 1,301 | jbfavre | oh, didn't saw it.
Thanks ! | 2017-09-28T11:34:20 | 2017-09-28T11:34:20 | {} | NONE |
ClickHouse | ClickHouse | 331,523,231 | 1,275 | RoyBellingan | Hi Ludv1x, yes the table is a MergeTree
Here is the full output
CREATE TABLE rt_analytics.impressions ( request_nr UInt64, timestamp UInt32, milliseconds UInt32, ip String, user_timezone_offset Int64, nation UInt64, subregion_0 UInt64, subregion_1 UInt64, city UInt64, carrier String, connection_type Strin... | 2017-09-22T18:19:10 | 2017-09-22T18:19:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 331,864,199 | 1,281 | KochetovNicolai | Hello,
Could you say which query has caused the crashing? Was it
`SELECT greatCircleDistance(NULL, 55.3, 38.7, 55.1) AS distance format JSONEachRow;`
?
Also, which version of ClickHouse do you use? In last both
`SELECT greatCircleDistance(NULL, 55.3, 38.7, 55.1) AS distance format JSONEachRow;`
and
`SELECT ... | 2017-09-25T12:22:34 | 2017-09-25T12:22:34 | {} | MEMBER |
ClickHouse | ClickHouse | 331,868,418 | 1,275 | George3d6 | I believe the problem may lie in the columns definition:
`extension Array(Tuple(String, String))`
MergeTrees don't allow for array of array type columns, but it doesn't complain for array of tuples... though maybe, even if the warning is not there, the same bugs can arise (if tuple is just a wrapper over the arra... | 2017-09-25T12:40:45 | 2017-09-25T12:40:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 331,983,637 | 1,225 | filimonov | ```
:) select toRelativeDayNum(toDateTime(1000000000), 'America/New_York');
SELECT toRelativeDayNum(toDateTime(1000000000), 'America/New_York')
┌─toRelativeDayNum(toDateTime(1000000000), \'America/New_York\')─┐
│ 11573 │
└────────────────────────────────... | 2017-09-25T19:15:29 | 2017-09-25T19:15:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 331,989,258 | 1,225 | filimonov | It will work correctly (also for fractional offsets) if you know exactly the offset, which is not a constant value in time. :/
So for example for central europenian time (CET) you should use +2 hours offset in the winter, and +1 in the summer, and you should know exactly when to switch. If you also have a historic... | 2017-09-25T19:38:19 | 2017-09-25T19:38:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 332,500,073 | 1,296 | KochetovNicolai | No, it's not a bug. It's because ClickHouse uses 32-bit type to store DateType.
https://en.wikipedia.org/wiki/Year_2038_problem | 2017-09-27T12:08:30 | 2017-09-27T12:08:30 | {} | MEMBER |
ClickHouse | ClickHouse | 332,543,591 | 703 | filimonov | And if you provide less columns, than declared to be inserted - it executes it w/o error (should be exception):
```
:) insert into test(id,date) select number from system.numbers limit 10;
INSERT INTO test (id, date) SELECT number
FROM system.numbers
LIMIT 10
Ok.
``` | 2017-09-27T14:40:26 | 2017-09-27T14:40:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 332,756,271 | 390 | erdebee | +1, I think this should be fixed, as it makes writing "INSERT INTO SELECT FROM" queries unnecessarily cumbersome.
I would suggest that it should be mapped by name, and that it should not be necesarry that all columns that are available on the SELECT must be used in the INSERT | 2017-09-28T07:46:58 | 2017-09-28T07:48:40 | {} | NONE |
ClickHouse | ClickHouse | 333,082,309 | 1,305 | KochetovNicolai | Hi,
Primary key in ClickHouse is a sorted tuple. That's why when you use sessionId in "where clause" it's fast, but when you use only clientIp, it isn't. In the first case ClickHouse reads only data from continuous sessionId interval, but in the second it have to do many small reads for sessionIds which have appropr... | 2017-09-29T09:44:51 | 2017-09-29T09:44:51 | {} | MEMBER |
ClickHouse | ClickHouse | 331,581,096 | 1,277 | ramazanpolat | Ok. That's good to know. | 2017-09-22T23:15:39 | 2017-09-22T23:15:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 331,676,068 | 1,255 | alexey-milovidov | Note: You should merge master branch into your branch before making a pull request, because otherwise the log is too long. | 2017-09-23T23:23:07 | 2017-09-23T23:23:07 | {} | MEMBER |
ClickHouse | ClickHouse | 331,706,123 | 1,280 | alexey-milovidov | http://downforeveryoneorjustme.com/clickhouse.yandex
Looks like it is just for you.
Please provide more details: your IP address, traceroute, etc. | 2017-09-24T12:15:05 | 2017-09-24T12:15:05 | {} | MEMBER |
ClickHouse | ClickHouse | 331,711,387 | 1,280 | bookin | russian - http://clickhouse.readthedocs.io/en/latest/reference_ru.html | 2017-09-24T13:50:21 | 2017-09-24T13:50:21 | {} | NONE |
ClickHouse | ClickHouse | 331,867,458 | 1,281 | silviucpp | Hello,
I retested now after updated my cluster at (1.1.54292)
```sql
SELECT greatCircleDistance(NULL, 55.3, 38.7, 55.1) AS distance format JSONEachRow; - works OK
```
Then:
```sql
SELECT floor(greatCircleDistance(NULL, 55.3, 38.7, 55.1)) AS distance format JSONEachRow;
```
Crashing with:
```sh
2... | 2017-09-25T12:36:25 | 2017-09-25T12:36:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 331,946,101 | 1,282 | KochetovNicolai | Hi,
Thank you for detailed explanation! Problem was reproduced, will try to fix it. | 2017-09-25T17:01:45 | 2017-09-25T17:01:45 | {} | MEMBER |
ClickHouse | ClickHouse | 332,215,532 | 1,281 | silviucpp | @KochetovNicolai are you able to replicate using my last steps ?
Silviu | 2017-09-26T14:25:16 | 2017-09-26T14:25:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 332,390,610 | 1,290 | kislenok-roman | ClickHouse server version 1.1.54164. So we need to update... ok, thank you | 2017-09-27T02:27:21 | 2017-09-27T02:27:21 | {} | NONE |
ClickHouse | ClickHouse | 332,775,702 | 1,261 | wangyinsheng | em....
while resharding dosen't work, what should i do when removing a server from cluster?
by the way, is there any monitor tool for clickhouse cluster? | 2017-09-28T09:04:52 | 2017-09-28T09:04:52 | {} | NONE |
ClickHouse | ClickHouse | 332,802,227 | 127 | mbobakov | Any estimates ? | 2017-09-28T10:59:29 | 2017-09-28T10:59:29 | {
"+1": 4
} | NONE |
ClickHouse | ClickHouse | 332,811,129 | 1,300 | proller | а можно еще показать CREATE TABLE IF NOT EXISTS advertiser
? | 2017-09-28T11:41:55 | 2017-09-28T11:41:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 333,212,730 | 1,307 | fleapapa | Thanks. I see. I asked the question due to a suspicion that the loss might come from (unexpected) de-duplication of duplicate primary keys in the table. From the original CSV files, there might be some duplicate primary keys but not as many as 170M row. Chance should be small for different taxi drivers to pick up and d... | 2017-09-29T19:08:49 | 2017-09-29T19:09:21 | {} | NONE |
ClickHouse | ClickHouse | 331,990,226 | 1,225 | silviucpp | Ohh right.. | 2017-09-25T19:42:10 | 2017-09-25T19:42:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 332,812,265 | 1,300 | kshvakov | Конечно
```sql
CREATE TABLE IF NOT EXISTS reports.advertiser (
action_date Date,
action_time DateTime DEFAULT toDateTime(action_date),
rtb_client_id UInt32,
advertiser_id UInt32,
campaign_id UInt32,
banner_id UInt32,
url_id UInt32,
forma... | 2017-09-28T11:47:28 | 2017-09-28T11:47:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 332,881,325 | 1,288 | proller | looks like signal loop -
```
void BaseDaemon::initialize(Application & self) {
...
add_signal_handler({SIGABRT, SIGSEGV, SIGILL, SIGBUS, SIGSYS, SIGFPE, SIGPIPE},
faultSignalHandler);
static void faultSignalHandler(int sig, siginfo_t * info, void * context) {
...
call_default_signal_handler(sig);
... | 2017-09-28T15:54:43 | 2017-09-28T15:54:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 332,979,683 | 1,301 | alexey-milovidov | Your solution is very hacky and must be discouraged.
Please implement it in another way. Example: creating a database in the test script. | 2017-09-28T22:19:39 | 2017-09-28T22:19:39 | {} | MEMBER |
ClickHouse | ClickHouse | 330,822,406 | 1,252 | santaux | Ok, thank you! I'll try it! | 2017-09-20T11:21:42 | 2017-09-20T11:21:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 330,938,960 | 1,231 | silviucpp | @alexey-milovidov should I close this bug or you want to close it when doc is updated ? | 2017-09-20T18:24:46 | 2017-09-20T18:24:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 331,665,985 | 1,268 | ludv1x | Note that it should be merged with test env. PR https://nda.ya.ru/3TRcsG
Otherwise, tests are not passed. | 2017-09-23T19:51:33 | 2017-09-23T22:31:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 331,676,122 | 1,120 | alexey-milovidov | I have to close, because no answer was provided. | 2017-09-23T23:24:40 | 2017-09-23T23:24:40 | {} | MEMBER |
ClickHouse | ClickHouse | 331,932,795 | 1,278 | ramazanpolat | I can't reproduce the problem. They have different versions (1.1.54276 and 1.1.54284 ) but it works now.
| 2017-09-25T16:14:36 | 2017-09-25T16:14:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 331,978,498 | 1,225 | filimonov | >Basically your example just convert a date into another timezone.
Yes, and by introducing the fixed seconds offset you're getting exactly the same effect. But without extra knowlegde about DST, historical changes of local laws connected to local time.
DateTime in clickhouse, AFAIK, is just a regular unix timest... | 2017-09-25T18:55:27 | 2017-09-25T19:01:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 331,448,799 | 1,275 | ludv1x | Hi!
Do you use MergeTree engine? Could you show `SHOW CREATE TABLE impressions`? | 2017-09-22T13:37:35 | 2017-09-22T13:37:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 331,675,835 | 1,276 | alexey-milovidov | Missing task number in commit message (note that you can create tasks by yourself). | 2017-09-23T23:17:04 | 2017-09-23T23:17:04 | {} | MEMBER |
ClickHouse | ClickHouse | 332,193,913 | 167 | quba | @alexey-milovidov any update on this? [GDPR is coming](https://gdpr-info.eu/art-17-gdpr/). | 2017-09-26T13:13:51 | 2017-09-26T13:13:51 | {} | NONE |
ClickHouse | ClickHouse | 332,581,504 | 1,225 | filimonov | Anyway - i just wanted to say that there is a quite simple (may be a bit tricky) workaround. And fixed timezone offset is connected with certain risks (user should be aware that offset can change). So not sure if that feature is really useful.
About fractional timezones - I've created separate ticket: #1298 | 2017-09-27T16:37:41 | 2017-09-27T16:38:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 331,134,936 | 1,270 | tobad357 | @filimonov we solved this by using views for the dictionaries to query
Would that work for you? | 2017-09-21T11:57:46 | 2017-09-21T11:57:46 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 331,912,105 | 1,225 | filimonov | >> for each timezone at this point you create an instance of that timezone (as a singleton of DateLUT) which means if we add seconds as second param we will have lot of singletons created for each offset (timezone names are limited - number of seconds somehow are not) you don't consider this as a problem ?
>I think,... | 2017-09-25T15:09:30 | 2017-09-25T15:09:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 332,020,504 | 1,234 | vbadanov | See this: https://github.com/gperftools/gperftools/issues/360 | 2017-09-25T21:34:42 | 2017-09-25T21:34:42 | {} | NONE |
ClickHouse | ClickHouse | 332,596,377 | 1,283 | hamsterready | ```
:) select toTypeName(ClientIpAddress) from pageviews_dist limit 1;
SELECT toTypeName(ClientIpAddress)
FROM pageviews_dist
LIMIT 1
┌─toTypeName(ClientIpAddress)─┐
│ UInt32 │
└─────────────────────────────┘
1 rows in set. Elapsed: 5.476 sec.
:) select toTypeName(topK(1)(ClientI... | 2017-09-27T17:30:26 | 2017-09-27T17:30:26 | {} | NONE |
ClickHouse | ClickHouse | 333,192,956 | 1,307 | ztlpn | Hi @fleapapa ! Can you provide the CREATE TABLE statement for the problematic table? Is it a simple MergeTree table or some variant such as ReplacingMergeTree?
Can you try to reproduce the unwanted behaviour with a small repeatable test case? You can trigger additional merges with [OPTIMIZE statement](https://clickh... | 2017-09-29T17:46:24 | 2017-09-29T17:46:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 333,255,819 | 1,307 | fleapapa | Thanks for the investigation. I wrote a python script to load the data. The reason why my row count was less is there were quite some rows in the source files that contains '-10' and empty strings for _c8 and _c13. Afterward i changed _c8 to Int32 and _c13 to nullable to ignore the errors. Right now i'm reloading the d... | 2017-09-29T22:42:58 | 2017-09-29T22:49:28 | {} | NONE |
ClickHouse | ClickHouse | 333,317,645 | 1,307 | fleapapa | BTW, i also changed type of the table from MergeTree to ReplacingMergeTree so will expect any deduplication this time. I may stick to this table type for a while, so let me close this issue first. In case i switch back MergeTree and the issue happens again, i will re-open it. Thanks a lot for the investigation! | 2017-09-30T15:57:39 | 2017-09-30T15:57:39 | {} | NONE |
ClickHouse | ClickHouse | 331,818,543 | 1,261 | ludv1x | Currently, the resharding is not supported and it doesn't work. | 2017-09-25T08:54:53 | 2017-09-25T08:54:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 331,924,814 | 1,285 | KochetovNicolai | Confirmed, thank you for report. | 2017-09-25T15:49:37 | 2017-09-25T15:49:37 | {} | MEMBER |
ClickHouse | ClickHouse | 332,035,778 | 1,283 | vavrusa | What's the server version?
```
SELECT version()
┌─version()─┐
│ 1.1.54290 │
└───────────┘
``` | 2017-09-25T22:47:54 | 2017-09-25T22:47:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 333,271,701 | 1,305 | wusihang | Store duplicate is OK for me. Is it possible to support multiple indexes independently? | 2017-09-30T01:15:43 | 2017-09-30T01:15:43 | {} | NONE |
ClickHouse | ClickHouse | 330,832,846 | 1,261 | ludv1x | I should just initialize required tables on new servers and update cluster configuration on each server.
Changes in the configuration files will be applied automatically.
If you worry about the non-uniform distribution of the data across the shards (the new one will be empty) then you could reallocate shard weights... | 2017-09-20T12:14:05 | 2017-09-20T12:14:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 331,981,703 | 1,225 | silviucpp | Ok,
Let's suppose the following use case. All data are stored in a table in UTC. Then later you want to do a count grouped daily between two dates.
What you do is :
select count(*) as count, toRelativeDayNum(date) - toRelativeDayNum(X) from table
when date >= X and date =< Y
In this moment your result will... | 2017-09-25T19:07:30 | 2017-09-25T19:07:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 330,830,805 | 1,265 | ludv1x | More details about higher-order functions: https://clickhouse.yandex/docs/en/single/#higher-order-functions | 2017-09-20T12:04:16 | 2017-09-20T12:04:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 330,947,540 | 1,231 | alexey-milovidov | Better to close. | 2017-09-20T18:55:03 | 2017-09-20T18:55:03 | {} | MEMBER |
ClickHouse | ClickHouse | 331,635,315 | 1,272 | jackpgao |
Found that the replication engine can solve data loss problem.
But the failover solution is still not clear. | 2017-09-23T13:18:58 | 2017-09-23T13:18:58 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 332,029,885 | 1,283 | hamsterready | Hi @vavrusa,
I run the `topK()` on nullable(string), string, datetime, uint... and it seems that the problem is somewhere with `topK(_)(String)`, see some results:
```
:) select IPv4NumToStringClassC(ClientIpAddress) from pageviews_dist where SessionId = UUIDStringToNum('6af5e8fa-2d9a-407c-b37f-fe562355077e')
... | 2017-09-25T22:16:46 | 2017-09-25T22:16:46 | {} | NONE |
ClickHouse | ClickHouse | 332,738,926 | 1,044 | PhantomPhreak | Interested in this feature too. 2 option is exactly what is needed.
Thanks in advance for this feature!
| 2017-09-28T06:19:13 | 2017-09-28T06:19:13 | {} | NONE |
ClickHouse | ClickHouse | 333,198,694 | 1,308 | fleapapa | Any sample include files for the two unfound includes will be appreciated :) | 2017-09-29T18:10:37 | 2017-09-29T18:10:37 | {} | NONE |
ClickHouse | ClickHouse | 333,242,042 | 1,307 | ztlpn | Well, I failed to reproduce it with clickhouse-server from our prebuilt packages. I created the table with the date column defined as `_cx Date default toDate(_c5)`, loaded all the data into it with the following statement:
```
for i in {1..12}; do tail -n +2 trip_data_$i.csv | clickhouse-client -q 'insert into trip(... | 2017-09-29T21:21:10 | 2017-09-29T21:21:10 | {} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.