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 | 445,941,966 | 3,792 | alexey-milovidov | This is very non-obvious behaviour.
Views will stop working when table that they depend on is dropped.
- that's Ok.
Views will not continue to work when new table with the same name is created.
- this is not Ok, because the user will look at the definition of materialized view and wonder.
After server restar... | 2018-12-10T19:30:38 | 2018-12-10T19:30:38 | {} | MEMBER |
ClickHouse | ClickHouse | 446,180,753 | 3,806 | filimonov | > The only concern is that the user may disable HTTP completely (they can leave HTTPS only).
(For example, this is the case for Yandex.Cloud).
Actually, every separate part can be disabled. `tcp_port` can be disabled too. Secure ports are not enabled by default. Message line "[Ready for connections.](https://github... | 2018-12-11T12:10:24 | 2018-12-11T12:15:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 446,958,122 | 2,709 | dqminh | I think the issue is not so relevant now since we are so far ahead now and not going to downgrade. I'm going to close this. | 2018-12-13T12:53:58 | 2018-12-13T12:53:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 447,350,791 | 3,829 | kotoyants | +1
After update from 18.10.3 to 18.14.18 dictionaries (clickhouse, postgresql) started to accidentally stop updating. Downgrade to 18.14.9 resolves the issue.
No errors in the log file, only one exeption for PDO dictionary:
> Poco::Exception. Code: 1000, e.code() = 11, e.displayText() = Timeout, e.what() = Timeout
... | 2018-12-14T15:01:40 | 2018-12-14T15:15:55 | {} | NONE |
ClickHouse | ClickHouse | 447,433,152 | 3,836 | alexey-milovidov | Missing test. | 2018-12-14T19:41:32 | 2018-12-14T19:41:32 | {} | MEMBER |
ClickHouse | ClickHouse | 447,626,476 | 3,829 | k0st1an | 18.16.0 also has problems. | 2018-12-16T08:43:21 | 2018-12-16T08:43:21 | {} | NONE |
ClickHouse | ClickHouse | 447,835,069 | 3,829 | k0st1an | `SYSTEM RELOAD DICTIONARIES` executed N count. no problem. Then at some point timeout. | 2018-12-17T12:50:49 | 2018-12-17T12:51:35 | {} | NONE |
ClickHouse | ClickHouse | 447,878,222 | 3,837 | alexey-milovidov | Thank you!
PS. Single integration test has failed - false positive. | 2018-12-17T15:09:32 | 2018-12-17T15:09:32 | {} | MEMBER |
ClickHouse | ClickHouse | 446,523,787 | 3,717 | alexey-milovidov | Ok. A setting looks reasonable. | 2018-12-12T09:39:15 | 2018-12-12T09:39:15 | {} | MEMBER |
ClickHouse | ClickHouse | 447,402,737 | 3,569 | amosbird | +1 to `__attribute__((optimize("fwrapv")))` . It's very descriptive. | 2018-12-14T17:56:22 | 2018-12-14T17:56:22 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 447,550,925 | 3,829 | k0st1an | +1 | 2018-12-15T08:48:09 | 2018-12-15T08:48:09 | {} | NONE |
ClickHouse | ClickHouse | 447,791,104 | 3,829 | k0st1an | Error appears when updating one of the dictionaries failed:
```
Poco::Exception. Code: 1000, e.code() = 11, e.displayText() = Timeout, e.what() = Timeout
```
After this the dictionaries are not updated. `SYSTEM RELOAD DICTIONARIES` is wery long... infinity (
If kill the `clickhouse odbc-bridge`, it will rest... | 2018-12-17T10:11:52 | 2018-12-17T10:11:52 | {
"+1": 2
} | NONE |
ClickHouse | ClickHouse | 447,983,815 | 3,857 | alesapin | I don't know why this didn't fire in daemon mode. Maybe because daemon process were started from shell. | 2018-12-17T20:16:48 | 2018-12-17T20:16:48 | {} | MEMBER |
ClickHouse | ClickHouse | 448,180,032 | 3,829 | alesapin | @k0st1an @angeldestroyer as workaround specify settings for odbc-bridge logging:
https://github.com/yandex/ClickHouse/issues/3825#issuecomment-448179911 | 2018-12-18T10:53:16 | 2018-12-18T10:54:25 | {} | MEMBER |
ClickHouse | ClickHouse | 445,932,077 | 3,800 | alexey-milovidov | I think, the user will expect, that `min_date`, `max_date` columns will be also filled with non-zero value, if there is `min_time`, `max_time`. | 2018-12-10T19:02:41 | 2018-12-10T19:02:41 | {} | MEMBER |
ClickHouse | ClickHouse | 447,406,751 | 3,831 | filimonov | Did you enable on disk group by and order by?
Check those settings:
max_bytes_before_external_sort
max_bytes_before_external_group_by
distributed_aggregation_memory_efficient | 2018-12-14T18:10:18 | 2018-12-14T18:10:18 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 447,550,953 | 3,829 | alesapin | @angeldestroyer In startup process clickhouse updates dictionaries for the first time. If I understood correctly clickhouse can start with that configuration, but cannot update dictionaries?
Can you set your log-level to trace? Also, can you provide information about your OS? Maybe you use clickhouse in docker? Also... | 2018-12-15T08:48:38 | 2018-12-15T08:48:38 | {} | MEMBER |
ClickHouse | ClickHouse | 447,833,338 | 3,829 | k0st1an | ```
┌─version()─┐
│ 18.16.0 │
└───────────┘
...
┌─name────────┬─type─────────────┬─bytes_allocated─┬─last_exception────────────────────────────────────────────────────────────────────────────┐
│ xxxxxxxxx │ Hashed │ 18200680 │ Poco::Exception. Code: 1000, e.code() = 11, e.displayText() = Time... | 2018-12-17T12:44:45 | 2018-12-17T12:44:45 | {} | NONE |
ClickHouse | ClickHouse | 447,838,951 | 3,843 | proller | 1. hotfix for you: add -DARCH_NATIVE=0 to cmake flags.
1.1 or try use gcc (but it usually does not work for me)
2. remove -DHAVE_SSE41=0 -DHAVE_SSE42=0 -DHAVE_SSSE3=0 -DHAVE_AVX2=0 -DHAVE_AVX=0 -DHAVE_POPCNT=0 -DHAVE_SSE2_INTRIN=0 -DSSE2~AG=' ' -DHAVE_PCLMULQDQ_INTRIN=0 -DPCLMULFLAG=' '"
- it should work automati... | 2018-12-17T13:05:51 | 2018-12-17T13:05:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 447,912,769 | 79 | filimonov | @DivyaRao30 please don't duplicate your comments. I've answered you in #3847 | 2018-12-17T16:43:47 | 2018-12-17T16:43:47 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 446,178,171 | 3,809 | blinkov | Let's keep this discussion on stackoverflow, no need to have two threads about same topic: https://stackoverflow.com/questions/53709001/carry-forward-rounding | 2018-12-11T12:00:46 | 2018-12-11T12:00:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 446,370,437 | 3,816 | alesapin | I'd like to link `libglibc-compatibility` to `clickhouse` binary only. However cmake reorder positions of libraries in linker command even when this library added with `target_link_libraries` in last position:
```
/usr/bin/g++-7 -fdiagnostics-color=always -pipe -msse4.1 -msse4.2 -mpopcnt -fno-omit-frame-pointer -... | 2018-12-11T21:34:08 | 2018-12-11T21:34:08 | {} | MEMBER |
ClickHouse | ClickHouse | 446,427,789 | 3,796 | amosbird | Updated. It seems unavoidable to touch the context class. | 2018-12-12T01:32:22 | 2018-12-12T01:32:22 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 446,463,896 | 3,796 | amosbird | Note, another reason for getDatabaseName in IStorage | 2018-12-12T05:08:37 | 2018-12-12T05:08:37 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 446,528,587 | 3,816 | alesapin | JFYI https://lwn.net/Articles/711013/ | 2018-12-12T09:55:16 | 2018-12-12T09:55:16 | {} | MEMBER |
ClickHouse | ClickHouse | 447,022,077 | 2,947 | ztlpn | Will look into this. The intention is that SessionExpired event should clear the znode cache and all needed znodes would be reloaded. | 2018-12-13T16:02:05 | 2018-12-13T16:02:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 447,098,447 | 3,360 | k0st1an | Solution: enable the ability to connect to PSQL without SSL and remove from config
`/etc/odbc.ini` line:
```
SSLmode = require
```
But in the old version, it works with SSL:
```
SELECT version()
┌─version()─┐
│ 18.4.0 │
└───────────┘
``` | 2018-12-13T19:52:15 | 2018-12-13T19:52:15 | {} | NONE |
ClickHouse | ClickHouse | 448,153,244 | 3,829 | snvad | I tried to "enable" dictionaries from small to large.
Up to this point, everything was fine:
```
┌─type─────────────┬─key──────┬─element_count─┬─formatReadableSize(bytes_allocated)─┬───────creation_time─┬─last_exception─┐
│ ComplexKeyHashed │ (String) │ 2 │ 18.37 KiB │ 2018-12-... | 2018-12-18T09:24:35 | 2018-12-18T09:24:35 | {} | NONE |
ClickHouse | ClickHouse | 446,278,894 | 3,815 | blinkov | There's a `WITH` clause https://github.com/yandex/ClickHouse/blob/master/dbms/tests/queries/0_stateless/00490_with_select.sql , but it has limitations that prevent its usage in specifically your case. | 2018-12-11T17:00:39 | 2018-12-11T17:00:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 447,347,173 | 3,569 | alexey-milovidov | Yes. But it violates the C++ standard and it will confuse compiler. And all tests should pass under UBSan (currently they don't; and we don't have autotests with UBSan). | 2018-12-14T14:49:27 | 2018-12-14T14:50:03 | {} | MEMBER |
ClickHouse | ClickHouse | 447,911,897 | 3,847 | filimonov | Don't you wrap whole command in some another bash call? (in that case double quote char can disappear, and you will get the similar error).
In case of direct call everything works.
```
CREATE DATABASE insightbasev2;
CREATE TABLE insightbasev2.company_temp
(
a UInt32,
b String,
c String,
... | 2018-12-17T16:41:34 | 2018-12-17T16:41:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 448,078,390 | 3,811 | alexey-milovidov | I will try to address all my concerns and merge this PR. | 2018-12-18T02:55:15 | 2018-12-18T02:55:15 | {} | MEMBER |
ClickHouse | ClickHouse | 446,128,620 | 3,584 | denisov-vlad | Thanks. `toNullable()` works.
```
CREATE TABLE local.apps_sessions (
app_id UInt32, app_name String, app_version String,
os_name Enum8('ios' = 0, 'android' = 1, 'windows' = 2),
os_version String, session_id UInt64, session_time DateTime,
profile_id Nullable(String), device_id_appmetrica UInt6... | 2018-12-11T09:17:24 | 2018-12-11T09:17:24 | {} | NONE |
ClickHouse | ClickHouse | 446,177,897 | 1,130 | 4ertus2 | Works in master. It needs insert_sample_with_metadata setting to be set.
```
cat ~/clickhouse/test/head.json | ./clickhouse-client --input_format_skip_unknown_fields=1 --insert_sample_with_metadata=1 -q "INSERT INTO commententry1 FORMAT JSONEachRow"
select created_utc, created_date from commententry1
SELECT
... | 2018-12-11T11:59:51 | 2018-12-11T11:59:51 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 446,636,159 | 2,085 | vmanyushin | Спасибо! | 2018-12-12T15:50:47 | 2018-12-12T15:50:47 | {} | NONE |
ClickHouse | ClickHouse | 447,115,163 | 3,775 | alex-krash | @alexey-milovidov , sorry for interrupting, just wondering. Why you prefer string representation of signature instead of having a separate method/visitor pattern/whatever with explicit programmatic specification?
(I have a doubt about "primitive obsession", but just want to have Guru opinion on this stuff). | 2018-12-13T20:47:09 | 2018-12-13T20:47:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 447,342,797 | 2,440 | alexey-milovidov | It is 5% done, at most. This PR has no future.
We will continue in a separate PR. | 2018-12-14T14:33:48 | 2018-12-14T14:33:48 | {} | MEMBER |
ClickHouse | ClickHouse | 447,343,885 | 2,441 | alexey-milovidov | Superseded by #2586 | 2018-12-14T14:37:41 | 2018-12-14T14:37:41 | {} | MEMBER |
ClickHouse | ClickHouse | 447,551,556 | 3,829 | k0st1an | OS: ubuntu 16.04.5 LTS (not docker)
```
SYSTEM RELOAD DICTIONARY <name>
```
long wait ... log:
```
2018.12.15 08:53:57.249988 [ 8869 ] {e4af30af-f454-49c2-92e8-5accefc67aab} <Debug> executeQuery: (from [::1]:50352) SYSTEM RELOAD DICTIONARY <name>
```
```
$ ps aux | grep clickhouse
clickho+ 45479 4.8 ... | 2018-12-15T09:00:12 | 2018-12-15T09:00:12 | {} | NONE |
ClickHouse | ClickHouse | 447,796,572 | 3,829 | k0st1an | ```
curl http://localhost:9018/ping
```
Not works if `last_exception` is `Poco::Exception. Code: 1000, e.code() = 11, e.displayText() = Timeout, e.what() = Timeout` | 2018-12-17T10:29:49 | 2018-12-17T10:29:49 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 447,821,828 | 3,829 | alesapin | Was `clickhouse odbc-bridge` alive when you have tried to `curl`? What exactly `curl http://localhost:9018/ping` output?
Anyway, need logs of `clickhouse-server` in **trace** level or reproducible example. Also I'd like to see dictionary config. | 2018-12-17T12:04:37 | 2018-12-17T12:12:58 | {} | MEMBER |
ClickHouse | ClickHouse | 447,089,478 | 3,360 | k0st1an | ```
SELECT version()
┌─version()─┐
│ 18.14.18 │
└───────────┘
```
postgresql log:
```
2018-12-13 19:19:00.750 UTC [13409] [unknown]@10.99.67.110/[unknown] LOG: could not accept SSL connection: EOF detected
2018-12-13 19:19:10.770 UTC [13416] [unknown]@10.99.67.110/[unknown] LOG: could not accept SSL con... | 2018-12-13T19:24:59 | 2018-12-13T19:25:17 | {} | NONE |
ClickHouse | ClickHouse | 446,868,616 | 3,705 | blinkov | @dfulachier could you please show some examples of queries with `TIMESTAMPADD` that you get generated? | 2018-12-13T07:27:03 | 2018-12-13T07:27:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 446,929,434 | 3,800 | alexey-milovidov | Ok. | 2018-12-13T11:05:29 | 2018-12-13T11:05:29 | {} | MEMBER |
ClickHouse | ClickHouse | 447,573,340 | 3,764 | shangshujie365 | @viongpanzi
not yet, if your column data has no null value, my fix can work.
If your data has null value, the null value result will be wrong. | 2018-12-15T14:42:11 | 2018-12-17T02:30:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 448,146,310 | 3,829 | snvad | The problem is not reproduced on a small number of dictionaries (or allocated_memory).
"Disabled" all dictionaries and restarted CH (I have 12 Hashed/ComplexKeyHashed dictionaries with size ranging from 100KB to 1,4GB). Enable two test-dictionaries.
```
┌─name─────────────────────┬─type─────────────┬─key──────┬─el... | 2018-12-18T09:00:57 | 2018-12-18T09:00:57 | {} | NONE |
ClickHouse | ClickHouse | 448,179,911 | 3,825 | alesapin | The bug fixed in https://github.com/yandex/ClickHouse/pull/3857. As workaround you can set logger settings for odbc-bridge in server `config.xml` (it will not hang in current version):
```
<logger>
<!-- Possible levels: https://github.com/pocoproject/poco/blob/develop/Foundation/include/Poco/Logger.h#L10... | 2018-12-18T10:52:52 | 2018-12-18T10:52:52 | {
"hooray": 1
} | MEMBER |
ClickHouse | ClickHouse | 446,539,930 | 2,586 | alesapin | @zhang2014 I'll finish this task, because we need to merge it in two weeks. If you have some unfinished work commit please. | 2018-12-12T10:32:57 | 2018-12-12T11:15:12 | {} | MEMBER |
ClickHouse | ClickHouse | 446,744,024 | 3,220 | alexey-milovidov | Fixed in master. | 2018-12-12T21:07:04 | 2018-12-12T21:07:04 | {} | MEMBER |
ClickHouse | ClickHouse | 446,906,509 | 3,764 | viongpanzi | hi @shangshujie365
We have the same problem, is the problem fixed now? | 2018-12-13T09:48:39 | 2018-12-13T09:48:39 | {} | NONE |
ClickHouse | ClickHouse | 400,305,255 | 2,529 | Joozty | @filimonov Thanks a lot. I finally found a solution. Actually `max_query_size` did work. The problem was that the number wasn't high enough. The number is in bytes, isn't it? Maybe you could write this to the documentation.
| 2018-06-26T13:20:53 | 2018-12-14T13:38:00 | {
"+1": 3
} | NONE |
ClickHouse | ClickHouse | 447,408,355 | 3,592 | alexey-milovidov | @abyss7 is on leave this week.
I hope, this task will be continued next week. | 2018-12-14T18:15:28 | 2018-12-14T18:15:28 | {} | MEMBER |
ClickHouse | ClickHouse | 447,909,057 | 3,849 | filimonov | related #1403 | 2018-12-17T16:33:48 | 2018-12-17T16:33:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 447,989,888 | 3,592 | kanterov | @alexey-milovidov thanks for the update! I did a test of this PR, in our case, there is a significant speedup between 2 and 8 times depending on queries and tables.
One thing I've noticed, that it's possible to create distributed tables without sharding key, and in this case, this code will cause a segmentation faul... | 2018-12-17T20:36:34 | 2018-12-17T20:36:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 446,137,790 | 3,800 | alexey-milovidov | > For my understanding, DateTime is timezone agnostic (it could be seen to any timezone, and still refer to the same instant). But Date is timezone dependant. So for converting a DateTime to Date, you need to specify a timezone. So here, which timezone could we use for converting the DateTime to Date ? The system timez... | 2018-12-11T09:46:04 | 2018-12-11T09:46:04 | {} | MEMBER |
ClickHouse | ClickHouse | 446,184,262 | 3,806 | filimonov | Also there is a bit hacky workaround - just to enable http_port by passing it via `clickhouse-server` command line paramaters. In that case some logic for choosing free port should be added to the script. | 2018-12-11T12:23:52 | 2018-12-11T12:23:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 445,970,505 | 3,792 | den-crane | My common usecase is:
T->( MV(T),MV1(T),MV2(T),MV4(T) )
stop load to T
rename T to T_arc201811;create table T as T_arc201811;
start load.
I don't really understand why CH needs to drop MV automatically. | 2018-12-10T20:59:05 | 2018-12-10T21:00:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 447,576,954 | 2,415 | salisbury-espinosa | extra info on the subject of issue
```sql
select arrayReduce(x -> x, [1, 2, 3])
```
> DB::Exception: Function arrayReduce can't have lambda-expressions as arguments. | 2018-12-15T15:37:45 | 2018-12-15T15:37:45 | {
"+1": 2
} | NONE |
ClickHouse | ClickHouse | 447,875,695 | 3,825 | ztlpn | @alesapin please take a look | 2018-12-17T15:02:16 | 2018-12-17T15:02:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 448,116,386 | 3,696 | yuliujuan | @blinkov
sorry for the late reply.
I will use custom partitioning.
Please close the issue. Thank you all.
Actually, I am thinking that if select * from t final can be used for getting the latest record, is it possible to make "optimize table t FINAL" to merge data from different partitions, 'cause the site doesn... | 2018-12-18T06:55:32 | 2018-12-18T06:55:32 | {} | NONE |
ClickHouse | ClickHouse | 448,125,544 | 3,811 | shangshujie365 | > I will try to address all my concerns and merge this PR.
That's cool | 2018-12-18T07:40:21 | 2018-12-18T07:40:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 448,141,897 | 3,849 | crakjie | @filimonov maybe related but it's not in HTTP.
In my case the query is finish and the connection is not closed. | 2018-12-18T08:45:31 | 2018-12-18T08:47:19 | {} | NONE |
ClickHouse | ClickHouse | 446,557,812 | 3,802 | simPod | Hm, you're right. Tried to create minimal example and it works :/ Closing as cannot reproduce-'. | 2018-12-12T11:38:07 | 2018-12-12T11:38:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 447,104,772 | 3,825 | iFoton | Same problem. Begins since 18.14.17 | 2018-12-13T20:11:54 | 2018-12-13T20:11:54 | {} | NONE |
ClickHouse | ClickHouse | 447,174,888 | 3,775 | alexey-milovidov | > @alexey-milovidov , sorry for interrupting, just wondering. Why you prefer string representation of signature instead of having a separate method/visitor pattern/whatever with explicit programmatic specification?
(I have a doubt about "primitive obsession", but just want to have Guru opinion on this stuff).
We ne... | 2018-12-14T00:52:17 | 2018-12-14T00:52:17 | {} | MEMBER |
ClickHouse | ClickHouse | 447,844,181 | 3,829 | alesapin | Sorry, this diagnosis is not enought. | 2018-12-17T13:24:34 | 2018-12-17T13:24:34 | {} | MEMBER |
ClickHouse | ClickHouse | 446,126,421 | 3,800 | edonin | For the tests, I will take a look. | 2018-12-11T09:10:35 | 2018-12-11T09:10:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 446,175,116 | 3,810 | blinkov | Duplicate for #3779 | 2018-12-11T11:48:57 | 2018-12-11T11:48:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 447,794,653 | 3,829 | alesapin | @k0st1an I.e. single update performed after odbc-bridge restart, but all consecutive updates hang?
`clickhouse-server` communicate with `clickhouse odbc-bridge` via http, using tcp port 9018 on localhost. Maybe you have some iptables rules which forbid these connections?
If you enable trace log level you should see f... | 2018-12-17T10:23:26 | 2018-12-17T10:24:17 | {} | MEMBER |
ClickHouse | ClickHouse | 447,914,516 | 3,847 | Crazylearner30 | Hi ,
I do wrap the commands in single line.
Understood , but I have exported the file from another database and using
it as input file for clickhouse which is in cvs format.
When I tried to load the input file data into clickhouse getting the
mentioned error.
Input file is huge like millions.
On Mon, 1... | 2018-12-17T16:48:29 | 2018-12-17T16:51:49 | {} | NONE |
ClickHouse | ClickHouse | 446,183,780 | 75 | 4ertus2 | Works in master. It needs insert_sample_with_metadata to be set.
```
DROP TABLE IF EXISTS test_table;
CREATE TABLE test_table (
a String,
b Date,
c DEFAULT sipHash64(a),
d MATERIALIZED sipHash64(a)
) ENGINE = MergeTree(b, a, 8192);
SET insert_sample_with_metadata=1;
INSERT INTO test_table FORMAT... | 2018-12-11T12:22:12 | 2018-12-11T12:22:12 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 446,185,508 | 1,870 | 4ertus2 | Works in master (with setting insert_sample_with_metadata set).
```
echo '{"_time":"2018-02-07T11:33:14","body":"something here"}' | ./clickhouse-client --insert_sample_with_metadata=1 -q "INSERT INTO timeseq FORMAT JSONEachRow"
./clickhouse-client -q 'select * from timeseq'
2018-02-07 11:33:14 2018-02-07 ... | 2018-12-11T12:28:46 | 2018-12-11T12:28:46 | {
"+1": 2,
"hooray": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 446,556,316 | 3,802 | simPod | Hi, sure!
`SELECT version()`
18.14.18 | 2018-12-12T11:32:42 | 2018-12-12T11:32:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 446,874,773 | 3,705 | dfulachier | Ivan, I get sometimes very complex queries with Tableau, concerning date grouping.
Samples provided in the page I mentionned before (http://kb.its.psu.edu/node/3349) are significant.
One sample generated from Tableau (through ClickHouse ODBC driver):
```
SELECT
SUM(`IuR_Interface_Huawei_30_Minutes`.`C_73393847`)... | 2018-12-13T07:55:25 | 2018-12-13T07:55:25 | {} | NONE |
ClickHouse | ClickHouse | 447,100,325 | 3,360 | k0st1an | Through `isql` also works without problems. | 2018-12-13T19:58:05 | 2018-12-13T19:58:05 | {} | NONE |
ClickHouse | ClickHouse | 447,398,155 | 3,569 | alexey-milovidov | We can avoid it in two possible ways:
1. Add `__attribute__((optimize("fwrapv")))` for function.
2. Parse in unsigned type. Then cast to signed, perform a branch for min int, invert. Or invert via bit not and plus one, and then do bit_cast to signed. | 2018-12-14T17:41:05 | 2018-12-14T17:41:44 | {} | MEMBER |
ClickHouse | ClickHouse | 447,781,584 | 3,829 | alesapin | Sorry, did not notice, that you use `complex_key_hashed` dictionary, but situation is similar with this config:
```
<dictionaries>
<dictionary>
<name>postgres_odbc_hashed</name>
<source>
<odbc>
<table>clickhouse.test_table</table>
<connection_strin... | 2018-12-17T09:43:02 | 2018-12-17T09:43:02 | {} | MEMBER |
ClickHouse | ClickHouse | 447,935,048 | 1,827 | proller | https://github.com/yandex/ClickHouse/pull/2443 | 2018-12-17T17:48:46 | 2018-12-17T17:48:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 446,263,584 | 3,814 | den-crane | kafka engine does not respect input_format_allow_errors_ratio input_format_allow_errors_num #2171 #2675
| 2018-12-11T16:19:50 | 2018-12-11T16:19:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 446,446,451 | 3,717 | amosbird | The performance degrade badly for small arrays. It makes sense to add an option then.
```
Running: array_join
[
{
"hostname": "201.nobida.cn",
"main_metric": "max_rows_per_second",
"num_cores": 20,
"num_threads": 40,
"ram": 134768111616,
"runs": [
{
"max_rows_... | 2018-12-12T03:13:42 | 2018-12-12T03:13:42 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 446,539,071 | 3,816 | alesapin | Make check -- unrelated `illegal instruction` with gcc-8 and old processor. | 2018-12-12T10:30:06 | 2018-12-12T10:30:06 | {} | MEMBER |
ClickHouse | ClickHouse | 446,957,941 | 2,947 | dqminh | We are still seeing this issue on v18.14.17.
An example of the timeline:
1. Node get SessionExpired
2. Some time after, to test this, i manually update the config in zookeeper, and check generated config in /etc/clickhouse-server/config-preprocessed.xml. Change is not propagated. This is tried a few times without... | 2018-12-13T12:53:12 | 2018-12-13T12:53:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 447,107,042 | 3,360 | alexey-milovidov | We've figured out the root cause and the way to fix it. | 2018-12-13T20:19:18 | 2018-12-13T20:19:18 | {} | MEMBER |
ClickHouse | ClickHouse | 447,458,499 | 3,835 | alexey-milovidov | In master. | 2018-12-14T20:22:54 | 2018-12-14T20:22:54 | {} | MEMBER |
ClickHouse | ClickHouse | 447,876,119 | 3,825 | alesapin | Looks same https://github.com/yandex/ClickHouse/issues/3829. | 2018-12-17T15:03:25 | 2018-12-17T15:03:25 | {} | MEMBER |
ClickHouse | ClickHouse | 448,195,689 | 3,529 | madm1keru | Need to add some details.
4 shards, 2 replicas, insert to Distributed table on round-robin servers, ReplicatedMergeTree as base storage, about a 6 materialized views - ReplicatedAggregatedMergeTree.
Data sharded about an equally per shard. About 15 original inserts per minute to cluster - and it's turns to signi... | 2018-12-18T11:54:19 | 2018-12-18T11:54:19 | {} | NONE |
ClickHouse | ClickHouse | 445,943,100 | 3,789 | Jack012a | Thank you! | 2018-12-10T19:33:46 | 2018-12-10T19:33:46 | {} | NONE |
ClickHouse | ClickHouse | 446,065,705 | 102 | TianyiLu | The [Size] parameter still cannot use refer to https://clickhouse.yandex/docs/en/query_language/functions/date_time_functions/#timeslotsstarttime-duration91-size93
SELECT timeSlots(toDateTime('2018-09-10 00:00:00'), toUInt32(6000), 300)
Received exception from server (version 1.1.54370):
Code: 42. DB::Exception:... | 2018-12-11T04:08:53 | 2018-12-11T04:10:03 | {} | NONE |
ClickHouse | ClickHouse | 446,129,306 | 3,800 | edonin | For the other issue, I have a doubt.
For my understanding, DateTime is timezone agnostic (it could be seen to any timezone, and still refer to the same instant). But Date is timezone dependant. So for converting a DateTime to Date, you need to specify a timezone. So here, which timezone could we use for converting ... | 2018-12-11T09:19:25 | 2018-12-11T09:19:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 446,172,628 | 3,802 | s-mx | Hi. I checked on my machine and that's ok. Could you provide more info about server version? It would be great if you can write small ddl script to demonstrate that behavior. | 2018-12-11T11:39:29 | 2018-12-11T11:39:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 446,634,619 | 2,085 | filimonov | in master #3724 | 2018-12-12T15:46:40 | 2018-12-12T15:46:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 446,657,588 | 3,800 | edonin | I added some tests.
Nevertheless, the result is a bit different of what i described in the issue. The column min_time and max_time are filled if the merge tree **partitionning index** contains a DateTime. It does not care about the merge tree **sorting index**.
Since these two columns is defined for each partitio... | 2018-12-12T16:46:49 | 2018-12-12T16:46:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 446,659,405 | 3,821 | filimonov | ```sql
drop database `test-test`;
-- or
drop database "test-test";
```
See https://clickhouse.yandex/docs/en/query_language/syntax/#identifiers | 2018-12-12T16:51:40 | 2018-12-12T16:51:40 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 446,702,366 | 3,783 | alexey-milovidov | @4ertus2 Please note that `__uint128_t` assumes 16-byte alignment.
```
template <typename T>
const char * ColumnDecimal<T>::deserializeAndInsertFromArena(const char * pos)
{
data.push_back(*reinterpret_cast<const T *>(pos));
return pos + sizeof(T);
}
```
is incorrect.
You should use `unalignedLoad`. | 2018-12-12T18:57:33 | 2018-12-12T18:57:33 | {} | MEMBER |
ClickHouse | ClickHouse | 447,135,353 | 3,756 | reflection | Hi, any update on this issue as it's been over a week? Thanks! | 2018-12-13T21:58:00 | 2018-12-13T21:58:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 447,552,943 | 3,829 | k0st1an | Process `clickhouse odbc-bridge` should to down after update dict? Checked at 18.16.0. SYSTEM RELOAD DICTIONARY has worked, but the process is hanging. Restarting works without problems. But this is a fresh install. Maybe when there is a timeout error then SYSTEM RELOAD DICTIONARY will not work. | 2018-12-15T09:19:59 | 2018-12-15T09:19:59 | {} | NONE |
ClickHouse | ClickHouse | 447,553,878 | 3,829 | alesapin | > Process clickhouse odbc-bridge should to down after update dict?
No, this process should stay alive with clickhouse server. | 2018-12-15T09:33:46 | 2018-12-15T09:37:35 | {} | MEMBER |
ClickHouse | ClickHouse | 447,626,084 | 3,829 | snvad | @alesapin, yes, at start all looks fine. But when the update should be performed - timeout occurs and nothing happens. Not docker, not vm. Package installed from the official repository. In the logs only that indicated when opening issue, without additional information.
OS: Ubuntu 16.04.3 LTS
```bash
76116 ? ... | 2018-12-16T08:35:15 | 2018-12-16T08:35:15 | {} | NONE |
ClickHouse | ClickHouse | 447,778,007 | 3,829 | alesapin | I have tried to reproduce the issue, but everything seems to work fine for me(.
My setup:
- ubuntu 16.04;
- clickhouse 18.14.18;
- postgres from docker image `docker run -p 5432:5432 --name some-postgres1 -e POSTGRES_PASSWORD=mysecretpassword -d postgres`.
- unixodbc 2.3.1-4.1
- odbc-postgresql:amd64 1:09.03.030... | 2018-12-17T09:31:04 | 2018-12-17T09:31:04 | {} | MEMBER |
ClickHouse | ClickHouse | 448,186,714 | 3,597 | brobusta | Hi @ztlpn @healiseu,
I tried mutation with SET and it is still failed.
```sql
DROP TABLE IF EXISTS default.test;
CREATE TABLE default.test (d Date default today(), i Int8, v Int8) Engine = MergeTree(d, (d,i), 8192);
INSERT INTO default.test (i, v) VALUES (0, 100);
INSERT INTO default.test (i, v) VALUES (1, 11... | 2018-12-18T11:17:30 | 2018-12-18T11:19:16 | {} | NONE |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.