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 | 378,929,563 | 2,010 | silviucpp | Is this supposed to be fixed ? I still get `Code: 48, e.displayText() = DB::Exception: Method getDataAt is not supported for Nullable(Float64), e.what() = DB::Exception` when doing `select uniq(arr) from test_array2` | 2018-04-05T13:07:56 | 2018-04-05T13:07:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 380,126,129 | 2,180 | daskol | I have finally built client for ClickHouse v1.1.54375-testing. I run `clickhouse-client` and quit then client throws exception with following stacktrace. I have found the same exception in older versions too.
```
Core was generated by `clickhouse-client'.
Program terminated with signal SIGABRT, Aborted.
#0 0x000... | 2018-04-10T14:42:26 | 2018-04-10T14:42:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 380,293,560 | 2,204 | pdeva | looking around it seems this could be due to #2086.
the table above only has those 11 rows total | 2018-04-11T01:03:29 | 2018-04-11T01:03:29 | {} | NONE |
ClickHouse | ClickHouse | 378,578,522 | 2,093 | vpanfilov | It would be great for determining region from telephone number. | 2018-04-04T12:12:52 | 2018-04-04T12:12:52 | {} | NONE |
ClickHouse | ClickHouse | 378,851,403 | 2,169 | vavrusa | @ennio1991 in your case the problem is that the event rate is low, so you have to adjust the settings accordingly to get lower latency. The `max_block_size` and `stream_flush_interval_ms` should work, but you need to set those in the environment before you create/attach tables. The `CREATE TABLE ... SETTINGS x=y` might... | 2018-04-05T07:56:50 | 2018-04-05T07:56:50 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,931,768 | 2,179 | tobad357 | As far as I know mid may - june is planned for some simple update/delete functionality
| 2018-04-05T13:15:36 | 2018-04-05T13:15:36 | {} | NONE |
ClickHouse | ClickHouse | 378,956,603 | 2,180 | proller | Our builds in travis near time limits.
When we add any new build - wait for test result increased by hours | 2018-04-05T14:31:16 | 2018-04-05T14:31:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,962,812 | 2,178 | jonatasfreitasv | Is the same version in all:
### Shard 1
ClickHouse client version 1.1.54362.
Connected to ClickHouse server version 1.1.54362.
### Replica 1
ClickHouse client version 1.1.54362.
Connected to ClickHouse server version 1.1.54362.
### Shard 2
ClickHouse client version 1.1.54362.
Connected to ClickHouse serv... | 2018-04-05T14:49:06 | 2018-04-05T14:49:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 379,002,718 | 2,180 | daskol | > When we add any new build - wait for test result increased by hours.
I mean for release candidates or stable releases.
> clickhouse-client throws exception double free or corruption
I rebuild Clickhouse client with `RelWithDebInfo` but it turns out that debug symbols are useless since it looks like exception... | 2018-04-05T16:51:07 | 2018-04-05T16:51:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 379,117,438 | 1,981 | alexey-milovidov | `ADDRESS_FAMILIES` means that this is related to IPv4 vs IPv6 addresses in configuration.
Typical Linux system have support both for IPv4 and IPv6 and you can use `::` wildcard address in config to listen.
But sometimes (in some cloud providers) they have a kernel with explicitly disabled IPv4 or IPv6.
In this c... | 2018-04-06T00:46:06 | 2018-04-06T00:46:06 | {} | MEMBER |
ClickHouse | ClickHouse | 379,182,829 | 2,180 | daskol | I have built dependency graph for target `clickhouse-client` with `--graphviz` CMake option. I wonder why client depends on heavy compiled full-featured parsers `clickhouse_functions`, `clickhouse_table_functions`, `clickhouse_aggregate_functions`.
![image](https://user-images.githubusercontent.com/9336514/38408769-... | 2018-04-06T08:18:43 | 2018-04-06T08:18:43 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 379,429,310 | 2,101 | alexey-milovidov | Fixed in different PR. | 2018-04-07T03:41:55 | 2018-04-07T03:41:55 | {} | MEMBER |
ClickHouse | ClickHouse | 379,500,654 | 2,086 | alexey-milovidov | OPTIMIZE FINAL is already doing re-merge even in case of single part (as it is documented).
Initially, it was intended for GraphiteMergeTree. But for other MergeTrees, it was read and rewrite all data on disk, doing no transforms. This case for OPTIMIZE FINAL was almost useless (actually it is useful to materialize ... | 2018-04-07T21:30:51 | 2018-04-07T21:31:50 | {} | MEMBER |
ClickHouse | ClickHouse | 379,501,896 | 2,086 | filimonov | I'm not sure if i understand and remember that properly, but i think on telegram chat somebody mentioned that they use ReplacingMergeTree and regularly fire OPTIMIZE FINAL (in cron or something like that), to have the state of table consistent. That's not too smart, but could work. After that change that will make huge... | 2018-04-07T21:53:42 | 2018-04-07T21:53:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,627,022 | 2,155 | tobad357 | I think this may fix #1793 which is that distributed ddl doesn't work if using nat interserver_http_host | 2018-04-04T14:48:19 | 2018-04-04T14:48:19 | {} | NONE |
ClickHouse | ClickHouse | 378,924,353 | 2,178 | jonatasfreitasv | Yes, but when i try this:
`DROP TABLE rt_analytics.activations ON CLUSTER cluster`
I receive this message:
`Code: 49, e.displayText() = DB::Exception: Query id cannot be empty, e.what() = DB::Exception` | 2018-04-05T12:49:23 | 2018-04-05T12:49:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 379,115,936 | 2,112 | alexey-milovidov | Although we can cast both sides to Float32 when comparing Float32 with Float64,
it will introduce another troubles, because:
1. It will differ from the behaviour of C++ (now it is consistent).
2. Some results will be wrong, for example:
```
SELECT CAST(inf AS Float32) = 1e100
```
now it returns 0 as expected.
... | 2018-04-06T00:34:17 | 2018-04-06T00:34:17 | {} | MEMBER |
ClickHouse | ClickHouse | 379,117,671 | 1,981 | alexey-milovidov | > I saw this error when I changed the location of the logs
Probably you have different error, could you please show it? | 2018-04-06T00:47:26 | 2018-04-06T00:47:26 | {} | MEMBER |
ClickHouse | ClickHouse | 379,149,953 | 1,846 | alexey-milovidov | Fixed in master. | 2018-04-06T05:13:08 | 2018-04-06T05:13:08 | {} | MEMBER |
ClickHouse | ClickHouse | 379,425,897 | 2,180 | alexey-milovidov | ClickHouse client is dependent on aggregate functions to allow read binary data with serialized aggregation states from the server (columns with type `AggregateFunction(...)`). This is useful, for example, to create data dumps of tables that have columns of such kind. But we can introduce a flag to disable this possibi... | 2018-04-07T02:26:55 | 2018-04-07T02:27:28 | {} | MEMBER |
ClickHouse | ClickHouse | 378,576,757 | 1,770 | alexm93 | То есть RowBinary форматом? А строки сами кодируете в нужный формат? Или может есть какой то encoder готовый? | 2018-04-04T12:05:33 | 2018-04-04T12:05:33 | {} | NONE |
ClickHouse | ClickHouse | 378,847,675 | 2,163 | vavrusa | @tobad357 any chance you could try investigate? It seems to be throwing an error, but maybe something needs to be checked in the readSuffix() before committing. https://github.com/yandex/ClickHouse/blob/master/dbms/src/Storages/StorageKafka.cpp#L185 | 2018-04-05T07:40:52 | 2018-04-05T07:40:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,986,103 | 2,181 | den-crane | https://github.com/yandex/ClickHouse/issues/2170
https://github.com/yandex/ClickHouse/pull/2173 | 2018-04-05T15:58:06 | 2018-04-05T15:58:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 379,066,561 | 1,282 | alexey-milovidov | Yes.
I've added a test just in case. | 2018-04-05T20:29:38 | 2018-04-05T20:29:38 | {} | MEMBER |
ClickHouse | ClickHouse | 379,114,452 | 2,113 | alexey-milovidov | Long time ago there was no higher order functions, and we have implemented "special" handling for arrays in the left hand side of IN:
`[elems...] IN (set)` means "any of elems is in set"
This special behaviour is deprecated.
Now it could be written more naturally, as `arrayExists(x -> x IN (set), arr)`
Recent... | 2018-04-06T00:23:25 | 2018-04-06T00:23:38 | {} | MEMBER |
ClickHouse | ClickHouse | 379,116,980 | 2,010 | alexey-milovidov | Not yet, I had almost fixed this but not finished. | 2018-04-06T00:42:35 | 2018-04-06T00:42:49 | {} | MEMBER |
ClickHouse | ClickHouse | 379,704,473 | 525 | battulgag | when support datetime with millisecond? | 2018-04-09T10:17:12 | 2018-04-09T10:17:12 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 378,943,113 | 2,180 | daskol | Click [here](https://pastebin.com/JKSJHdym). | 2018-04-05T13:51:11 | 2018-04-05T13:51:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 379,768,398 | 2,187 | goranc | Thanks for info about stability.
We should take care of production systems testing results, maybe on Altinity web site.
g.c. | 2018-04-09T14:19:03 | 2018-04-09T14:19:03 | {} | NONE |
ClickHouse | ClickHouse | 378,586,056 | 1,770 | deonwise | На golang драйвер передает, я использую tcp подключение, не http | 2018-04-04T12:42:00 | 2018-04-04T12:42:00 | {} | NONE |
ClickHouse | ClickHouse | 378,630,758 | 2,155 | jpiper | This PR is garbage though, so it won’t work :p but something similar should be implemented if we want CH to work behind NAT (or on clickhouse). I’ve had to resort to using hostNetwork for the timebeing | 2018-04-04T14:58:07 | 2018-04-04T14:58:07 | {} | NONE |
ClickHouse | ClickHouse | 379,079,713 | 2,109 | alexey-milovidov | Look at `graphite` sections in `config.xml`:
```
<!-- Sending data to Graphite for monitoring. Several sections can be defined. -->
<!--
interval - send every X second
root_path - prefix for keys
hostname_in_path - append hostname to root_path (default = true)
metrics - send... | 2018-04-05T21:17:32 | 2018-04-05T21:17:32 | {} | MEMBER |
ClickHouse | ClickHouse | 379,704,580 | 57 | battulgag | when support datetime with millisecond? | 2018-04-09T10:17:40 | 2018-04-09T10:17:40 | {
"+1": 15
} | NONE |
ClickHouse | ClickHouse | 380,420,441 | 2,207 | ludv1x | You should add ```<internal_replication>true</internal_replication>``` to the cluster definition. Like that:
https://github.com/yandex/ClickHouse/blob/bf1c4d156a72f915a78428dbc8c5493eaba67db2/dbms/tests/integration/test_distributed_ddl/configs/config.d/clusters.xml#L7 | 2018-04-11T11:33:42 | 2018-04-11T11:34:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,641,056 | 2,155 | derekperkins | @jpiper Other than this issue, has running ClickHouse on Kubernetes been a good experience for you? | 2018-04-04T15:25:38 | 2018-04-04T15:25:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,931,140 | 936 | kiruxan | @dolbyzerr I've got the same error, one guy in the CH Telegram channel told me that ALTER TABLE ON CLUSTER is not yet supported for Distributed tables. It looks for me that recreating Distributed table is the easiest way in cases when you need to alter it. | 2018-04-05T13:13:30 | 2018-04-05T13:13:30 | {} | NONE |
ClickHouse | ClickHouse | 378,941,524 | 2,180 | proller | fix second error: cmake -DNO_WERROR=1 | 2018-04-05T13:46:24 | 2018-04-05T13:47:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 379,070,480 | 797 | alexey-milovidov | Fixed long time ago.
Added a test just in case. | 2018-04-05T20:43:40 | 2018-04-05T20:43:40 | {} | MEMBER |
ClickHouse | ClickHouse | 379,114,772 | 2,113 | alexey-milovidov | 6452a3bb768712505b2cd74390cb0727fc29610e
30dd11f8caf9eae20926335f16231132424f6f7f | 2018-04-06T00:25:46 | 2018-04-06T00:25:46 | {} | MEMBER |
ClickHouse | ClickHouse | 379,296,337 | 2,180 | filimonov | currently (and historically) clickhouse-client is rather fat client tightly integrated with server code. And there is only one query parser used both on client and server.
It has some advantages (DRY principle), and some drawbacks (client is fat). As I know changing that is not the part of (nearest) plans of Yandex... | 2018-04-06T15:52:25 | 2018-04-06T15:57:53 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 379,901,377 | 360 | buyology | I've been trying hard to build on High Sierra but keeps hitting that exact same error as above.
Has anyone done this successfully? Would love to hear what setup you have + what `cmake` cmd you are running :)
https://gist.github.com/buyology/07c314992496c14dab18f997ef02bd76
```bash
--- clickhouse/build ‹mast... | 2018-04-09T21:35:02 | 2018-04-09T21:35:33 | {} | NONE |
ClickHouse | ClickHouse | 378,647,884 | 2,155 | jpiper | Overall, we’re very pleased with our setup. The configuration just requires a bit of effort to work around the fact that Clickhouse wasn’t designed to be deployed this way (and it depends what your SLA neess to be - some people might be less fussy when it comes to failure domains etc and would be happy not needing to a... | 2018-04-04T15:45:12 | 2018-04-04T15:45:12 | {} | NONE |
ClickHouse | ClickHouse | 378,781,627 | 2,155 | jpiper | The problem with local PVC is that you still don’t know about the placement before you schedule so configuring your sharding schema before you deploy isn’t possible in order to spread replicas across failure domains. Local PVs are also in beta, and I’m wary of putting all my state into a beta feature!
| 2018-04-05T00:00:30 | 2018-04-05T00:00:30 | {} | NONE |
ClickHouse | ClickHouse | 378,845,360 | 2,165 | vavrusa | I didn't have any immediate plans, as the Kafka engine doesn't support writing yet. I have some plans to add support for that so that ClickHouse can be used as a stream processor reading/write to Kafka, but probably late April/May as I'm a bit busy with other work stuff. I can help reviewing/helping out with PRs though... | 2018-04-05T07:30:01 | 2018-04-05T07:30:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,915,097 | 2,161 | jonatasfreitasv | Guy's!!
I'm apply this architecture and work perfect!!!! | 2018-04-05T12:12:36 | 2018-04-05T12:12:36 | {
"+1": 12,
"heart": 2,
"hooray": 3,
"laugh": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,931,507 | 2,174 | tobad357 | Great news that works. so it solves my problem but is still a bit of a papercut.
| 2018-04-05T13:14:45 | 2018-04-05T13:14:45 | {} | NONE |
ClickHouse | ClickHouse | 379,063,248 | 2,176 | alexey-milovidov | Можно просто заменить восьмеричные escape-последовательности на hex: `\x01`, `\x02`. | 2018-04-05T20:17:38 | 2018-04-05T20:17:38 | {} | MEMBER |
ClickHouse | ClickHouse | 379,065,635 | 1,360 | alexey-milovidov | Resolved long time ago. | 2018-04-05T20:26:19 | 2018-04-05T20:26:19 | {} | MEMBER |
ClickHouse | ClickHouse | 379,448,056 | 2,178 | jonatasfreitasv | The last version for YUM is 1.1.54362 =/ | 2018-04-07T08:13:36 | 2018-04-07T08:13:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 379,883,196 | 2,195 | filimonov | A lot of options...
```sql
SELECT
client_id,
argMax(event, created_at),
max(created_at),
FROM table
GROUP BY client_id
-- or
SELECT
client_id,
anyLast(event),
anyLast(created_at)
FROM
(SELECT client_id,event, created_at FROM table ORDER BY created_at)
GROUP BY client_id
```
Also soluti... | 2018-04-09T20:27:00 | 2018-04-09T20:27:00 | {
"+1": 4,
"hooray": 2,
"laugh": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,578,480 | 2,161 | jonatasfreitasv | With Kafka, the final intention is to make real-time analytics/dashboard.
Nice @filimonov thx a lot guy!!!!!
And the last question, for now, hehehe.
I'm going to build a RESTapi to deliver the data to the front end, run the querys in the database and deliver the result.
As I'm going to have two servers in c... | 2018-04-04T12:12:40 | 2018-04-04T12:14:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,669,711 | 2,161 | filimonov | If you create Distibuted table on both servers - you can connect to any of them and it will get data from the clickhouse cluster behind.
If one server will fall you need to switch to another server.
You can use some proxy in front of clickhouse (there is also nice proxy especially Clickhouse https://github.com/Verta... | 2018-04-04T16:52:48 | 2018-04-04T16:52:48 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,713,858 | 2,147 | den-crane | seems first part of issue (0 rows) fixed in ClickHouse server version 1.1.54373.
But weird issue still persists
```
drop table trololo;
CREATE TABLE trololo (a date, b Int, c Int) ENGINE = MergeTree order by (a,b);
INSERT INTO trololo VALUES (0, 1, 3);
SELECT * FROM trololo WHERE (a, b) IN (SELECT (a, b... | 2018-04-04T19:15:06 | 2018-04-04T19:15:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,941,881 | 2,180 | proller | can you show full cmake output ? | 2018-04-05T13:47:30 | 2018-04-05T13:47:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 379,189,788 | 2,169 | ennio1991 | Hi @vavrusa I didn't find any info about the SETTINGS syntax, can you give me more info?
```
CREATE TABLE tests.games_transactions (
day Date,
UserId UInt32,
Amount Float32,
CurrencyId UInt8,
timevalue DateTime,
ActivityType UInt8
)
ENGINE = MergeTree(day, (day, UserId), 8192)
SETTINGS max_block_size=10
... | 2018-04-06T08:48:20 | 2018-04-06T08:48:28 | {} | NONE |
ClickHouse | ClickHouse | 379,534,026 | 2,178 | lamberken | @jonatasfreitasv, you can use docker https://hub.docker.com/r/yandex/clickhouse-server/ | 2018-04-08T09:07:56 | 2018-04-08T09:07:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 380,088,733 | 2,199 | fahrettinf89 | @jonatasfreitasv I was inserting to Distributed data already as I mention above. Do you mean "I have to have separate Distributed tables on both remote and localhost?" | 2018-04-10T12:58:03 | 2018-04-10T12:58:03 | {} | NONE |
ClickHouse | ClickHouse | 378,566,631 | 1,770 | deonwise | По моему при вставке можно просто строку указать, без UUIDStringToNum. Но это не точно | 2018-04-04T11:22:10 | 2018-04-04T11:22:32 | {} | NONE |
ClickHouse | ClickHouse | 378,575,300 | 268 | imissyouso | @Slach `Error response from daemon: pull access denied for yandex/clickhouse-builder, repository does not exist or may require 'docker login'.` | 2018-04-04T11:59:58 | 2018-04-04T12:00:25 | {} | NONE |
ClickHouse | ClickHouse | 378,641,308 | 2,170 | den-crane | Probably this query causes a segfault:
SELECT count(*) FROM testSegmtFault PREWHERE id1 IN (1) and key1 = 0 | 2018-04-04T15:26:22 | 2018-04-04T15:26:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,641,600 | 2,155 | tobad357 | @derekperkins we run it on Kubernetes without issues
Instead of using hostnetwork we created some python scripts that can run DDL on all servers in a cluster
You point it to a node and it will query system table to get all the other nodes and then run the sql | 2018-04-04T15:27:22 | 2018-04-04T15:27:22 | {} | NONE |
ClickHouse | ClickHouse | 378,670,203 | 2,155 | derekperkins | Thanks for the feedback. With 1.10 supporting local PVC, I think a lot of your setup workarounds can be done more Kubernetes-y which could help you feel better about it. :)
I had to do a few similar un-k8s workarounds when trying to get GitHub Orchestrator running on Kubernetes, using a single service per StatefulSe... | 2018-04-04T16:54:29 | 2018-04-04T17:08:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 379,116,752 | 2,066 | alexey-milovidov | I've reproduced similar issue with:
```
USE test;
DROP TABLE IF EXISTS nested;
CREATE TABLE nested
(
column Nested
(
name String,
names Array(String),
types Array(Enum8('PU' = 1, 'US' = 2, 'OTHER' = 3))
)
) ENGINE = MergeTree ORDER BY tuple();
INSERT INTO nested VA... | 2018-04-06T00:40:45 | 2018-04-06T00:40:45 | {} | MEMBER |
ClickHouse | ClickHouse | 379,191,613 | 2,169 | ennio1991 | @javisantana
I have a single m4.xlarge machine on aws.
Here the structure:
```
CREATE TABLE games (
UserId UInt32,
ActivityType UInt8,
Amount Float32,
CurrencyId UInt8,
Date String
) ENGINE = Kafka('XXXX.eu-west-1.compute.amazonaws.com:9092,XXXX.eu-west-1.compute.amazonaws.com:909... | 2018-04-06T08:55:48 | 2018-04-06T08:55:48 | {} | NONE |
ClickHouse | ClickHouse | 378,631,847 | 2,155 | jpiper | On on kubernetes* | 2018-04-04T15:00:51 | 2018-04-04T15:00:51 | {} | NONE |
ClickHouse | ClickHouse | 378,846,114 | 2,174 | vavrusa | Does it work when you attach the second MV to `foo_mv`? e.g.
```CREATE MATERIALIZED VIEW IF NOT EXISTS foo_2_mv TO foo_2 AS SELECT date, id FROM foo_mv```
I think I've encountered that as well recently, the behavior must have changed. | 2018-04-05T07:33:45 | 2018-04-05T07:33:45 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 380,381,481 | 2,205 | filimonov | Related to https://github.com/yandex/ClickHouse/issues/2039 | 2018-04-11T09:04:46 | 2018-04-11T09:04:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 379,085,287 | 2,002 | alexey-milovidov | This issue isn't present in recent versions. | 2018-04-05T21:40:17 | 2018-04-05T21:40:17 | {} | MEMBER |
ClickHouse | ClickHouse | 379,131,832 | 1,981 | derekperkins | I figured it out, it was just a permissions error, but it also shows that error if the directory doesn't exist. I had to run `chown clickhouse:clickhouse` to fix it. | 2018-04-06T02:33:55 | 2018-04-06T02:33:55 | {
"+1": 9
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 379,209,581 | 2,183 | meolu | Thank you : ) | 2018-04-06T10:07:25 | 2018-04-06T10:07:25 | {} | NONE |
ClickHouse | ClickHouse | 380,125,019 | 2,199 | filimonov | Your data is always inserted to one shard because you are using constant value as sharding expression (last parameter in Distributed table parameters). It should be some column name or function name based on which clickhouse will decide to each of the servers to send that data. For example you can use `random()` to spl... | 2018-04-10T14:39:49 | 2018-04-10T15:37:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,810,471 | 2,169 | jonatasfreitasv | Hi @ennio1991.
I'm too use Kafka Engine with MergeTree Engine in topic with a lot of messages per second, 10-20k, in real-time.
But have a lot of variables to get real-time performance.
Some questions:
What hardware with ClickHouse Server?
What is the MergeTree and Kafka Engine table struct?
How many ... | 2018-04-05T03:30:27 | 2018-04-05T03:31:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,874,050 | 2,170 | KochetovNicolai | https://github.com/yandex/ClickHouse/pull/2173 | 2018-04-05T09:23:29 | 2018-04-05T09:23:29 | {} | MEMBER |
ClickHouse | ClickHouse | 378,980,987 | 2,178 | lamberken | try version 1.1.54370.
I met the same error before, I use 1.1.54370 currently, it's ok | 2018-04-05T15:42:51 | 2018-04-05T15:44:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 379,069,034 | 1,260 | alexey-milovidov | Actually it works right now
(but I won't add this as a test case). | 2018-04-05T20:38:12 | 2018-04-05T20:38:12 | {} | MEMBER |
ClickHouse | ClickHouse | 379,428,071 | 2,086 | filimonov | Hmm... Wouldn't it lead to big performance downgrade in some scenarios? Nop operation will be changed to big remerge?
May be it will be safer to make separate syntax for that case? Like
```sql
OPTIMIZE ... FINAL REMERGE
-- or
OPTIMIZE ... FINAL SETTINGS optimize_single_part = 1
-- or something like that?
``` | 2018-04-07T03:14:08 | 2018-04-07T03:16:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 379,384,710 | 2,086 | alexey-milovidov | Recently FINAL (both SELECT with FINAL and OPTIMIZE with FINAL) wasn't doing any transformations while merging in the case of single data part.
This is considered Ok because it's expected that data modifications are inserted in separate INSERTions (so the data that should be replaced/collapsed/aggregated/etc is loc... | 2018-04-06T21:23:04 | 2018-04-07T21:32:18 | {
"+1": 6
} | MEMBER |
ClickHouse | ClickHouse | 379,507,380 | 2,086 | alexey-milovidov | OPTIMIZE FINAL is already doing full data rewrite at that moment.
Please check the commit. It only removes optimization of calculations inside merge. | 2018-04-07T23:49:59 | 2018-04-07T23:50:14 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 379,886,277 | 2,195 | jonatasfreitasv | Man, you are a God!!! Thx a lot!!! | 2018-04-09T20:37:54 | 2018-04-09T20:37:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 380,080,924 | 2,199 | jonatasfreitasv | Insert data into a Distributed table to insert in both shards. | 2018-04-10T12:29:48 | 2018-04-10T12:29:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,611,105 | 2,165 | tobad357 | It would be pretty great to be able to get CapnProto or other structured formats back from the DB
That way integration with any server side would be very straight forward | 2018-04-04T14:03:58 | 2018-04-04T14:03:58 | {} | NONE |
ClickHouse | ClickHouse | 378,949,496 | 2,178 | lamberken | hi，it needs the version of server and client are same | 2018-04-05T14:10:07 | 2018-04-05T14:10:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 379,115,240 | 2,113 | alexey-milovidov | Fixed in master.
Added your example as a test case. | 2018-04-06T00:29:14 | 2018-04-06T00:29:14 | {} | MEMBER |
ClickHouse | ClickHouse | 379,458,457 | 2,171 | linuxgood1230 | input_format_allow_errors_ratio
input_format_allow_errors_num
Do not effect StoreKafka | 2018-04-07T10:06:12 | 2018-04-07T10:06:12 | {} | NONE |
ClickHouse | ClickHouse | 380,086,043 | 360 | silviucpp | Use
```sh
cmake .. -DCMAKE_CXX_COMPILER=`which g++-7` -DCMAKE_C_COMPILER=`which gcc-7` -DCMAKE_OSX_DEPLOYMENT_TARGET=10.12 -DCMAKE_CXX_FLAGS="-I/usr/local/include" -DCMAKE_BUILD_TYPE=Release
```
Or look where that file is located on your pc:
```sh
locate sqlext.h
/usr/local/Cellar/unixodbc/2.3.5_1/include... | 2018-04-10T12:48:22 | 2018-04-10T12:48:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,645,551 | 2,155 | jpiper | It’s not that easy if you want to use the replicated tables as you have to specify the sharding schema up front with the hostnames of each node in the cluster. This is fine if you’re happy using a stateful set with persistent volumes claims but I want more control as I need replicas to be spread across failure domains ... | 2018-04-04T15:38:46 | 2018-04-04T15:38:46 | {} | NONE |
ClickHouse | ClickHouse | 378,711,543 | 2,170 | den-crane | Same issue in master / commit e203457d5547afd2e1c4934fe79b3df7d74e9fd6
Received exception from server (version 1.1.54373):
Code: 49. DB::Exception: Received from localhost:9000, ::1. DB::Exception: Unexpected stack size in PKCondition::mayBeTrueInRange. | 2018-04-04T19:07:25 | 2018-04-04T19:07:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,923,547 | 2,178 | alex-zaitsev | Look at 'ON CLUSTER' syntax.
https://clickhouse.yandex/docs/en/single/index.html#distributed-ddl-queries-on-cluster-clause | 2018-04-05T12:46:27 | 2018-04-05T12:46:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,944,515 | 2,180 | daskol | @proller Do you plan make build matrix for `clickhouse-client` in Travis CI. | 2018-04-05T13:55:24 | 2018-04-05T14:15:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,999,888 | 2,183 | filimonov | Generally Clickhouse need some memory to process data in efficient way. So if you have big dataset, it's better to have lot of ram - it need to store indexes, metadata etc. If you will limit the memory too much clickhouse will not be 100% efficient. So just add more RAM, it doesn't cost much nowdays.
Check that reco... | 2018-04-05T16:41:29 | 2018-04-05T16:42:39 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 379,474,540 | 2,187 | alex-zaitsev | We are waiting for the latest stable version. 1.1.54370 is not stable -- too many bugs. | 2018-04-07T14:46:19 | 2018-04-07T14:46:32 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 379,800,854 | 2,141 | ludv1x | > Ah, didn't even know that ClickHouse had DNS caching. Would be useful to disable this in situations where the active endpoint for a service (e.g. an external dictionaries endpoint) is managed via DNS with a short TTL.
In the next release, it will be possible to set `<disable_internal_dns_cache>1</disable_internal_... | 2018-04-09T15:53:04 | 2018-04-09T15:54:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 380,037,471 | 2,164 | decaseal | **Changes in StorageFile**
Сreating a table from a file in the current database directory now allowed. | 2018-04-10T09:36:27 | 2018-04-10T09:36:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 380,096,319 | 1,722 | silviucpp | Branch merged with master | 2018-04-10T13:23:36 | 2018-04-10T13:23:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 378,656,966 | 2,171 | linuxgood1230 | Found It
input_format_allow_errors_ratio | 2018-04-04T16:11:11 | 2018-04-04T16:11:11 | {} | NONE |
ClickHouse | ClickHouse | 378,671,871 | 2,161 | jonatasfreitasv | I'll probably use direct calls to the servers using some resilient http lib.
Thx a looooot again @filimonov!!!! | 2018-04-04T16:59:53 | 2018-04-04T16:59:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 379,510,300 | 2,163 | tobad357 | Will try to investigate.
My C++ chops isn't the greatest, as haven't used it in 18 years | 2018-04-08T00:57:25 | 2018-04-08T00:57:25 | {} | NONE |
ClickHouse | ClickHouse | 379,562,004 | 2,188 | sleptor | Probably, this issue connected to #2146 | 2018-04-08T16:13:27 | 2018-04-08T16:13:27 | {} | NONE |
ClickHouse | ClickHouse | 380,344,736 | 2,199 | fahrettinf89 | @filimonov I create my Distributed table with following syntax and it seems work;
`create table d_y as boncuk ENGINE=Distributed(mavi,default,boncuk,rand())`
Thx.
But I think sharding key part in documentation should be more detailed. I would never understand the problem if I didn't ask this question.
Thank yo... | 2018-04-11T06:43:04 | 2018-04-11T06:43:20 | {} | NONE |
ClickHouse | ClickHouse | 380,377,770 | 2,205 | jpiper | This looks like a more general problem that the result of toDateTime(0) behaves badly before the epoch
```
:) SELECT toDateTime(0)+100, toDateTime(0) -1
SELECT
toDateTime(0) + 100,
toDateTime(0) - 1
┌─plus(toDateTime(0), 100)─┬─minus(toDateTime(0), 1)─┐
│ 1970-01-01 00:01:40 │ 1970-01-01... | 2018-04-11T08:52:16 | 2018-04-11T08:52:16 | {} | NONE |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.