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 | 792,303,071 | 21,406 | vzakaznikov | @hexiaoting, a similar problem with FixedString keys.
```
echo -e "SELECT m['1'] AS m FROM table_map_select_key_with_key_string_19c4a804_7f5f_11eb_9322_0173346139d7 FORMAT JSONEachRow" | clickhouse client -n --allow_experimental_map_type "1"
3ms ... | 2021-03-07T16:12:10 | 2021-03-07T16:12:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 792,419,053 | 21,209 | cyberhuman | any news here? | 2021-03-08T02:36:26 | 2021-03-08T02:36:26 | {} | NONE |
ClickHouse | ClickHouse | 792,499,191 | 21,473 | Zhile | Thanks @den-crane , that's really helpful! | 2021-03-08T06:18:23 | 2021-03-08T06:18:23 | {} | NONE |
ClickHouse | ClickHouse | 791,610,441 | 20,477 | akuzm | > If we decide that we want to introduce the optimization barrier and have a proper separation between the primary and secondary query, can we use the remote query executor, but in a zero-copy fashion (just pass the blocks in memory)?
I mean, we could make `createLocalPlan` start a proper secondary query, but not ma... | 2021-03-05T18:46:28 | 2021-03-05T18:46:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 791,940,675 | 21,425 | alesapin | I've added sleeps and really, kazoo just ignore heart beats:
```
2021.03.06 16:28:44.745204 [ 13 ] {} <Trace> NuKeeperTCPHandlerFactory: NuKeeper request. Address: 192.168.144.1:35340
2021.03.06 16:28:44.745912 [ 13 ] {} <Information> NuKeeperTCPHandler: Requesting session ID for the new client
2021.03.06 16:28:44.... | 2021-03-06T13:30:43 | 2021-03-06T13:30:43 | {} | MEMBER |
ClickHouse | ClickHouse | 791,971,610 | 21,491 | den-crane | Yes, this is the known issue. Tableau does not support Clickhouse SQL. | 2021-03-06T15:27:29 | 2021-03-06T15:27:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 792,225,408 | 3,474 | alexey-milovidov | 1 is implemented with:
`EXPLAIN AST your query` | 2021-03-07T06:22:17 | 2021-03-07T06:22:17 | {} | MEMBER |
ClickHouse | ClickHouse | 792,225,769 | 7,047 | alexey-milovidov | This task is trivial but no one is interested. | 2021-03-07T06:25:26 | 2021-03-07T06:25:26 | {} | MEMBER |
ClickHouse | ClickHouse | 792,293,496 | 14,995 | traceon | @vzakaznikov will it fix the failures or the hanged runs? | 2021-03-07T15:06:16 | 2021-03-07T15:06:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 791,674,041 | 21,136 | chenziliang | we have implemented something to stream back the progress status (bytes / row scan etc) via HTTP interface by using chunk format before sending the real query result. Clients need do some special parsing to decide if it is progress response or query result. It works pretty fine for us so far. If this solution sounds go... | 2021-03-05T20:38:45 | 2021-03-05T20:40:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 791,897,012 | 21,436 | azat | >Integration tests (asan) — fail: 2, passed: 1156, error: 1
>Integration tests (release) — fail: 1, passed: 1157, error: 0
#21461
>Integration tests (thread) — fail: 1, passed: 1040, error: 24
#21452
>Performance — 7 errors, 6 faster, 4 slower, 57 unstable
#21447
>Marker check Pending — Waiting to... | 2021-03-06T08:43:35 | 2021-03-06T08:43:52 | {} | MEMBER |
ClickHouse | ClickHouse | 791,939,978 | 21,425 | alesapin | Hmmm, looks like a bug.... But I don't understand where and cannot reproduce it. For some reason when the follower node down leader tries to send a 0-level snapshot (fake) with each `append_entries` call. Possibly it's OK, but 100% shouldn't affect `append_entries` performance.
Also, for some reason kazoo client doe... | 2021-03-06T13:26:30 | 2021-03-06T13:26:30 | {} | MEMBER |
ClickHouse | ClickHouse | 792,284,017 | 21,425 | alesapin | Stress test:
```
2021.03.07 12:53:55.398258 [ 296 ] {} <Debug> Application: Received termination signal.
2021.03.07 12:53:55.398366 [ 296 ] {} <Debug> Application: Waiting for current connections to close.
<...>
2021.03.07 12:54:01.288076 [ 296 ] {} <Information> Application: Closed connections. But 9 remain. Tip:... | 2021-03-07T14:01:05 | 2021-03-07T14:23:38 | {} | MEMBER |
ClickHouse | ClickHouse | 792,350,899 | 21,499 | alexey-milovidov | Experiment has proved that gcc's full LTO has no advantage over clang's ThinLTO. | 2021-03-07T20:52:57 | 2021-03-07T20:52:57 | {} | MEMBER |
ClickHouse | ClickHouse | 792,311,132 | 21,502 | phamtai97 | I am sory, data_time is date_time field.
Time is int64 type, date_time is DateTime type and date_time = time/1000.
I use field date_time to partition for table. | 2021-03-07T17:00:24 | 2021-03-07T23:56:21 | {} | NONE |
ClickHouse | ClickHouse | 792,458,298 | 21,468 | ghost | @alesapin
> java.io.IOException: Len error _more_than_one_megabyte_
No log likes `_more_than_one_megabyte_` or "1 MB".
@filimonov
ZooKeeper Exceptions:
```
BadVersion
Got user-level KeeperException when processing sessionid:0x509a345825207e0 type:multi cxid:0x6e9a6 zxid:0xa2b7585e7 txntype:-1 reqpat... | 2021-03-08T04:44:59 | 2021-03-08T05:15:54 | {} | NONE |
ClickHouse | ClickHouse | 792,523,757 | 17,623 | kitaisreal | @filimonov
> * Dictionary improvements
>
> * access all dictionary attributes with a single call (return named tuple?) - avoiding several dictGet calls to get different attributes in the same query (I guess can be useful for `direct`, but even accessing cache several times may be quite expensive comparing to s... | 2021-03-08T07:04:30 | 2021-03-08T07:04:30 | {
"hooray": 1,
"rocket": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 791,491,466 | 19,303 | qoega | https://clickhouse-test-reports.s3.yandex.net/0/9a9d5a74ba901ee65b330f47423bbb019d82b2c2/fuzzer_debug/report.html#fail1 | 2021-03-05T15:29:18 | 2021-03-05T15:29:18 | {} | MEMBER |
ClickHouse | ClickHouse | 791,891,552 | 20,608 | azat | >Only crash I found is clickhouse-client on asan servers (cannot reproduce on other versions):
Actually you need only client with ASAN, but since there is one binary for everything, it does not matter.
> when I just type in client /*. it crashes on . letter
Basically any char after `/*` will trigger it, should... | 2021-03-06T07:54:01 | 2021-03-06T07:54:01 | {} | MEMBER |
ClickHouse | ClickHouse | 792,303,820 | 21,406 | vzakaznikov | @hexiaoting, similar issue with Nullable.
```
Mar 07,2021 11:15:13 ⟥ By executing command
echo -e "SELECT m['1'] AS m FROM table_map_select_key_with_key_string_4b750155_7f60_11eb_9322_0173346139d7 FORMAT JSONEachRow" | clickhouse client -n --allow_experimental_ma... | 2021-03-07T16:17:08 | 2021-03-07T16:17:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 792,359,083 | 14,995 | vzakaznikov | > @vzakaznikov will it fix the failures or the hanged runs?
It will not fix the timeout. The fix only applies to the failing LDAP authentication performance test. | 2021-03-07T21:47:07 | 2021-03-07T21:47:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 792,400,866 | 21,220 | alexey-milovidov | It did not. | 2021-03-08T01:25:26 | 2021-03-08T01:25:26 | {} | MEMBER |
ClickHouse | ClickHouse | 791,808,559 | 21,136 | traceon | > Maybe implement WebSockets protocol?
Boost has an implementation, by the way. On top of Boost.Asio. It has HTTP implementation too. | 2021-03-06T00:22:07 | 2021-03-06T00:22:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 791,850,134 | 21,443 | ywxkz | CREATE USER xxx IDENTIFIED WITH ldap_server BY 'my_ldap_server';
If I change uid={user_... | 2021-03-06T03:04:02 | 2021-03-06T03:04:02 | {} | NONE |
ClickHouse | ClickHouse | 792,288,796 | 21,425 | alesapin | This replica was created for 5 minutes before the shutdown.
```
2021.03.07 12:50:50.216462 [ 466338 ] {85c2dd98-5b7d-4cee-a65b-36a7a777f362} <Trace> ContextAccess (default): Access granted: CREATE TABLE ON test_25.quorum1
2021.03.07 1... | 2021-03-07T14:34:13 | 2021-03-07T14:34:36 | {} | MEMBER |
ClickHouse | ClickHouse | 792,311,872 | 21,503 | phamtai97 | I try query this SQL on superset tool on Mac PC, the result is not change. | 2021-03-07T17:05:08 | 2021-03-07T17:05:08 | {} | NONE |
ClickHouse | ClickHouse | 792,372,924 | 21,502 | phamtai97 | > You'd better focus on PK index because you wasted it by placing high-cardinality column time to a leading position.
I do not understand your idea. Do you mean that you should use field times and PK fields to query instead of using other fields?
| 2021-03-07T23:19:08 | 2021-03-08T00:00:11 | {} | NONE |
ClickHouse | ClickHouse | 792,420,439 | 17,638 | zhicwu | It's an issue in JDBC driver(see ClickHouse/clickhouse-jdbc#590) and it will be fixed in 0.3.0. | 2021-03-08T02:40:48 | 2021-03-08T02:40:48 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 792,474,720 | 21,468 | ghost | @den-crane
@alesapin
I wonder why "Len error" occurs in v21.1.2.15 but not in v20.8.2.3.
Are there any changes to it? | 2021-03-08T05:28:48 | 2021-03-08T05:28:48 | {} | NONE |
ClickHouse | ClickHouse | 792,527,090 | 21,209 | cyberhuman | @kitaisreal thank you very much for your answer!
I'm really looking forward to see Nullable, arrays and enums being supported! | 2021-03-08T07:10:22 | 2021-03-08T07:10:22 | {} | NONE |
ClickHouse | ClickHouse | 791,445,142 | 21,468 | alesapin | Something like:
```
java.io.IOException: Len error _more_than_one_megabyte_
``` | 2021-03-05T14:14:51 | 2021-03-05T14:14:51 | {} | MEMBER |
ClickHouse | ClickHouse | 791,601,015 | 20,477 | akuzm | > Right, but interfaces for query execution via pipeline and via executeQueryImpl is different, that's why I've added this abstraction.
Ah, so the local query is embedded as a part of the pipeline and this is why we don't see it in the query log? Is it possible that we optimize the pipeline to the point where we can... | 2021-03-05T18:28:08 | 2021-03-05T18:28:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 791,915,270 | 14,995 | zvonand | > @zvonand so tests are failing?..
Yes, the LDAP tests, not Kerberos, I guess @vzakaznikov will know better | 2021-03-06T11:13:19 | 2021-03-06T11:13:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 791,918,306 | 21,443 | traceon | I see. What is the user name in the ClickHouse (the `xxx` placeholder)?
ClickHouse user name will be placed in `{user_name}` so they must match for bind to succeed. | 2021-03-06T11:24:10 | 2021-03-06T11:24:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 792,177,231 | 4,548 | zhicwu | ClickHouse JDBC driver uses HTTP interface. I didn't find a way to specify client/custom timezone in request, so I had to parse DateTime string returned from server(because we're using TabSeparatedWithNamesAndTypes at this point), and then re-format based using client/custom timezone... Would be great if we can add a s... | 2021-03-07T02:37:42 | 2021-03-07T02:37:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 792,285,586 | 21,503 | phamtai97 | I use lib:
```pom
<dependency>
<groupId>ru.yandex.clickhouse</groupId>
<artifactId>clickhouse-jdbc</artifactId>
</dependency>
``` | 2021-03-07T14:11:57 | 2021-03-07T14:11:57 | {} | NONE |
ClickHouse | ClickHouse | 792,389,784 | 21,220 | alexey-milovidov | Maybe rebase will help. | 2021-03-08T00:41:01 | 2021-03-08T00:41:01 | {} | MEMBER |
ClickHouse | ClickHouse | 792,470,713 | 21,468 | ghost | @den-crane
@alesapin
"Len error"s exists at that time. I will try #11933 (comment)
Thank you. | 2021-03-08T05:19:43 | 2021-03-08T05:21:47 | {} | NONE |
ClickHouse | ClickHouse | 791,592,137 | 20,477 | azat | >Do I get it right that the basic idea is to make queries that go to local replica look like remote queries in the query log? E.g. have their own query id and process list entry.
Absolutely correct.
> This is basically what executeQueryImpl would do if we passed internal = false to it, right?
Right, but inter... | 2021-03-05T18:11:32 | 2021-03-05T18:11:32 | {} | MEMBER |
ClickHouse | ClickHouse | 792,016,292 | 8,832 | alexey-milovidov | This issue is related to generic support for Nullable types in dictionaries, not just for ODBC.
It has been addressed here: https://github.com/ClickHouse/ClickHouse/pull/18236
To enable support for NULL, you should specify your data types as Nullable, e.g. Nullable(String). | 2021-03-06T18:43:26 | 2021-03-06T18:43:26 | {} | MEMBER |
ClickHouse | ClickHouse | 792,205,767 | 21,478 | amosbird | You can try dumping out the `column_names` here https://github.com/ClickHouse/ClickHouse/blob/master/src/Storages/StorageMergeTree.cpp#L178ckHouse/blob/master/src/Storages/StorageMergeTree.cpp#L178 . Columns listed in query_info might not be the actual access ones. | 2021-03-07T04:39:22 | 2021-03-07T04:39:52 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 792,225,452 | 3,474 | alexey-milovidov | 2 is not implemented. | 2021-03-07T06:22:37 | 2021-03-07T06:22:37 | {} | MEMBER |
ClickHouse | ClickHouse | 792,294,168 | 10,298 | Slach | @alexey-milovidov yep I think a separate function will better | 2021-03-07T15:10:47 | 2021-03-07T15:10:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 792,299,664 | 21,504 | den-crane | 1. Can you show 'show create table yourReplicatedReplacingMergeTree'
2. Can you show 'select * from system.macros'
3. Can you show 'select * from system.clusters' | 2021-03-07T15:49:07 | 2021-03-07T15:49:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 792,334,143 | 21,425 | alesapin | I have some bug with snapshots:
```
2021.03.07 21:45:48.326377 [ 72262 ] {} <Fatal> BaseDaemon: ########################################
2021.03.07 21:45:48.326433 [ 72262 ] {} <Fatal> BaseDaemon: (version 21.4.1.6189, build id: 9CA33DF56446D8AFAA017C202F16256573DB58C1) (from thread 72184) (no query) Received signal... | 2021-03-07T19:02:52 | 2021-03-07T19:02:52 | {} | MEMBER |
ClickHouse | ClickHouse | 792,371,773 | 21,502 | den-crane | Skip indexes are not inverted indexes like you used to with other databases. They can make queries slower , because skip index processing is hard task and it requires a lot of CPU.
Try different settings bloom_filter(0.01) (0.3) (0.5) GRANULARITY 4 - GRANULARITY 10.
But do not expect much. Skip indexes help in lim... | 2021-03-07T23:11:42 | 2021-03-07T23:13:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 792,356,064 | 21,502 | phamtai97 | But, i need query field1 or other fields such as field3 or field4 that is String type. Do I need to use the index for fields 3 and 4 ? Type index ?
Why did I use bloom filter for field1 the performance was not good ? | 2021-03-07T21:25:20 | 2021-03-07T23:15:27 | {} | NONE |
ClickHouse | ClickHouse | 792,460,961 | 21,468 | ghost | There are only one writer process for one shard and the `Writer-Process` always send `INSERT QUERY` to `ClickHouse Replica 1`.
```
insert_quorum = 1
(Writer-Process) --INSERT QUERY--> (ClickHouse Replica 1)
|
(Clic... | 2021-03-08T04:53:29 | 2021-03-08T05:16:34 | {} | NONE |
ClickHouse | ClickHouse | 792,525,225 | 21,209 | kitaisreal | Hello, @cyberhuman.
Currently we support:
1. UInt8, UInt16, UInt32, UInt64, UInt128
2. Int8, Int16, Int32, Int64
3. Float32, Float64, Decimal32, Decimal64, Decimal128
4. String
5. Nullable types are supported in flat dictionary, hashed, complex key hashed, direct, complex key direct
6. PolygonDictionary suppor... | 2021-03-08T07:07:04 | 2021-03-08T07:07:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 791,755,103 | 21,136 | alexey-milovidov | > Adding a separate mode/format/switch that will make ClickHouse to respond in HTTP with Server-sent events?
I imagined it as a framing format for server responses (activated by URL parameter).
But server sent events don't support binary data: https://stackoverflow.com/questions/33529685/using-sse-to-send-image-d... | 2021-03-05T22:28:45 | 2021-03-05T22:33:08 | {} | MEMBER |
ClickHouse | ClickHouse | 791,773,828 | 21,454 | alexey-milovidov | Not merged correctly. | 2021-03-05T23:00:14 | 2021-03-05T23:00:14 | {} | MEMBER |
ClickHouse | ClickHouse | 792,285,340 | 21,506 | vzakaznikov | Closing this PR as it is not based on master and includes changes that have not been merged. | 2021-03-07T14:10:02 | 2021-03-07T14:10:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 792,295,690 | 21,502 | den-crane | >toYYYYMM(data_time)
I don't see **data_time** column.
Why do you have 2 different time column date_time and time ?
| 2021-03-07T15:21:12 | 2021-03-07T15:21:12 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 792,371,608 | 21,503 | phamtai97 | Thank you, I understand this issue. | 2021-03-07T23:10:36 | 2021-03-07T23:11:07 | {} | NONE |
ClickHouse | ClickHouse | 791,603,991 | 20,477 | akuzm | If we decide that we want to introduce the optimization barrier and have a proper separation between the primary and secondary query, can we use the remote query executor, but in a zero-copy fashion (just pass the blocks in memory)? This patch looks overly complicated -- QueryProcess is a complex thing that does a lot ... | 2021-03-05T18:33:35 | 2021-03-05T18:33:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 791,684,307 | 21,382 | sevirov | @tavplubix Please, review this PR. | 2021-03-05T20:58:34 | 2021-03-05T20:58:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 791,914,350 | 14,995 | traceon | @zvonand so tests are failing?.. | 2021-03-06T11:11:54 | 2021-03-06T11:11:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 792,019,764 | 13,014 | alexey-milovidov | @abyss7 Still relevant in 21.3. | 2021-03-06T18:48:50 | 2021-03-06T18:48:50 | {} | MEMBER |
ClickHouse | ClickHouse | 792,313,045 | 21,503 | phamtai97 | So how the data is actually stored? Is it because lib JDBC has converted the time to the correct local time?
What SQL do I have to use to compare the date_time properly? | 2021-03-07T17:13:13 | 2021-03-07T17:13:13 | {} | NONE |
ClickHouse | ClickHouse | 792,370,824 | 5,582 | Felixoid | I've started some research for the issue, and I'm able to reproduce the issue quite quick:
- Precision 4
- Insert each metric every 2 seconds with values 1 and 3
```xml
<graphite_rollup_example>
<default>
<function>avg</function>
<retention>
<age>0</age>
... | 2021-03-07T23:05:11 | 2021-03-07T23:09:33 | {} | MEMBER |
ClickHouse | ClickHouse | 791,542,511 | 14,995 | traceon | @zvonand I don't see Kerberos test cases in the TestFlows run. Shouldn't they be there?
@vzakaznikov I saw that some LDAP tests are failing here, do you know what's the problem with them? Is it some pending adjustment that should be done to them in the scope of this PR or is it instability or a regression? | 2021-03-05T16:48:01 | 2021-03-05T16:48:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 791,556,850 | 14,995 | zvonand | > @zvonand I don't see Kerberos test cases in the TestFlows run. Shouldn't they be there?
It's okay, the tests are executed in same order they are defined in regression, if one fails, next ones are not executed | 2021-03-05T17:11:23 | 2021-03-05T17:11:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 792,285,292 | 21,503 | phamtai97 | The result of your SQL:
|timezone()| now() |toString(now())|
| ----------- | ----------- | ----------- |
|Asia/Ho_Chi_Minh | 2021-03-07T14:08:03+00:00 | 2021-03-07 21:08:03
| 2021-03-07T14:09:46 | 2021-03-07T14:10:58 | {} | NONE |
ClickHouse | ClickHouse | 792,311,616 | 21,503 | phamtai97 | I develop aplication by Java anh use lib Clickhouse JDBC. Then, I deploy app on Linux server. | 2021-03-07T17:03:29 | 2021-03-07T17:03:54 | {} | NONE |
ClickHouse | ClickHouse | 792,325,195 | 21,503 | den-crane | >So how the data is actually stored?
it depends on how you insert data.
>Is it because lib JDBC has converted the time to the correct local time?
yes. JDBC converts datetime to the local TZ. It should work OK if you run JAVA app at the server with Asia/Ho_Chi_Minh.
>What SQL do I have to use to compare the... | 2021-03-07T18:04:00 | 2021-03-07T18:04:00 | {
"heart": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 792,406,163 | 21,502 | phamtai97 | Ok, I agree with your idea.
So, I have many SQL queries that use different conditions (field1, field2, field3, ....), the above SQL statement is just one of them. But, most of my queries use the `time` field. I cannot leave all of those fields as primary keys. And I use PK in three fields: time, field1, field2 to pe... | 2021-03-08T01:48:08 | 2021-03-08T01:52:16 | {} | NONE |
ClickHouse | ClickHouse | 792,419,194 | 21,504 | 15537261738 | CREATE TABLE kafka_queue
(
`message` String
)
ENGINE = Kafka
SETTINGS kafka_broker_list = 'kafka-1:9092,kafka-2:9092,kafka-3:9092', kafka_topic_list = 'topic', kafka_group_name = 'clickhouse', kafka_format = 'TSVRaw'
CREATE MATERIALIZED VIEW kafka_consumer TO local
AS
SELECT
replace(replace(splitBy... | 2021-03-08T02:36:52 | 2021-03-08T02:36:52 | {} | NONE |
ClickHouse | ClickHouse | 792,430,904 | 16,240 | Sallery-X | > > How to solve the cache consistency problem of different replicas?
>
> In Clickhouse data is "append only". Sometimes entry part can be deleted. So could you please describe where you see a problem?
If multiple replicas share the same storage, how to resolve the share file confliction when the diffrent replic... | 2021-03-08T03:14:50 | 2021-03-08T03:14:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 792,439,797 | 21,502 | phamtai97 | After I remove the bloom filter index, should I run analyze SQL? | 2021-03-08T03:38:30 | 2021-03-08T03:38:30 | {} | NONE |
ClickHouse | ClickHouse | 791,468,667 | 21,456 | nikitamikhaylov | Ready to merge | 2021-03-05T14:53:25 | 2021-03-05T14:53:25 | {} | MEMBER |
ClickHouse | ClickHouse | 791,577,311 | 21,468 | den-crane | @alesapin but drop partition should not touch all partitions?
Len error / DJute https://github.com/ClickHouse/ClickHouse/issues/11933#issuecomment-661310032 | 2021-03-05T17:45:07 | 2021-03-05T17:45:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 791,596,621 | 21,432 | azat | >Integration tests (release) — fail: 1, passed: 1157, error: 0
>Integration tests (thread) — fail: 1, passed: 1150, error: 1
```
test_distributed_ddl/test.py::test_simple_alters[configs]
test_distributed_ddl/test.py::test_kill_query[configs]
```
Addressed in #21461
>Performance — 7 errors, 7 faster, 4 slo... | 2021-03-05T18:19:51 | 2021-03-05T18:19:51 | {} | MEMBER |
ClickHouse | ClickHouse | 791,473,750 | 21,476 | wuzedewuze | thanks
Maybe I needs try to build clickhouse without libunwind follow #20221
I‘m not faminiar with c++ , how to build clickhouse without libunwind ? | 2021-03-05T15:01:19 | 2021-03-06T01:26:15 | {} | NONE |
ClickHouse | ClickHouse | 791,873,240 | 20,477 | azat | >Ah, so the local query is embedded as a part of the pipeline and this is why we don't see it in the query log?
Yes
> Is it possible that we optimize the pipeline to the point where we can't tell where the primary query begins and the secondary starts?
It is not part of pipeline/plan actually, it attached sepa... | 2021-03-06T04:50:58 | 2021-03-06T04:50:58 | {} | MEMBER |
ClickHouse | ClickHouse | 791,965,031 | 21,481 | ucasfl | Should be fixed in #21498. | 2021-03-06T14:45:03 | 2021-03-06T14:45:03 | {
"heart": 1
} | COLLABORATOR |
ClickHouse | ClickHouse | 792,018,514 | 10,298 | alexey-milovidov | @Slach We can name it `decodeURLFormComponent`. | 2021-03-06T18:46:08 | 2021-03-06T18:46:08 | {} | MEMBER |
ClickHouse | ClickHouse | 792,067,758 | 20,182 | alexey-milovidov | Prestable version has been released. | 2021-03-06T21:45:39 | 2021-03-06T21:45:39 | {} | MEMBER |
ClickHouse | ClickHouse | 792,231,251 | 21,425 | alesapin | `test_materialize_mysql_database`
Not related | 2021-03-07T07:21:32 | 2021-03-07T07:21:32 | {} | MEMBER |
ClickHouse | ClickHouse | 792,295,376 | 21,503 | den-crane | What tool do use to query data using JDBC ? DataGrip?
What timezone at you local computer ? Windows/MACos -- Asia/Ho_Chi_Minh ? | 2021-03-07T15:19:00 | 2021-03-07T15:19:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 791,415,980 | 21,443 | traceon | Could you please post your LDAP config in ClickHouse here? | 2021-03-05T13:22:12 | 2021-03-05T14:39:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 791,949,542 | 21,425 | alesapin | Wow, seems like our zookeeper client sends redundant heartbeats. Each operation in ZK considered as a heartbeat and we don't need to send them by timeout if have other ops (https://github.com/apache/zookeeper/blob/master/zookeeper-server/src/main/java/org/apache/zookeeper/server/ZooKeeperServer.java#L1150). Need to cha... | 2021-03-06T13:54:28 | 2021-03-06T13:54:28 | {} | MEMBER |
ClickHouse | ClickHouse | 792,058,266 | 21,497 | olgarev | @vdimir Hi, will you review this PR, please?
Concerning the "Using S3 for Data Storage" section.
This is an "Edit and translate" ticket, so I've just translated the english text. But are you sure the section about S3 is in the right place? Seems that the next "Details" section (it has level h3) is not related to ... | 2021-03-06T21:32:41 | 2021-03-06T21:33:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 792,472,499 | 18,583 | alexey-milovidov | AMD Ryzen does not have `erms` and in fact `rep movsb` is working up to 3 times slower than unrolled aligned SSE2 mov operations if copying within L1 or L2 cache using 16 threads. | 2021-03-08T05:23:46 | 2021-03-08T05:23:46 | {} | MEMBER |
ClickHouse | ClickHouse | 792,285,113 | 14,995 | vzakaznikov | > > @zvonand so tests are failing?..
>
> Yes, the LDAP tests, not Kerberos, I guess @vzakaznikov will know better
Please apply fix from https://github.com/ClickHouse/ClickHouse/pull/21507
| 2021-03-07T14:08:38 | 2021-03-07T14:15:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 792,382,607 | 9,404 | alexey-milovidov | `toRelativeDayNum` should be extended from UInt16. | 2021-03-08T00:15:11 | 2021-03-08T00:15:11 | {} | MEMBER |
ClickHouse | ClickHouse | 792,422,721 | 21,512 | amosbird | Shard pruning can be improved after this pr https://github.com/ClickHouse/ClickHouse/pull/21401 . | 2021-03-08T02:49:05 | 2021-03-08T02:49:05 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 792,281,785 | 21,501 | den-crane | https://github.com/ClickHouse/ClickHouse/blob/master/docs/en/whats-new/changelog/2020.md#new-feature-3
this feature has been added in 20.9.2.20 | 2021-03-07T13:44:30 | 2021-03-07T13:44:30 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 792,359,588 | 21,499 | traceon | @alexey-milovidov not exactly. So far, this shows:
1. The way ThinLTO was originally enabled in cmake files is probably not entirely correct.
2. Clang's FullLTO takes indefinitely long (so far) to complete.
3. GCC's LTO doesn't build right now (some slight tuning of the build system is required, probably)
4. Clang's T... | 2021-03-07T21:50:56 | 2021-03-07T21:50:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 792,420,324 | 21,351 | fuwhu | @nikitamikhaylov @alexey-milovidov
hi,
it seems the test "Marker check" can not finish. any suggestion? | 2021-03-08T02:40:25 | 2021-03-08T02:40:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 791,511,496 | 21,425 | alesapin | Seems like after query:
```
SELECT
[257], ... | 2021-03-05T16:00:02 | 2021-03-05T16:00:24 | {} | MEMBER |
ClickHouse | ClickHouse | 791,844,131 | 21,464 | lcwxz1989 | @filimonov @den-crane hi, I have updated scenario and DDL info, thanks a lot | 2021-03-06T02:29:12 | 2021-03-06T02:29:12 | {} | NONE |
ClickHouse | ClickHouse | 792,131,034 | 9,404 | alexey-milovidov | `toStartOfDay` does not work properly:
```
SELECT toStartOfDay(toDateTime64(-1262285400, 0, 'Africa/Accra'))
``` | 2021-03-07T00:36:08 | 2021-03-07T00:36:08 | {} | MEMBER |
ClickHouse | ClickHouse | 792,326,593 | 21,502 | den-crane | You don't need ' date_time DateTime64' column
And you can use
```
PARTITION BY toYYYYMM(fromUnixTimestamp64Milli(time))
ORDER BY (toDate(fromUnixTimestamp64Milli(time)), field1, field2)
```
And remove bloom_filter skip indexes.
| 2021-03-07T18:12:36 | 2021-03-07T18:12:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 791,685,881 | 21,136 | traceon | > BTW, the status can be only streamed back before query result is produced. Once query result starts streaming back to HTTP client, no further status will be reported and obviously this feature is only turned on by using a setting flag.
That's the same limitation we want to work around right now. Currently, the pro... | 2021-03-05T21:00:52 | 2021-03-05T21:00:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 791,766,254 | 21,087 | alexey-milovidov | Writing SHALL in requirements to just document current behaviour looks wrong.
I recomment to replace it with MAY. | 2021-03-05T22:40:38 | 2021-03-05T22:40:38 | {} | MEMBER |
ClickHouse | ClickHouse | 791,782,360 | 19,257 | nikitamikhaylov | I had a talk with a teacher of geometry at Moscow State University and he told that there is no way to map a sphere to a plane map a sphere to a plane with the segments saved as is. Segments will map into straight lines or circular arcs always because of the curvature of the sphere.
That's why I will revert all my co... | 2021-03-05T23:13:44 | 2021-03-05T23:13:44 | {} | MEMBER |
ClickHouse | ClickHouse | 792,381,568 | 21,502 | den-crane | >ORDER BY (time, field1, field2)
This looks completely wrong for me.
The order of columns in PK and set of columns is the most important thing.
To develop the best PK need to understand the nature of the columns, and need to know the queries you are going to run.
Try `ORDER BY (field1, field2, time) ` `ORDER BY... | 2021-03-08T00:10:17 | 2021-03-08T00:10:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 768,239,036 | 19,705 | alexey-milovidov | > 2021.01.27 13:21:59.718256 [ 26522 ] {0a71f8b0-00e2-4c7a-9efd-430a2ca873f4} DynamicQueryHandler: Code: 24, e.displayText() = DB::Exception: Cannot write to ostream at offset 8388608: While executing TabSeparatedRowOutputFormat
Cannot send query result to the client. When sending data, client disconnected. | 2021-01-27T12:01:36 | 2021-01-27T12:01:36 | {} | MEMBER |
ClickHouse | ClickHouse | 768,381,021 | 19,518 | vdimir | I'm still not sure that `strcat` is good, maybe if previous version with string optimized enough it will perform not worse than current version.
I'm trying look how it is compiled and compare it here https://godbolt.org/z/96q77x
| 2021-01-27T15:52:14 | 2021-01-27T15:52:14 | {} | MEMBER |
ClickHouse | ClickHouse | 768,415,457 | 18,247 | SaltTan | @ghuname How many DDL tasks are stored in ZooKeeper?
```
SELECT count()
FROM system.zookeeper
WHERE path = '/clickhouse/task_queue/ddl/'
SELECT
ctime,
name,
value
FROM system.zookeeper
WHERE path = '/clickhouse/task_queue/ddl/'
ORDER BY ctime ASC
LIMIT 3
SELECT
ctime,
name,
v... | 2021-01-27T16:43:44 | 2021-01-27T16:43:44 | {} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.