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 | 485,612,499 | 5,014 | github1youlc | > I am not a ClickHouse expert (one of a recent user), but I have seen a scenario like this before.
>
> At that time, "XID overflow" was the direct cause of this issue.
> (Restarting the server will resolve the error (and this is what I did at that time), but some auto-recovery mechanism should be prepared against ... | 2019-04-23T01:54:23 | 2019-04-23T01:54:23 | {} | NONE |
ClickHouse | ClickHouse | 485,910,787 | 4,945 | alexey-milovidov | > I solve it by adding order by name,category,total when I read the link
the name is uniq, and the categgory/total will be the same value
Yes, this is the answer for this issue. | 2019-04-23T18:00:39 | 2019-04-23T18:00:39 | {} | MEMBER |
ClickHouse | ClickHouse | 485,919,364 | 5,071 | alexey-milovidov | Don't you think that predefined `Enum8('false' = 0, 'true' = 1)` as a synonim for Bool type can be easier and better solution? | 2019-04-23T18:23:51 | 2019-04-23T18:23:51 | {} | MEMBER |
ClickHouse | ClickHouse | 486,016,822 | 5,030 | alexey-milovidov | Need to use more loose compiler warning flags for compiling auto-generated file `tldLookup.cpp` | 2019-04-24T00:00:09 | 2019-04-24T00:00:09 | {} | MEMBER |
ClickHouse | ClickHouse | 486,134,728 | 5,003 | javisantana | Talking about geospatial index:
- I think [ZCurve](https://github.com/yandex/ClickHouse/pull/4538) is a much better approach than geohash because the filters are explicit and data locality is higher. It works with any kind coordinate system (it does not matter if that's lon, lat or x,y).
- With string values you ... | 2019-04-24T09:03:14 | 2019-04-24T09:03:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 486,663,585 | 5,101 | bzon | @den-crane thanks for the answer. I ran the command and I got the following.
```
# hostname -f
clickhouse-01-1.clickhouses.default.svc.cluster.local
``` | 2019-04-25T13:04:16 | 2019-04-25T13:04:16 | {} | NONE |
ClickHouse | ClickHouse | 485,605,949 | 5,087 | AlexZFX | @neverlee Thank you for your help during my translation。 | 2019-04-23T01:15:49 | 2019-04-23T01:15:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 485,724,801 | 4,586 | filimonov | ```
Received exception from server (version 19.4.2):
Code: 48. DB::Exception: Received from localhost:9000, 127.0.0.1. DB::Exception: DatabaseDictionary: removeTable() is not supported.
``` | 2019-04-23T09:26:14 | 2019-04-23T09:26:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 486,339,631 | 5,030 | alexey-milovidov | Performance tests are Ok. | 2019-04-24T17:17:40 | 2019-04-24T17:17:40 | {} | MEMBER |
ClickHouse | ClickHouse | 486,527,199 | 5,099 | Crazylearner30 | Thanks for the information Naveen,
Little confused on this.
How many machines to be considered for better performance ?
Thanks,
Divya
| 2019-04-25T05:41:47 | 2019-04-25T05:41:47 | {} | NONE |
ClickHouse | ClickHouse | 485,745,349 | 5,019 | bzon | @4ertus2 I managed to get it working fast by including the partitioning key in the query.
```sql
SELECT COUNT(*) FROM dist_table
WHERE employee_id ='xxxxxxxxx'
AND date BETWEEN '2019-04-16' AND '2019-04-16'
AND org_id IN('11111')
AND request_time BETWEEN '2019-04-16 00:00:00' AND '2019-04-16 23:59:59'
``` | 2019-04-23T10:22:02 | 2019-04-23T10:22:02 | {} | NONE |
ClickHouse | ClickHouse | 486,005,385 | 5,045 | alexey-milovidov | Now the tests queries became 30% slower:
https://clickhouse-test-reports.s3.yandex.net/5045/8b0dda39e4e268af282e77c2d24f89f04b3f64c6/performance_test/comparison_min_time_gcc_8.html
I have to either do some magic or just feel fine about it. | 2019-04-23T23:03:21 | 2019-04-23T23:03:21 | {} | MEMBER |
ClickHouse | ClickHouse | 485,519,107 | 5,045 | alexey-milovidov | Performance tests are Ok. | 2019-04-22T19:15:03 | 2019-04-22T19:15:03 | {} | MEMBER |
ClickHouse | ClickHouse | 485,917,591 | 5,071 | alexey-milovidov | Now it looks Ok.
BTW, Should we support case insensitive values: `TRUE`, `True`, etc?
It depends on the origin of the motivation for this data type.
Also please make some research about Bool type in other DBMS.
Maybe we should peek synonims for this data type, like BIT, Boolean?
But if we do, it can be incompa... | 2019-04-23T18:19:09 | 2019-04-23T18:20:08 | {} | MEMBER |
ClickHouse | ClickHouse | 486,049,023 | 5,071 | hczhcz | @alexey-milovidov
The motivation (and the implementation) is similar to IPv4/IPv6 domains in #3669 .
* Bool is useful when we match the type systems between ClickHouse and other systems like JSON, programming language bindings, etc.
* It is also good for documentation (in comparison to a raw UInt8)
* It is pai... | 2019-04-24T03:02:09 | 2019-04-24T03:02:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 486,125,995 | 5,088 | YiuRULE | Ok thanks, seem to be normally adapted to our case :)
So if I understand well, it will work in a similar way compared to when we insert a CSV with clickhouse with a first argument who is the format of the file and the second argument the defiintion of the datas right ? | 2019-04-24T08:37:35 | 2019-04-24T08:37:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 486,190,014 | 3,882 | bzon | @shurshun what is your workaround for this problem in Kubernetes? I'm also having the same issue and `system drop dns cache` does not solve the problem. I always have to restart the pod. | 2019-04-24T11:49:46 | 2019-04-24T11:49:46 | {} | NONE |
ClickHouse | ClickHouse | 486,190,518 | 5,088 | alexey-milovidov | Right. It will allow to to arbitrary data transformations on INSERT.
| 2019-04-24T11:51:29 | 2019-04-24T11:51:29 | {
"hooray": 2
} | MEMBER |
ClickHouse | ClickHouse | 486,339,989 | 5,030 | alexey-milovidov | > We don't have anymore build problem using Clang, does this fix seem ok ?
Yes.
| 2019-04-24T17:18:44 | 2019-04-24T17:18:44 | {} | MEMBER |
ClickHouse | ClickHouse | 486,411,288 | 5,066 | alexey-milovidov | https://github.com/yandex/ClickHouse/pull/4535 | 2019-04-24T20:24:10 | 2019-04-24T20:24:51 | {} | MEMBER |
ClickHouse | ClickHouse | 486,454,768 | 5,105 | alexey-milovidov | Consider the following expression:
`arrayMap(x -> f(x, b), a)`
To calculate this expression, temporary array join by `a` is performed:
```
a[1], b
a[2], b
a[3], b
...
```
and expression is evaluated:
```
a[1], b, f(a[1], b)
a[2], b, f(a[2], b)
a[3], b, f(a[3], b)
...
```
then the result is col... | 2019-04-24T22:42:21 | 2019-04-24T22:42:21 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 486,470,614 | 5,105 | den-crane | Thank you.
It seems I always used arrayEnumerate to iterate & filter through nested (K,V) by mistake.
arrayFilter is able to do pure magic.
```
select arrayFilter((v, k) -> k = 'a', values, keys) from
(select ['a','a','b','a'] keys, [1,2,3,4] values)
--
[1,2,4]
```
select arrayFilter((v, k)-> k = '777',... | 2019-04-25T00:02:36 | 2019-04-25T00:05:16 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 486,711,938 | 5,071 | hczhcz | @orloffv `Nullable(Bool)` works. | 2019-04-25T15:03:16 | 2019-04-25T15:03:16 | {
"rocket": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 486,856,047 | 5,045 | geldot | Does this fix cutToFirstSignificantSubdomain() etc? Covering those in the tests too will be great (and if not I guess we should fix those too, maybe as a separate issue)
We have `cutToFirstSignificantSubdomain(concatAssumeInjective('https://',domain))` in our queries right now, looking forward to removing. | 2019-04-25T22:10:39 | 2019-04-25T22:11:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 485,538,174 | 5,045 | alexey-milovidov | It should give correct results for both:
```
https://yandex.ru/ -> yandex.ru
yandex.ru:80 -> yandex.ru
mailto:hello@world ->
magnet:test.test ->
``` | 2019-04-22T20:18:30 | 2019-04-22T20:18:30 | {} | MEMBER |
ClickHouse | ClickHouse | 485,668,765 | 4,211 | code-of-kpp | the only workaround for now is to set those settings for default user in users.xml and restart the server | 2019-04-23T07:00:03 | 2019-04-23T07:00:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 485,910,457 | 4,945 | alexey-milovidov | > "limit 0, 10" working incorrect
"limit 0 offset 10" working fine
Both variants work correct, exactly as expected.
`limit 0, 10` is equivalent to `LIMIT 10 OFFSET 0` or just `LIMIT 10`
`limit 0 offset 10` is equivalent to `LIMIT 10, 0` or just `LIMIT 0` | 2019-04-23T17:59:42 | 2019-04-23T17:59:42 | {} | MEMBER |
ClickHouse | ClickHouse | 485,918,403 | 5,071 | alexey-milovidov | Should this type also represent a constraint?
(what will happen if we write `INSERT INTO t SELECT 5 AS b`) | 2019-04-23T18:21:23 | 2019-04-23T18:21:30 | {} | MEMBER |
ClickHouse | ClickHouse | 486,004,808 | 5,030 | alexey-milovidov | @alesapin I've made a commit. | 2019-04-23T23:00:22 | 2019-04-23T23:00:22 | {} | MEMBER |
ClickHouse | ClickHouse | 486,067,680 | 5,094 | blinkov | cc @zhang2014 @sundy-li | 2019-04-24T04:56:39 | 2019-04-24T04:56:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 485,967,006 | 5,089 | filimonov | Broken since 19.1.
May be seconds vs microseconds issue. Jdbc bridge need to be checked. | 2019-04-23T20:43:50 | 2019-04-24T18:32:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 485,500,734 | 5,002 | ipoese | Good thinking !
We've been down this road aswell. The server has 768Gbyte of ECC memory, we've run memory tests and CPU tests aswell - none of which turned anything up. | 2019-04-22T18:17:29 | 2019-04-22T18:17:29 | {} | NONE |
ClickHouse | ClickHouse | 485,538,238 | 5,074 | proller | Also we don't support building contrib/mariadb-connector-c under freebsd.
| 2019-04-22T20:18:41 | 2019-04-22T20:18:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 485,860,522 | 4,945 | georgeakulov | "limit 0, 10" working incorrect
"limit 0 offset 10" working fine | 2019-04-23T15:45:28 | 2019-04-23T15:46:08 | {} | NONE |
ClickHouse | ClickHouse | 486,454,840 | 5,105 | alexey-milovidov | Possible workarounds:
1. Lower `max_block_size`. | 2019-04-24T22:42:41 | 2019-04-24T22:42:41 | {} | MEMBER |
ClickHouse | ClickHouse | 486,456,785 | 5,105 | alexey-milovidov | create table X(A Array(String)) engine = **TinyLog**;
insert into X select arrayMap(x->toString (x) , range(1000)) from numbers(1000);
SET send_logs_level = 'trace', **max_block_size = 1**;
select arrayFilter(x->A[x]='777', arrayEnumerate(A)) from X format Null;
Peak memory usage (for query): 18.08 MiB. | 2019-04-24T22:51:27 | 2019-04-24T22:51:27 | {} | MEMBER |
ClickHouse | ClickHouse | 486,711,155 | 5,071 | hczhcz | @alexey-milovidov
For now, both of them generate type errors.
We may use `select x;` for bool values and `select x = toIPv4('127.0.0.1')` for ipv4 values.
Two things might be useful:
1. Add a `toBool(x)` function to perform type conversions
2. Return bool type in comparison/logic functions | 2019-04-25T15:01:17 | 2019-04-25T15:01:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 486,761,158 | 4,616 | kmatt | v19.5.3 has been more stable for me (at 48 hours into testing). Raised some similar concerns in #5085 | 2019-04-25T17:17:17 | 2019-04-25T17:17:17 | {} | NONE |
ClickHouse | ClickHouse | 486,765,388 | 4,970 | den-crane | related to https://github.com/yandex/ClickHouse/issues/5101 ( interserver_http_host helped ) | 2019-04-25T17:28:15 | 2019-04-25T17:28:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 486,772,205 | 5,066 | abyss7 | Found the culprit code - trying to fix. | 2019-04-25T17:47:46 | 2019-04-25T17:47:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 485,535,948 | 5,045 | alexey-milovidov | Now it gives somewhat wrong result for `firstSignificantSubdomain('magnet:yandex.abc')`.
| 2019-04-22T20:11:23 | 2019-04-22T20:15:51 | {} | MEMBER |
ClickHouse | ClickHouse | 485,596,860 | 5,067 | kmatt | Created #5085 related to server process halting under much lower message load. | 2019-04-23T00:20:29 | 2019-04-23T00:20:55 | {} | NONE |
ClickHouse | ClickHouse | 485,952,980 | 5,030 | alesapin | @YiuRULE unfortunately in our CI system Docker images have to be pushed by hands. I pushed images containing gperf. Can you merge with master or do any insignificant commit? | 2019-04-23T20:02:33 | 2019-04-23T20:02:33 | {} | MEMBER |
ClickHouse | ClickHouse | 486,298,176 | 5,100 | abyss7 | We already realized that there is a trade-off in this situation: if we read a very big message and partially handle it, then on server crash we may end with a committed message and partially lost data, or with uncommitted message and process some data twice. We'll introduce the setting to choose the desired behavior. | 2019-04-24T15:44:47 | 2019-04-24T15:44:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 486,003,436 | 5,019 | bzon | @alexey-milovidov thanks for the answers! 🎉
Sorry but I want to be enlightened about the following:
* Why do we have to ORDER BY **org_id** instead of **employee_id**? In our case, **org_id** has a lower cardinality than **employeed_id**. Is that why you recommended it?
* What's the benefit of using **LowCardi... | 2019-04-23T22:53:58 | 2019-04-23T22:53:58 | {} | NONE |
ClickHouse | ClickHouse | 486,003,930 | 5,070 | alexey-milovidov | Code cleanups required to make all the variants to follow some general logic. | 2019-04-23T22:56:11 | 2019-04-23T22:56:11 | {} | MEMBER |
ClickHouse | ClickHouse | 486,080,792 | 4,658 | ekbfh | Have this problem too on `19.4.3 revision 54416` | 2019-04-24T06:04:26 | 2019-04-24T06:04:26 | {} | NONE |
ClickHouse | ClickHouse | 486,921,414 | 5,045 | YiuRULE | @geldot Yep, the patch fix also your case, I forget to add test and just added some test for the function `cutToFirstSignificantSubdomain`, thanks ! | 2019-04-26T04:27:33 | 2019-04-26T04:27:33 | {
"rocket": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 485,595,388 | 5,085 | kmatt | It appears that SSL failures can bring a running server down as well. Just observed a node that was previously ingesting messages stop without after emitting only these two messages into the logs:
```
2019.04.23 00:08:49.359147 [ 1 ] {} <Error> Application: Caught exception while loading metadata: std::exception. C... | 2019-04-23T00:11:51 | 2019-04-23T00:17:29 | {} | NONE |
ClickHouse | ClickHouse | 485,738,816 | 5,003 | alex-zaitsev | @alexey-milovidov, numeric version will make things much more complicated. In particular, geohash consists of two parts: value and precision, for string precision is just a string length. You can fit it into UInt64 (12x5 = 60 bits for value and 4 for precision), but all other operations would become more difficult. Uin... | 2019-04-23T10:02:25 | 2019-04-23T10:07:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 485,757,932 | 4,686 | vitlibar | @hczhcz Can you fix build? I'd like to merge your PR but all checks should pass. | 2019-04-23T11:07:09 | 2019-04-23T11:07:09 | {} | MEMBER |
ClickHouse | ClickHouse | 485,788,639 | 4,686 | vitlibar | simdjson shows good performance:
```
SELECT any(jsonExtractInt(json, 'a')) from (select '{"a":2,"b":35.4,"c":[1,2,{"e":"f\\n"}], "d":"qq"}' as json from system.numbers limit 200000000)
1 rows in set. Elapsed: 1.377 sec. Processed 200.02 million rows, 1.60 GB (145.20 million rows/s., 1.16 GB/s.)
```
```
SELECT... | 2019-04-23T12:49:32 | 2019-04-23T12:51:45 | {} | MEMBER |
ClickHouse | ClickHouse | 486,154,734 | 4,047 | bzon | I'm using version 19.4.3. This issue still happens even after running `SYSTEM DROP DNS CACHE` and `<keep_alive_timeout>` in `config.xml` with 8 seconds.
```
<Error> default.table_sharded (StorageReplicatedMergeTree): DB::StorageReplicatedMergeTree::queueTask()::<lambda(DB::StorageReplicatedMergeTree::LogEntryPtr&)>... | 2019-04-24T09:54:49 | 2019-04-24T09:54:49 | {} | NONE |
ClickHouse | ClickHouse | 486,573,133 | 5,100 | minaevmike | Also i wan't to clarify, what should happen with block of data when server restarts(`systemctl restart`). This block must be written or it would be lost? | 2019-04-25T08:35:02 | 2019-04-25T08:43:52 | {} | NONE |
ClickHouse | ClickHouse | 486,702,133 | 5,101 | bzon | I'm now changing my Clickhouse Docker entrypoint script to make the interserver_http_host dynamically configurable upon the Pod runtime. I will update this thread for the result. | 2019-04-25T14:43:19 | 2019-04-25T14:43:19 | {} | NONE |
ClickHouse | ClickHouse | 485,731,032 | 1,148 | filimonov | > FixedStrings as dictionary keys are still not supported.
>
> But you can load FixedString content as a String.
> Simply specify `String` as a type in dictionary structure and it will load FixedStrings from the target table.
That doesn't work. I.e. it properly creates the dictionary and loads the data, but you ... | 2019-04-23T09:43:02 | 2019-04-23T09:43:02 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 486,554,884 | 5,099 | krnaveen14 | Depends on the use case and usage pattern (Latency sensitive / Throughout sensitive / Parallel Queries / Response Time / IO Limits / No.of databases / Tables / Partition keys) etc...
I would suggest configure the EC2 instance in staging environment, simulate the production workload in staging environment and analyse... | 2019-04-25T07:37:58 | 2019-04-25T07:37:58 | {} | NONE |
ClickHouse | ClickHouse | 485,601,211 | 5,086 | AlexZFX | @neverlee Thank you for your help during my translation。 | 2019-04-23T00:47:56 | 2019-04-23T00:47:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 483,830,038 | 5,019 | 4ertus2 | Your request is on request_time that is not in order key. So it's a full scan with filter on second column in order by key. It cannot be fast. | 2019-04-16T20:26:14 | 2019-04-23T18:05:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 486,370,273 | 5,071 | alexey-milovidov | We have to check if `x = 'true'` works in SQL query.
BTW, does `x = '127.0.0.1'` work for IPv4 domain?
If not, it is the cause of major usability issues. | 2019-04-24T18:24:33 | 2019-04-24T20:01:23 | {} | MEMBER |
ClickHouse | ClickHouse | 486,434,821 | 4,993 | Gladdy | @4ertus2 Found some time and ran some basic tests: the join performance does seem to be dependent on the data. It might be an option to come up with some representative use cases (which the one below is not) and then go for a single implementation, or add the join algorithm as a setting? (in my testing I just passed it... | 2019-04-24T21:31:08 | 2019-04-24T21:31:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 486,662,739 | 5,101 | den-crane | interserver_http_host -> host_name: ch-01-1.default.svc.cluster.local
```
<!-- Hostname that is used by other replicas to request this server.
If not specified, than it is determined analoguous to 'hostname -f' command.
This setting could be used to switch replication to another network i... | 2019-04-25T13:02:01 | 2019-04-25T13:03:38 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 486,671,319 | 3,143 | citrin | I also see a negative value with ClickHouse 19.5.3.8
```
┌─metric───────────────────────────────────┬─formatReadableSize(value)─┐
│ MemoryTracking │ 18.63 KiB │
│ MemoryTrackingInBackgroundProcessingPool │ -3.58 MiB │
│ MemoryTrackingInBackgroundSchedulePoo... | 2019-04-25T13:25:02 | 2019-04-25T13:25:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 486,725,121 | 5,101 | bzon | @den-crane it worked! 🎉
For each Kubernetes Pod replica, the docker_related_config.xml now looks like:
```xml
<yandex>
<!-- Listen wildcard address to allow accepting connections from other containers and host network. -->
<listen_host>::</listen_host>
<listen_host>0.0.0.0</listen_host>
<listen_try>1</... | 2019-04-25T15:36:31 | 2019-04-25T15:36:31 | {
"+1": 2
} | NONE |
ClickHouse | ClickHouse | 486,852,924 | 5,095 | geldot | Would it make sense to return an (IPv4,IPv4) tuple rather than a (UInt32,UInt32) tuple? That would make the interface more intuitive for display/interactive use. Specific use cases can always cast to UInt32 afterwards if needed. | 2019-04-25T21:58:52 | 2019-04-25T21:58:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 485,517,405 | 5,003 | alexey-milovidov | If you want to store geohashes in ClickHouse, the numeric version is preferred.
In the same way as UInt32 is preferred to store IPv4 addresses. | 2019-04-22T19:09:24 | 2019-04-22T19:09:24 | {} | MEMBER |
ClickHouse | ClickHouse | 485,609,197 | 5,087 | neverlee | Please, don't modify the link `[Original article](https://clickhouse.yandex/docs/en/operations/table_engines/dictionary/)`. Is is used for the source article without translation.` | 2019-04-23T01:35:26 | 2019-04-23T01:35:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 485,828,175 | 4,456 | AntonVozdvizhenskiy | On 19.4.2 revision 54416, with same simptoms, helped command `system reload dictionaries`. | 2019-04-23T14:31:05 | 2019-04-23T14:31:05 | {
"+1": 4,
"heart": 1
} | NONE |
ClickHouse | ClickHouse | 485,855,646 | 5,085 | kmatt | Entire clickhouse-server.err.log from other failed node:
```
2019.04.23 14:59:32.878876 [ 2 ] {} <Error> BaseDaemon: ########################################
2019.04.23 14:59:32.879034 [ 2 ] {} <Error> BaseDaemon: (from thread 24) Received signal Segmentation fault (11).
2019.04.23 14:59:32.879061 [ 2 ] {} <Error... | 2019-04-23T15:33:51 | 2019-04-23T15:33:51 | {} | NONE |
ClickHouse | ClickHouse | 485,914,166 | 5,088 | alexey-milovidov | We have a different feature in plan that will cover your needs.
It looks like this:
```
INSERT INTO test SELECT * FROM input(CSV, 'a UInt8') LIMIT 10 OFFSET 5
``` | 2019-04-23T18:09:48 | 2019-04-23T18:09:48 | {} | MEMBER |
ClickHouse | ClickHouse | 485,518,732 | 5,045 | alexey-milovidov | Stateful tests should be updated. You can download test data according to the instruction:
```
sudo apt install wget xz-utils
wget https://clickhouse-datasets.s3.yandex.net/hits/tsv/hits_v1.tsv.xz
wget https://clickhouse-datasets.s3.yandex.net/visits/tsv/visits_v1.tsv.xz
xz -v -d hits_v1.tsv.xz
xz -v -d vis... | 2019-04-22T19:13:55 | 2019-04-23T23:00:46 | {} | MEMBER |
ClickHouse | ClickHouse | 486,455,569 | 5,105 | alexey-milovidov | Possible solutions:
1. Automatically use `LowCardinality` data type for array joining. But LowCardinality arrays are not supported. | 2019-04-24T22:45:52 | 2019-04-24T22:47:36 | {} | MEMBER |
ClickHouse | ClickHouse | 486,504,618 | 5,097 | zhang2014 | > I cannot instantly figure out from the diff, what exact changes were made.
This fix is a bit of a hack, avoiding deadlock by changing the order in which insert and select executed, and I've added some comments to the code.
https://github.com/yandex/ClickHouse/pull/5097/files#diff-0a9ad6b0fe82092d9bf054174f8bac5... | 2019-04-25T03:15:06 | 2019-04-25T03:17:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 486,870,847 | 5,071 | geldot | In both MySQL and PostgreSQL, TRUE/FALSE are keywords, which makes sense independently of this PR. MySQL:
```
SELECT true, false, TRUE, FALSE, True, False;
-- 1 0 1 0 1 0
```
In MySQL it's a plain 1/0 bit, in PostgreSQL it's a special type similar to this patch with pretty formatting.
PostgreSQL accepts TRU... | 2019-04-25T23:16:29 | 2019-04-25T23:16:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 485,734,016 | 5,087 | blinkov | cc @zhang2014 @sundy-li | 2019-04-23T09:50:01 | 2019-04-23T09:50:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 485,589,748 | 4,680 | vitlibar | I've compared performance of this approach with [PR #4686](https://github.com/yandex/ClickHouse/pull/4686):
[PR #4680 Implement jsonAny, jsonAll, jsonAnyArray, jsonAllArrays, jsonCount](https://github.com/yandex/ClickHouse/pull/4680):
```
SELECT any(jsonAny(json, 'a')) from (select '{"a":2,"b":35.4,"c":[1,2,{"e":"... | 2019-04-22T23:40:55 | 2019-04-23T12:45:43 | {} | MEMBER |
ClickHouse | ClickHouse | 485,913,152 | 5,019 | alexey-milovidov | 1. Your order key should be `(org_id, request_time)` instead of `(date, employee_id)`.
2. Change
```
employee_id String,
org_id String,
```
to
```
employee_id LowCardinality(String),
org_id LowCardinality(String),
```
| 2019-04-23T18:07:20 | 2019-04-23T18:07:20 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 486,008,122 | 5,070 | danlark1 | > Code cleanups required to make all the variants to follow some general logic.
Yeah, I know but was unsuccessful the last day :( | 2019-04-23T23:16:09 | 2019-04-23T23:16:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 486,091,325 | 5,094 | sundy-li | LGTM | 2019-04-24T06:48:50 | 2019-04-24T06:48:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 486,452,567 | 5,104 | alexey-milovidov | It's a server-side error (not something depend on client), so 5xx code should be sent.
`503 Service Unavailable` looks reasonable. | 2019-04-24T22:32:38 | 2019-04-24T22:32:38 | {} | MEMBER |
ClickHouse | ClickHouse | 485,761,404 | 5,024 | bzon | @den-crane can you still modify these table settings after table creation? | 2019-04-23T11:19:41 | 2019-04-23T11:19:41 | {} | NONE |
ClickHouse | ClickHouse | 485,878,946 | 5,024 | den-crane | @bzon of course not (it has no sense it requires to remerge everything).
I am using https://clickhouse.yandex/docs/en/operations/table_engines/merge/ (Example 2:) | 2019-04-23T16:32:39 | 2019-04-23T16:32:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 486,288,465 | 5,088 | alexey-milovidov | @palasonic1 will do this task. | 2019-04-24T15:21:58 | 2019-04-24T15:21:58 | {} | MEMBER |
ClickHouse | ClickHouse | 485,520,191 | 2,829 | alexey-milovidov | This issue is not on the roadmap. But it's pretty easy in implementation and I almost forgot about it. | 2019-04-22T19:18:35 | 2019-04-22T19:18:35 | {
"eyes": 3
} | MEMBER |
ClickHouse | ClickHouse | 485,537,506 | 5,045 | alexey-milovidov | It is definitely not the expected behaviour:
```
SELECT
domain(URL) AS domain,
URL
FROM test.hits
WHERE (URL NOT LIKE '%://%') AND (domain != '')
LIMIT 10
┌─domain─┬─URL─────────┐
│ about │ about:blank │
│ about │ about:blank │
└────────┴─────────────┘
┌─domain─┬─URL───────────┐
│ tel │... | 2019-04-22T20:16:22 | 2019-04-22T20:16:22 | {} | MEMBER |
ClickHouse | ClickHouse | 485,843,046 | 5,085 | kmatt | Switched to testing with Docker yandex/clickhouse-server:19.1.14, pointing to the same local data volumes and config.xml:
```
docker run -d --name clickhouse --net host --ulimit nofile=262144:262144 \
-v /etc/clickhouse-server:/etc/clickhouse-server \
-v /localdata/clickhouse:/localdata/clickhouse \
... | 2019-04-23T15:05:06 | 2019-04-23T15:11:07 | {} | NONE |
ClickHouse | ClickHouse | 486,005,832 | 5,045 | alexey-milovidov | @YiuRULE
```
https_edit.do?res=5&uinfo=ww-980-wh-660-fps-v-krasnoyaroslavy-ispytyvat-i-table_key=5&has_photelnoe => https_edit.do
http-www-1349-wh-957-fw-1038-03-28.0395806813113 => http-www-1349-wh-957-fw-1038-03-28.0395806813113
http-www-1349-wh-957-fw-1038-03-28.0395806813113 => http-www-1349-wh-957-fw-1038-03-... | 2019-04-23T23:05:10 | 2019-04-23T23:05:10 | {} | MEMBER |
ClickHouse | ClickHouse | 486,312,499 | 5,071 | alexey-milovidov | Ok. | 2019-04-24T16:17:59 | 2019-04-24T16:17:59 | {} | MEMBER |
ClickHouse | ClickHouse | 485,499,005 | 2,829 | champtar | @alexey-milovidov is this issue somewhere on the roadmap ? | 2019-04-22T18:12:05 | 2019-04-22T18:12:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 485,657,614 | 5,045 | YiuRULE | Ok thanks, I applied a patch for the required fix and up to date the tests.
I just have few updates, 5 tests with the `SELECT ... test.hits` have a match.
```https_edit.do?res=5&uinfo=ww-980-wh-660-fps-v-krasnoyaroslavy-ispytyvat-i-table_key=5&has_photelnoe => https_edit.do
https_edit.do?res=5&uinfo=ww-980-wh-66... | 2019-04-23T06:14:13 | 2019-04-23T07:33:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 485,762,160 | 4,686 | hczhcz | @vitlibar Sure. I will try to fix it. | 2019-04-23T11:22:23 | 2019-04-23T11:22:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 485,881,304 | 5,085 | kmatt |
```
cat /var/log/clickhouse-server.err.log
2019.04.23 16:20:24.528645 [ 1 ] {} <Error> Application: Caught exception while loading metadata: std::exception. Code: 1001, type: cppkafka::Exception, e.what() = Failed to create consumer handle: ssl.ca.location failed: ../contrib/ssl/crypto/x509/by_file.c:254: error:0... | 2019-04-23T16:39:13 | 2019-04-23T16:39:13 | {} | NONE |
ClickHouse | ClickHouse | 486,190,687 | 5,030 | YiuRULE | Sorry just waited that the CI build finish, I was able to fix the warning problems with modifying a bit the `gperf` file, the generated file have now on the beginning of the file this pragma instructions
```
#pragma GCC diagnostic push
#pragma GCC diagnostic ignored "-Wimplicit-fallthrough"
#pragma GCC diagnostic... | 2019-04-24T11:52:05 | 2019-04-24T11:52:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 486,228,815 | 5,099 | krnaveen14 | r5.xlarge or r5d.xlarge - 4 vCPU, 32GB RAM, 150GB Instance Storage (r5d)
8 x 300GB GP2 EBS Volumes Striped (RAID0)
150GB Instance Storage as Temporary location for External GROUP BY
CPU and RAM requirements depends on the usage pattern. Above is just an approximation based on t2.2xlarge. If vCPU count of r5 is n... | 2019-04-24T13:27:27 | 2019-04-24T13:27:27 | {} | NONE |
ClickHouse | ClickHouse | 486,371,980 | 5,097 | alexey-milovidov | I cannot instantly figure out from the diff, what exact changes were made. | 2019-04-24T18:29:09 | 2019-04-24T18:29:09 | {} | MEMBER |
ClickHouse | ClickHouse | 486,400,455 | 5,071 | orloffv | is nullable supported? | 2019-04-24T19:52:13 | 2019-04-24T19:52:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 486,562,899 | 974 | dongchaohui | how to fix this issue ? | 2019-04-25T08:04:02 | 2019-04-25T08:04:02 | {} | NONE |
ClickHouse | ClickHouse | 486,639,667 | 5,107 | blinkov | cc @zhang2014 @sundy-li | 2019-04-25T11:48:55 | 2019-04-25T11:48:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 486,659,366 | 5,107 | AlexZFX | i'm sorry for the check error and I'm trying to fix it.
I have just build successfully with parameter “--skip-pdf --skip-single-page”. | 2019-04-25T12:52:13 | 2019-04-25T12:52:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 486,666,118 | 5,101 | den-crane | So you probably better to specify **interserver_http_host** in config.xml though I am not sure it will work without CH restart than guessing what `hostname -f` is applied in CH unix process.
| 2019-04-25T13:11:06 | 2019-04-25T13:11:06 | {} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.