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 | 527,474,051 | 6,707 | jcdang | 1. Is send_progress_in_http_headers even something that's wanted if server-sent-events gets implemented?
2. I'm thinking that if a query goes over the buffer size, it should fail the multiquery call. Is this unreasonable?
4. I saw this being sent but it was repeating with the same value. | 2019-09-03T14:06:20 | 2019-09-03T14:06:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,716,652 | 6,809 | den-crane | It could be bounded by disk IO (check `sar -d 1 100`) or network (check `iftop`) | 2019-09-04T02:48:19 | 2019-09-04T02:48:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,919,214 | 6,761 | yuzhichang | Please review pr #6819. | 2019-09-04T14:12:31 | 2019-09-04T14:12:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,010,666 | 6,822 | Jack012a | Hello den-crane,
I use ABC to specify a different format than the standard one. It seems that clickhouse doesn't allow us to specify the input format.
In your example, where I should add parseDateTimeBestEffort to? Can we add this conversion into the table definition?
Thank you | 2019-09-04T17:48:57 | 2019-09-04T17:53:24 | {} | NONE |
ClickHouse | ClickHouse | 528,045,114 | 6,817 | amosbird | I use shared build a lot and haven't encountered any issues. Perhaps using `env LD_PRELOAD=$BUILD_DIR/libs/libcommon/libcommon.so
clickhouse server` can resolve your issue. BTW, you might want to try https://github.com/yandex/ClickHouse/pull/6602 which significantly reduces linking time. | 2019-09-04T19:17:42 | 2019-09-04T19:17:42 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 528,075,835 | 6,826 | Akazz | Threads stalling in a typical mutual DL
```
#2 __pthread_cond_wait (cond=0x7f44cec00030, mutex=0x7f44d179ee90) at pthread_cond_wait.c:655
#3 0x00005605da61bf5c in std::condition_variable::wait(std::unique_lock<std::mutex>&) ()
#4 0x00005605d98a524e in std::condition_variable::wait<DB::RWLockImpl::getLock(DB::RWL... | 2019-09-04T20:35:51 | 2019-09-04T20:35:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,092,762 | 6,822 | den-crane | https://clickhouse.yandex/docs/en/operations/settings/settings/#settings-date_time_input_format
clickhouse-client --**date_time_input_format=best_effort**
Or use ENGINE = Null
```
create table Hello_null (
timestamp UInt64,
current_date_str String
) ENGINE = Null;
create table Hello (
timestamp UInt6... | 2019-09-04T21:22:40 | 2019-09-04T21:43:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,275,582 | 6,795 | xuxudede | Has been resolved, thank you！ | 2019-09-05T09:12:52 | 2019-09-05T09:12:52 | {} | NONE |
ClickHouse | ClickHouse | 528,323,507 | 6,830 | blinkov | There were some examples in this talk (slides are in description), but most of it is in Russian: https://www.youtube.com/watch?v=7pyNmackUTw | 2019-09-05T11:39:22 | 2019-09-05T11:39:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,426,770 | 6,784 | BayoNet | @zhang2014 Hello. I'm trying to create a table with functions that optimized for using with indexes. Can you help me with bloom_filter index? Which of the functions can be used effectively with bloom_filter? | 2019-09-05T15:32:43 | 2019-09-05T15:32:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,491,695 | 6,472 | xmariachi | > @xmariachi Yes, you should upgrade the remaining replicas.
Yep, I no longer see the message. Thanks. | 2019-09-05T17:39:06 | 2019-09-05T17:39:06 | {} | NONE |
ClickHouse | ClickHouse | 527,538,921 | 6,803 | simPod | Ah, Datagrip displayed it as null
thanks | 2019-09-03T16:37:47 | 2019-09-03T16:37:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,539,674 | 6,803 | den-crane | > Ah, Datagrip displayed it as null
OK. I would open a ticket in Datagrip support. | 2019-09-03T16:39:46 | 2019-09-03T16:40:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,540,151 | 6,803 | simPod | Not a bug then. Thanks for quick response. | 2019-09-03T16:41:01 | 2019-09-03T16:41:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,599,215 | 6,007 | alexey-milovidov | @KochetovNicolai Let's pull master and we'll continue reviewing. | 2019-09-03T19:14:29 | 2019-09-03T19:14:29 | {} | MEMBER |
ClickHouse | ClickHouse | 527,711,253 | 6,809 | filimonov | It is expected. ClickHouse uses server resources very efficiently. When you run single query - it uses 100% of your server resources. If you run 2 queries in parallel - each query can get about 50%, and will be ~ twice slower. | 2019-09-04T02:19:55 | 2019-09-04T02:19:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,968,626 | 1,784 | xmariachi | Thanks.
I've found this issue in the logs - may it be the cause? I replied in that permalink where you can see the problem:
https://github.com/yandex/ClickHouse/issues/6472#issuecomment-527967770 | 2019-09-04T16:01:49 | 2019-09-04T16:01:49 | {} | NONE |
ClickHouse | ClickHouse | 528,378,009 | 6,814 | vitlibar | @zhang2014 Can you explain please a bit what you're implementing here? Currently it doesn't seem very clear to me. What is the `SmallestJSON` data type? Why can't you just store an ordinary JSON in a `ColumnString`? | 2019-09-05T13:58:57 | 2019-09-05T14:01:35 | {} | MEMBER |
ClickHouse | ClickHouse | 528,538,344 | 6,834 | den-crane | https://clickhouse.yandex/docs/en/operations/table_engines/replication/
Data blocks are deduplicated. For multiple writes of the same data block (data blocks of the same size containing the same rows in the same order), the block is only written once. The reason for this is in case of network failures when the clien... | 2019-09-05T19:15:29 | 2019-09-05T19:17:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,443,662 | 6,667 | akuzm | I was able to reproduce the improvement for DISTINCT that we see in CI performance test (median QPS for 7 runs of 137 seconds, the runs with different server versions are interleaved):
```
SELECT DISTINCT URL, ngramDistanceUTF8(URL, 'как дела') AS distance FROM hits_100m_single ORDER BY distance ASC LIMIT 50
┏━━━━... | 2019-09-03T12:49:45 | 2019-09-03T12:49:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,683,894 | 6,739 | alexey-milovidov | I have some initial thoughts, and this algorithm looks easy to implement. It may look like adaptive variant of QuantileTiming. | 2019-09-03T23:54:54 | 2019-09-03T23:54:54 | {
"heart": 2
} | MEMBER |
ClickHouse | ClickHouse | 527,967,770 | 6,472 | xmariachi | I'm finding this issue too. I have two replicas. One of them is showing it at a rate of 13 occurrences per minute, and the other one at a rate of 15 per hour.
Servers have different versions (by mistake):
- Connected to ClickHouse server version 19.13.3 revision 54425
- Connected to ClickHouse server version... | 2019-09-04T15:59:53 | 2019-09-04T15:59:53 | {} | NONE |
ClickHouse | ClickHouse | 527,835,523 | 6,812 | alesapin | Also, which version do you use? | 2019-09-04T10:07:30 | 2019-09-04T10:07:30 | {} | MEMBER |
ClickHouse | ClickHouse | 528,628,842 | 6,661 | filimonov | > @filimonov up
>
> on 19.13.3.26 it like 50% of dups
@bespilotnayakartoshka Can you try on 19.11.9 (just released, it has important Kafka fixes)?
Also what's about @den-crane question? | 2019-09-05T23:01:41 | 2019-09-05T23:02:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,403,729 | 6,629 | CurtizJ | Probably it's a CI problem.
```
2019-09-03 03:56:08 00838_system_tables_drop_table_race: [ FAIL ] - having stderror:
2019-09-03 03:56:08 [af65ce602906] 2019.09.03 03:56:02.871285 {783dcb4e-b51b-487d-ae46-1d681d1bcb10} [ 101 ] <Error> executeQuery: Poco::Exception. Code: 1000, e.cod... | 2019-09-03T10:38:03 | 2019-09-03T10:38:03 | {} | MEMBER |
ClickHouse | ClickHouse | 527,540,080 | 6,688 | BayoNet | Сделаю. | 2019-09-03T16:40:49 | 2019-09-03T16:40:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,580,085 | 6,773 | alexey-milovidov | > The log indicates that the standalone node receives the query which aim to 127.0.0.2, without listening on 127.0.0.2.
We run ClickHouse that listen to `::` (everything) to run tests with `shard` tag. | 2019-09-03T18:24:49 | 2019-09-03T18:24:49 | {} | MEMBER |
ClickHouse | ClickHouse | 526,899,309 | 6,707 | filimonov | Yes. That's one of the thing which is done bad (IMHO) in current http protocol implementation. And that multiquery will make even more mess there...
Please check other http related issues, lot of them mention that problem
https://github.com/yandex/ClickHouse/labels/comp-http
So IMHO we first need to work on th... | 2019-09-01T08:36:06 | 2019-09-03T20:46:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,730,111 | 6,735 | volfco | I'm going to close this as it looks like the result of unbalanced shards | 2019-09-04T04:02:46 | 2019-09-04T04:02:46 | {} | NONE |
ClickHouse | ClickHouse | 527,875,503 | 6,807 | proller | Do you really need llvm for your build?
Actual problem is:
CMake Error: The following variables are used in this project, but they are set to NOTFOUND.
Please set them or make sure they are set and tested correctly in the CMake files:
CYRUS_SASL_SHARED_LIB (ADVANCED)
linked by target "orc" in directory /home/m... | 2019-09-04T12:19:38 | 2019-09-04T12:19:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,885,260 | 6,399 | akuzm | The docs should be updated to describe the new settings and the new fields in system.clusters. | 2019-09-04T12:48:23 | 2019-09-04T12:48:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,939,540 | 6,818 | alexey-milovidov | CC @yuzhichang | 2019-09-04T14:57:19 | 2019-09-04T14:57:19 | {} | MEMBER |
ClickHouse | ClickHouse | 527,433,797 | 6,798 | den-crane | duplicate for https://github.com/yandex/ClickHouse/issues/5670
fixed https://github.com/yandex/ClickHouse/pull/5757
19.6 is not supported. 19.10, 19.11, 19.13 is
| 2019-09-03T12:20:06 | 2019-09-03T12:21:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,538,095 | 6,803 | den-crane | Or `WHERE dt != 0`
or `dt Nullable(DateTime)` + `where dt is not null`
```
┌──────────────────dt─┬─attr─┬─sum─┐
│ 1970-01-01 00:00:01 │ S102 │ 80 │
│ 1970-01-01 00:00:01 │ S100 │ 10 │
│ 1970-01-01 00:00:01 │ S101 │ 15 │
└─────────────────────┴──────┴─────┘
┌──────────────────dt─┬─attr─┬─sum─┐
│ 1970-01... | 2019-09-03T16:35:40 | 2019-09-03T16:38:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,555,733 | 6,638 | den-crane | reproduced with 19.14.1.1246 (official build)
uname -a
Linux XXXX 3.16.0-4-amd64 #1 SMP Debian 3.16.43-2+deb8u2 (2017-06-26) x86_64 GNU/Linux
Need **two** different CH hosts ( test_cluster_two_shards_localhost does not reproduce).
```
SELECT query_id AS f
FROM test.test6638
GROUP BY f
FORMAT Null
↘ Prog... | 2019-09-03T17:21:13 | 2019-09-03T17:23:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,587,511 | 6,796 | CurtizJ | It shouldn't be backported, because affects only 19.14. | 2019-09-03T18:44:11 | 2019-09-03T18:44:11 | {
"+1": 2
} | MEMBER |
ClickHouse | ClickHouse | 527,594,472 | 6,602 | alexey-milovidov | Is it still in progress? | 2019-09-03T19:02:17 | 2019-09-03T19:02:17 | {} | MEMBER |
ClickHouse | ClickHouse | 527,715,800 | 6,809 | qietingfengling | I have 16 cores per server. I observed that when running a query, there is often only a single cpu usage rate higher. Isn't clickhouse useful of all cpu resources? | 2019-09-04T02:43:55 | 2019-09-04T02:43:55 | {} | NONE |
ClickHouse | ClickHouse | 528,063,308 | 6,294 | jcdang | Is the preferred workflow to have X+1 requests or only one?
* For X+1 requests -- X number of submissions and another for listening for events?
* For one request -- The problem is that the default web browser implementation of EventSource does not support POST. https://github.com/mpetazzoni/sse.js has been able t... | 2019-09-04T20:05:24 | 2019-09-04T20:05:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,595,716 | 6,638 | akuzm | I added two remote hosts but didn't add the local one. I could reproduce it after adding the local host as well. | 2019-09-03T19:05:22 | 2019-09-03T19:05:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,197,158 | 6,760 | alexey-milovidov | @vzakaznikov You can add condition on `__GNUC__` version in code. | 2019-09-05T04:58:17 | 2019-09-05T04:58:23 | {} | MEMBER |
ClickHouse | ClickHouse | 528,197,267 | 6,760 | alexey-milovidov | But we already have it:
```
/// For the expansion of gtest macros.
#if defined(__clang__)
#pragma clang diagnostic ignored "-Wdeprecated"
#elif defined (__GNUC__) && __GNUC__ >= 9
#pragma GCC diagnostic ignored "-Wdeprecated-copy"
#endif
``` | 2019-09-05T04:58:52 | 2019-09-05T04:58:52 | {} | MEMBER |
ClickHouse | ClickHouse | 528,376,671 | 6,716 | ghost | Going to close this issue as I believe I was reading the benchmarking results incorrectly. While a local table query showed X QPS and the distributed query showed X/3, the reported total QPS across all nodes was X, which makes sense.
Thanks for your help. | 2019-09-05T13:56:07 | 2019-09-05T13:56:07 | {} | NONE |
ClickHouse | ClickHouse | 528,579,908 | 5,966 | abyss7 | Checked locally that all Kafka tests pass | 2019-09-05T21:01:00 | 2019-09-05T21:01:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,599,728 | 5,773 | alexey-milovidov | It requires to always use libc++ from contrib and it depends on using LLVM from contrib,
on the other hand, it depends on updating LLVM to at least version 8, and @orivej is currently doing it. | 2019-09-03T19:15:48 | 2019-09-03T19:15:59 | {} | MEMBER |
ClickHouse | ClickHouse | 527,903,862 | 4,944 | alesapin | Seems like it's a bug in obsolete Altinity build. Now rpm packages available in yandex repo https://repo.yandex.ru/clickhouse/rpm/stable/x86_64/. | 2019-09-04T13:36:24 | 2019-09-04T13:36:24 | {} | MEMBER |
ClickHouse | ClickHouse | 527,906,911 | 6,602 | amosbird | Now all singletons (except Suggest.h) are placed in dedicated TUs respectively. | 2019-09-04T13:43:47 | 2019-09-04T13:43:47 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 528,392,055 | 2,878 | javisantana | @RouR proposed workaround -> https://github.com/yandex/ClickHouse/issues/3238#issuecomment-528391725
| 2019-09-05T14:31:23 | 2019-09-05T14:31:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,442,076 | 6,814 | zhang2014 | @vitlibar
SmallestJSON DataType always splits JSON data into different column stores when writing
For example:
```
INSERT INTO xxx(smallest_json_column) VALUES('true')('1')('1.23')('-1')('"test_string"')
('{"url": "yandex.ru"}')('{"session_info": {"cookies":"abcdef"}}')
```
It will be split into the floowin... | 2019-09-05T16:08:39 | 2019-09-05T16:08:39 | {
"+1": 3
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,400,934 | 2,542 | filimonov | Also hit that today on 19.13.3, this time with mutations:
```
create table delete_test Engine=MergeTree() ORDER BY id AS SELECT number as id from numbers(10000);
create table delete_ids Engine=Set() AS SELECT number*100 from numbers(10);
ALTER TABLE delete_test
DELETE WHERE id IN (delete_ids)
Received... | 2019-09-03T10:28:51 | 2019-09-03T10:28:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,478,485 | 3,538 | leow3811 | Hello, buddy, I want to use kettle to store data from mysql to CK , Can you tell me how to do ? | 2019-09-03T14:16:40 | 2019-09-03T14:16:40 | {} | NONE |
ClickHouse | ClickHouse | 527,585,151 | 6,607 | vitlibar | Fixed in https://github.com/yandex/ClickHouse/pull/6082
The fix has been in the release starting from 19.11.4.24. | 2019-09-03T18:37:49 | 2019-09-03T18:39:25 | {} | MEMBER |
ClickHouse | ClickHouse | 527,586,459 | 6,667 | akuzm | Also seeing a marginal improvement for `SELECT SearchEngineID AS k, uniqExact(UserID) FROM hits_100m_single GROUP BY k` (0.58 vs 0.56 qps).
Will merge this now as discussed offline, because we are seeing slight improvements, no regressions and the rationale sounds correct. | 2019-09-03T18:41:18 | 2019-09-03T18:41:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,876,174 | 6,816 | alesapin | The existing integration test already verifies this patch. | 2019-09-04T12:21:37 | 2019-09-04T12:21:37 | {} | MEMBER |
ClickHouse | ClickHouse | 527,978,329 | 6,472 | alexey-milovidov | @xmariachi Yes, you should upgrade the remaining replicas. | 2019-09-04T16:23:49 | 2019-09-04T16:23:49 | {} | MEMBER |
ClickHouse | ClickHouse | 527,444,451 | 6,667 | akuzm | No meaningful changes for hll12:
```
SELECT 1 AS k, uniqHLL12(UserID) FROM hits_100m_single GROUP BY k
┏━━━━━━━━━━━━━━━┳━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━┓
┃ version ┃ count() ┃ median qps ┃ spread ┃
┡━━━━━━━━━━━━━━━╇━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━┩
│ ./ch-prefault │ 7 │ 12.652 │ 1.11 │
├───────... | 2019-09-03T12:51:52 | 2019-09-03T12:51:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,533,900 | 6,783 | BayoNet | Записал | 2019-09-03T16:24:28 | 2019-09-03T16:24:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,577,446 | 6,607 | alexey-milovidov | @vitlibar Please confirm that it is fixed in recent versions. | 2019-09-03T18:18:03 | 2019-09-03T18:18:03 | {} | MEMBER |
ClickHouse | ClickHouse | 527,578,394 | 6,802 | zhang2014 | @Akazz
How are the questions listed here for https://github.com/yandex/ClickHouse/pull/5097#issuecomment-524422831 considered? Sorry, I haven't replied since I haven't come up with a good idea | 2019-09-03T18:20:32 | 2019-09-03T18:20:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,043,698 | 6,399 | Enmk | Docs updated, settings renamed. | 2019-09-04T19:13:30 | 2019-09-04T19:13:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,243,130 | 6,666 | millin | I have 2 servers with 19.13 version and the same timezone in config, but ` timezone()` function returns a different result.
```
<timezone>Europe/Moscow</timezone>
```
First:
```
SELECT timezone()
┌─timezone()─┐
│ UTC │
└────────────┘
```
Second:
```
SELECT timezone()
┌─timezone()────┐
│ Euro... | 2019-09-05T07:46:08 | 2019-09-05T07:51:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,449,317 | 6,732 | foxit337 | 2G | 2019-09-05T16:27:19 | 2019-09-05T16:27:19 | {} | NONE |
ClickHouse | ClickHouse | 527,589,645 | 1,784 | den-crane | @xmariachi try
in clickhouse-client:
```
set optimize_throw_if_noop = 1;
OPTIMIZE TABLE tbl FINAL;
```
In this case you'll get a message with a reason why merge could not start. | 2019-09-03T18:50:19 | 2019-09-03T18:50:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,674,760 | 5,710 | alexey-milovidov | https://github.com/yandex/ClickHouse/commit/9959e8d3e1582458271e5f12902919fd0489689c
https://github.com/yandex/ClickHouse/commit/843f8303b019b3cb7d04ec20d4d0387a1e9f601a
https://github.com/yandex/ClickHouse/commit/ece7a05cec2c1eb5df2296e342f5c7ae8c9fa01d
https://github.com/yandex/ClickHouse/commit/617604193531842d5e... | 2019-09-03T23:09:52 | 2019-09-03T23:09:52 | {} | MEMBER |
ClickHouse | ClickHouse | 527,749,289 | 6,335 | filimonov | Maybe some check can be added... | 2019-09-04T05:41:13 | 2019-09-04T05:41:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,214,054 | 6,818 | yuzhichang | @joebling, I have the following questions:
- What's the relationship between `fd_qmkg_abtest_kboss_user_locate` and `fd_qmkg_abtest_kboss_user_locate_bitmap`?
- Is `uid` an UInt32 column?
- Can you try following query?
`select bitmapCardinality(groupBitmapMergeState(uid_bitmap)) as nuid_bitmap from qmkg_abtest.f... | 2019-09-05T06:12:37 | 2019-09-05T06:34:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,227,573 | 3,538 | leow3811 | i have dealt with this problem ,thanks anyway | 2019-09-05T06:59:11 | 2019-09-05T06:59:11 | {} | NONE |
ClickHouse | ClickHouse | 528,347,237 | 6,232 | abyss7 | Backported to 19.12 with https://github.com/yandex/ClickHouse/commit/2e18a7d2c393762f0888f1c4815a2c195971629e | 2019-09-05T12:48:26 | 2019-09-05T12:48:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,635,108 | 4,736 | filimonov | @borismol can you please check it on 19.11.9 (just released, it has important Kafka fixes)? | 2019-09-05T23:04:37 | 2019-09-05T23:04:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,581,856 | 6,800 | alexey-milovidov | BTW, Conditions for row-level security don't use index at all.
To mitigate this issue, the user should provide index conditions in a query regardless to row-level security settings.
This issue should be addressed nevertheless. | 2019-09-03T18:29:11 | 2019-09-03T18:29:11 | {} | MEMBER |
ClickHouse | ClickHouse | 527,638,001 | 6,294 | filimonov | We can start with simpler and cleaner options, improving / adding complexity later:
1. Brotli compression - we definitely can start w/o that
2. Buffering - in general is good as will allow to use socket more efficiently, but from the other side if we will not flush it for longer time - client will not get any events.... | 2019-09-03T21:00:37 | 2019-09-03T21:06:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,929,504 | 6,817 | alexey-milovidov | It means that the object was allocated with different allocator and then freed with jemalloc.
The root cause looks like the order of loading of shared libraries (this is one of the typical issues with shared linking).
Possible solutions:
- fix the order of loading shared libraries;
- always link jemalloc statical... | 2019-09-04T14:35:09 | 2019-09-04T14:35:09 | {} | MEMBER |
ClickHouse | ClickHouse | 528,373,860 | 6,827 | alexey-milovidov | The motivation is clear, we need to do it. | 2019-09-05T13:49:35 | 2019-09-05T13:49:35 | {} | MEMBER |
ClickHouse | ClickHouse | 528,439,942 | 6,807 | ildus | Yeah, problem with SASL is there, but still, it could not find LLVM. | 2019-09-05T16:03:12 | 2019-09-05T16:03:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,441,944 | 5,773 | proller | Please make submodule | 2019-09-05T16:08:17 | 2019-09-05T16:08:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,450,174 | 6,784 | zhang2014 | @BayoNet It can only work in `equals`, `notEquals`, `in` and `notIn` function at present. | 2019-09-05T16:29:41 | 2019-09-05T16:29:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,395,357 | 6,778 | filimonov | Try that:
```
alter table SomeTable update Device=CAST('Other', 'Enum8(\'Other\' = 0, \'Phone\' = 1, \'Tablet\' = 2, \'Desktop\' = 3)') ...
``` | 2019-09-03T10:10:51 | 2019-09-03T10:10:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,408,298 | 6,795 | filimonov | Can you show a sample row of data and table schema? | 2019-09-03T10:53:25 | 2019-09-03T10:53:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,515,255 | 6,661 | den-crane | @bespilotnayakartoshka It's very unusual to use Distributed tables in this workflow.
Can you show my_cluster description? Are you sure you used `<internal_replication>true</internal_replication>` and Distributed is not the root cause of the issue?
| 2019-09-03T15:39:50 | 2019-09-03T15:39:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,590,912 | 6,802 | Akazz | > @Akazz
> How are the questions listed here for [#5097 (comment)](https://github.com/yandex/ClickHouse/pull/5097#issuecomment-524422831) considered?
We decided to proceed with your PR because it fixed deadlock scenarios as much as for Log + StripeLog engines. Commits are to follow soon
> Sorry, I haven't replie... | 2019-09-03T18:53:39 | 2019-09-03T18:54:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,052,849 | 6,817 | abyss7 | The order of shared libraries should be important only if there are some allocations before `main()`. Do we have any?
Even if there are some, I think we should fix the build in the way that all shared libraries will be linked directly with jemalloc. | 2019-09-04T19:36:43 | 2019-09-04T19:36:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,502,314 | 1,784 | xmariachi | I upgraded servers to match the same version. I no longer see those errors.
However, the behaviour is still the same: data doesn't get compacted.
The query finishes fast:
```
OPTIMIZE TABLE tbl FINAL
Ok.
0 rows in set. Elapsed: 2.502 sec.
```
However, the table does not do data compaction, or at least no... | 2019-09-05T17:57:27 | 2019-09-05T17:57:39 | {} | NONE |
ClickHouse | ClickHouse | 528,525,610 | 6,050 | fessmage | I got NOT_LOADED state for every dictionary after update of clickhouse-server from 18.* to 19.* and only thing that worked for me - advice from @vitlibar. `system reload dictionary <dictionary_name>` for every dictionary - fixes problem. | 2019-09-05T18:56:39 | 2019-09-05T18:56:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,403,439 | 6,629 | CurtizJ | Kafka integration tests are still broken in master. | 2019-09-03T10:37:01 | 2019-09-03T10:37:47 | {} | MEMBER |
ClickHouse | ClickHouse | 527,463,407 | 6,661 | mikekytyzov | @filimonov up
on 19.13.3.26 it like 50% of dups | 2019-09-03T13:41:10 | 2019-09-03T13:41:10 | {} | NONE |
ClickHouse | ClickHouse | 527,465,628 | 6,294 | jcdang | What are the features that need to be there in this streaming API?
- Compression -- Is Brotli really needed?
- Is buffering to file needed if we're flushing frequently?
- What are the event types we want and what do they produce?
summary, progress,...
Here is my initial go at it:
https://github.com/jcdang/Cli... | 2019-09-03T13:46:36 | 2019-09-03T13:46:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,825,779 | 5,568 | gle4er | > May be related to #4167 and #3983
Yes, #3983 is same.
> What build do you use?
I use official build:
```
$ clickhouse server -V
ClickHouse server version 19.7.3.9 (official build).
```
Also, this is build from docker registry. | 2019-09-04T09:40:21 | 2019-09-04T09:41:36 | {} | NONE |
ClickHouse | ClickHouse | 527,995,193 | 6,822 | den-crane | ABC ?
```
select parseDateTimeBestEffort( '2018/07/03T08:48:01')
┌─parseDateTimeBestEffort('2018/07/03T08:48:01')─┐
│ 2018-07-03 08:48:01 │
└────────────────────────────────────────────────┘
```
https://clickhouse.yandex/docs/en/operations/settings/settings/#settings-date_time_inpu... | 2019-09-04T17:07:54 | 2019-09-04T17:08:04 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,087,335 | 6,760 | vzakaznikov | Here are the details to reproduce the issue with GCC 8 trying to build the latest stable
release v19.13.3.26-stable.
The build env is described by the following Dockerfile
```
FROM ubuntu:18.04
RUN apt-get update -y \
&& apt-get install -y software-properties-common \
&& add-apt-repository ppa:u... | 2019-09-04T21:06:59 | 2019-09-04T21:06:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,391,725 | 3,238 | javisantana | Proposed workaround, create a matview from a matview:
```
:) create table A (a Int32) Engine=MergeTree() order by tuple();
:) create materialized view B Engine=MergeTree() order by tuple() as select a*2 as b from A where a > 10;
:) create materialized view C Engine=MergeTree() order by tuple() as select b*2 as ... | 2019-09-05T14:30:42 | 2019-09-05T14:31:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,623,028 | 4,497 | filimonov | Fixed in https://github.com/yandex/ClickHouse/commit/44d3e1e837fd5e804c2d0452e8a8f1b4af6f93f8 | 2019-09-05T22:54:04 | 2019-09-05T22:54:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,581,515 | 1,784 | xmariachi | I found the same issue here, though in my case I am finding duplicates in a ReplicatedReplacingMergeTree even when running it as `OPTIMIZE TABLE tbl FINAL`.
Optimization returns OK but no changes are made on duplicates, still the same number. Is this because of being a Replicated* table?
`Connected to ClickHouse serv... | 2019-09-03T18:28:13 | 2019-09-03T18:29:16 | {} | NONE |
ClickHouse | ClickHouse | 527,591,216 | 6,760 | Enmk | @vzakaznikov could you please share the command you are running to set the compiler and do the build ? | 2019-09-03T18:54:28 | 2019-09-03T18:54:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,632,694 | 6,707 | filimonov | 1. With server-sent-events - all that extra metadata should be send in events, not in headers. And it should be much easier to implement multiquery / final stats / progress and other messages etc. It's possible that binary formats will be send less efficiently in that case (all newlines should be replaced to `\ndata: `... | 2019-09-03T20:46:08 | 2019-09-03T20:46:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,834,290 | 6,812 | alesapin | Can you provide data or create a reproducible example? | 2019-09-04T10:04:06 | 2019-09-04T10:04:06 | {} | MEMBER |
ClickHouse | ClickHouse | 527,876,031 | 6,710 | MrBuddyCasino | Could this be useful for migrations? CH executes them very fast, but the number of migrations is always growing, and avoiding many roundtrips might safe us time for executing integration tests. | 2019-09-04T12:21:15 | 2019-09-04T12:21:15 | {} | NONE |
ClickHouse | ClickHouse | 527,912,906 | 4,944 | den-crane | dup: https://github.com/yandex/ClickHouse/issues/6401
fixed: https://github.com/yandex/ClickHouse/pull/6409 | 2019-09-04T13:58:19 | 2019-09-04T13:58:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,844,747 | 5,097 | zhang2014 | Continued by @Akazz in #6802 | 2019-09-04T10:36:38 | 2019-09-04T10:36:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 528,059,805 | 6,817 | alexey-milovidov | I am sure that we have allocations before `main`. | 2019-09-04T19:55:49 | 2019-09-04T19:55:49 | {} | MEMBER |
ClickHouse | ClickHouse | 528,658,223 | 6,835 | alexey-milovidov | orc_load test has failed. | 2019-09-06T00:27:45 | 2019-09-06T00:27:45 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.