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 | 660,335,466 | 12,570 | alexey-milovidov | ```
#0 __libc_write (nbytes=9437145, buf=0x7ffa24980700, fd=1) at ../sysdeps/unix/sysv/linux/write.c:26
#1 __libc_write (fd=1, buf=0x7ffa24980700, nbytes=9437145) at ../sysdeps/unix/sysv/linux/write.c:24
#2 0x00000000105958d5 in replxx::Terminal::write8 (this=<optimized out>,
data_=0x7ffa24980700 "SELECT \0... | 2020-07-17T20:59:52 | 2020-07-17T20:59:52 | {} | MEMBER |
ClickHouse | ClickHouse | 660,890,419 | 12,442 | qoega | Probably we need to fix docs also
`grep 0000-00-00`
```
fr/sql-reference/data-types/date.md
fr/sql-reference/statements/create.md
fa/operations/system-tables.md
fr/operations/system-tables.md
fr/interfaces/formats.md
fa/sql-reference/data-types/date.md
fa/sql-reference/statements/create.md
en/sql-reference/st... | 2020-07-20T08:41:32 | 2020-07-20T08:41:32 | {} | MEMBER |
ClickHouse | ClickHouse | 661,035,342 | 12,571 | 4ertus2 | I've tryed to reproduce the issue on my laptop. First of all it returns some data immediately. But I think there's no guaranty for it.
```
│ 1294337 │ 1304334 │
│ 1294337 │ 1304335 │
│ 1294337 │ 1304336 │
└─────────┴─────────┘
Cancelling query.
Showed first 10000.
Query was cancelled.
110000000 rows in s... | 2020-07-20T13:20:58 | 2020-07-20T13:20:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 660,174,278 | 12,566 | alesapin | Also, setting `max_threads_for_updates` cannot be tuned both from SQL query and xml config because of https://github.com/ClickHouse/ClickHouse/blob/master/src/Dictionaries/DictionaryFactory.cpp#L37-L41. | 2020-07-17T15:35:17 | 2020-07-17T15:35:17 | {} | MEMBER |
ClickHouse | ClickHouse | 660,344,896 | 11,546 | den-crane | @RonanMorgan you can't set min_insert_block_size_rows / max_insert_block_size in server. They should be set as query properties in JDBC query.
I think your issue is because of JSONEachRow / parralel parsing, try to set `input_format_parallel_parsing=0` in JDBC | 2020-07-17T21:27:45 | 2020-07-17T21:27:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 660,655,960 | 11,933 | den-crane | >2809 partititions
You mean parts? `select count() from system.parts where active=1 and table='table 2'`
Did you try to increase `-Djute.maxbuffer` https://github.com/ClickHouse/ClickHouse/issues/11933#issuecomment-654286261 | 2020-07-19T14:45:04 | 2020-07-19T14:45:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 660,663,023 | 11,933 | den-crane | As I can see you have by hour partitions. CH was designed for monthly partitioning.
What was your goal to use hour partitions?
And what is the data retention? For how long do you store your data (week ? month?)?
With such partitioning you will have a hard journey.
| 2020-07-19T15:25:50 | 2020-07-19T15:26:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 661,006,459 | 11,933 | keliss | We do data aggregation hourly and it is convenient to do the partitioning using the same time interval. Plus querying the database for data for some hour should be faster if I understand correctly how it works. The data retention is set to be three months generally. But can you please tell me what are the disadvantages... | 2020-07-20T12:39:54 | 2020-07-20T12:39:54 | {} | NONE |
ClickHouse | ClickHouse | 661,010,350 | 8,406 | KochetovNicolai | Must have been fixed by #8406 | 2020-07-20T12:43:11 | 2020-07-20T12:43:11 | {
"confused": 1
} | MEMBER |
ClickHouse | ClickHouse | 661,269,657 | 12,589 | KochetovNicolai | Oops, set index is not thread-safe because of this:
https://github.com/ClickHouse/ClickHouse/blob/3854ce6d8467ca29b7261f9420e25d8551a19bd3/src/Interpreters/Set.cpp#L476-L478
@CurtizJ how do you think it will affect performance if we move those lines back to `checkInRange`? | 2020-07-20T18:50:08 | 2020-07-20T18:50:08 | {} | MEMBER |
ClickHouse | ClickHouse | 661,615,239 | 12,583 | JTCunning | @den-crane: Thank you very much for the quick response and reproduction.
I'm going to hold off on setting any increased limit in the efforts of upgrading. After reading your script, it seems as if the only trigger to reset `MemoryTracking` is an exception, and we get along with 20.3 just fine on these machines and 2... | 2020-07-21T03:50:36 | 2020-07-21T03:50:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 660,044,557 | 12,541 | alexey-milovidov | @KochetovNicolai Should backport to 20.6. | 2020-07-17T11:04:37 | 2020-07-17T11:04:37 | {} | MEMBER |
ClickHouse | ClickHouse | 660,216,398 | 12,566 | alesapin | If we remove these two unlockers https://github.com/ClickHouse/ClickHouse/blob/master/src/Dictionaries/CacheDictionary.cpp#L882-L892 than it should work as expected, but the whole idea of non-blocking read will be broken.
When I add more granular locks (both read for update and write for rest) my server hangs withou... | 2020-07-17T16:45:49 | 2020-07-17T16:48:18 | {} | MEMBER |
ClickHouse | ClickHouse | 660,536,168 | 12,570 | alexey-milovidov | The issue reproduces even if `mc` is run under `kitty`. | 2020-07-18T20:14:29 | 2020-07-18T20:14:29 | {} | MEMBER |
ClickHouse | ClickHouse | 660,766,966 | 11,955 | bobrik | @alexey-milovidov this sped up my query from 4.250s to 3.911s 🎉 | 2020-07-20T02:35:00 | 2020-07-20T02:35:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 661,057,622 | 12,571 | den-crane | @4ertus2 it does not.
The right table is huge
formatReadableSize(sum(data_uncompressed_bytes))
19.17 GiB
it does not fit into memory, join with a single row (left table)
version 20.7.1
set max_memory_usage = 30000000000
set join_algorithm = 'hash'
SET send_logs_level = 'trace' ;
select q.id, p.id ... | 2020-07-20T13:59:39 | 2020-07-20T13:59:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 660,158,048 | 12,564 | den-crane | https://github.com/ClickHouse/ClickHouse/issues/9115#issuecomment-586534439 | 2020-07-17T15:05:13 | 2020-07-17T15:05:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 660,170,605 | 12,459 | akuzm | Now it shows positives very often -- this is expected, because the queries were unstable with up to 40%. I added max_ignored_relative_change=0.4 in master. | 2020-07-17T15:28:37 | 2020-07-17T15:28:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 660,954,105 | 1,118 | KochetovNicolai | > I want to know the explain that quering on cluster
Explain for queries on cluster is not implemented yet. This feature will be added later.
> what is ParallelAggregating,what is PartialSorting and Concat
This is a different parts of query pipeline. Those classes may be found in the source code and are useful... | 2020-07-20T10:50:06 | 2020-07-20T10:50:06 | {} | MEMBER |
ClickHouse | ClickHouse | 660,992,008 | 12,327 | ghost | @nikitamikhaylov
I fixed codes and problems with style.
| 2020-07-20T12:19:14 | 2020-07-20T12:19:32 | {} | NONE |
ClickHouse | ClickHouse | 661,296,113 | 12,571 | 4ertus2 | OK, it looks I've reproduced the issue.
@gkristic could you run the query with `max_threads = 1` set. Would it fail with it or not?
```
select q.id, p.id, p.name
from xs q, xs p
where startsWith(q.name, p.name)
settings max_threads = 1
``` | 2020-07-20T19:45:29 | 2020-07-20T19:48:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 661,310,032 | 11,933 | den-crane | repro:
```
20.7.1
CREATE TABLE testAlter( A Int64, T DateTime, X Float64)
ENGINE = ReplicatedReplacingMergeTree ('/clickhouse/testAlter','xxx')
partition by toStartOfHour(T) ORDER BY A;
set max_partitions_per_insert_block=0
insert into testAlter select number, toDateTime('2000-01-01 00:00:00') + numbe... | 2020-07-20T20:15:32 | 2020-07-20T21:11:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 661,701,510 | 12,612 | alexey-milovidov | It is not officially supported and will not compile and work out of the box.
It should be possible with some modifications.
As far as I know, no one tried it yet.
Adding a new platform (64bit little endian platform with modern compilers support) for ClickHouse is mostly straightforward if you disable some librarie... | 2020-07-21T08:01:53 | 2020-07-21T08:01:53 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 661,724,678 | 12,613 | alexey-milovidov | OUTPUT_IS_NOT_SORTED can be removed from ErrorCodes.cpp too. | 2020-07-21T08:48:57 | 2020-07-21T08:48:57 | {} | MEMBER |
ClickHouse | ClickHouse | 660,530,202 | 12,570 | alexey-milovidov | Looks like a bug in "Konsole" terminal emulator that I use.
If I use "Kitty" it works perfectly. | 2020-07-18T19:26:43 | 2020-07-18T19:26:43 | {} | MEMBER |
ClickHouse | ClickHouse | 660,565,691 | 7,404 | alexey-milovidov | Now it works as you expect. | 2020-07-19T00:58:02 | 2020-07-19T00:58:02 | {} | MEMBER |
ClickHouse | ClickHouse | 661,030,743 | 12,540 | azat | >@azat do we have opened issues now?
Yep:
```
Code: 49, e.displayText() = DB::Exception: Pipeline stuck. Current state:
digraph
{
n140039724560416[label="Remote(257 jobs) (Finished)"];
n140046543991576[label="Converting(256 jobs) (Finished)"];
n140046543993816[label="ExpressionTransform(256 jobs) (Finished)... | 2020-07-20T13:12:46 | 2020-07-20T13:12:46 | {} | MEMBER |
ClickHouse | ClickHouse | 661,263,392 | 12,594 | 4ertus2 | We should fix it but it's not a very good idea to use dots in db or table name in production. Dots are used for database and table split, table and column split, column and nested column split. All this things depends on context where the name is present. And sometimes we do not know the contexts. | 2020-07-20T18:37:31 | 2020-07-20T18:37:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 661,589,769 | 12,583 | den-crane | It seems memory accounting (MemoryTracking) is still wrong for inserts.
@JTCunning you can set
<max_server_memory_usage_to_ram_ratio>2</max_server_memory_usage_to_ram_ratio>
related https://github.com/ClickHouse/ClickHouse/issues/11153
it was wrong in 20.3.9.70 as well but there was no total limit (max_serve... | 2020-07-21T02:48:31 | 2020-07-21T02:52:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 660,394,981 | 12,510 | ritaank | The posted `CANNOT INSERT` issue seems to be a wraparound case as resolved in #11581.
However, this [comment](https://github.com/ClickHouse/ClickHouse/issues/11581#issuecomment-642618638) from that same issue says that returning exitcode 0 is a bug. This happens in the `CREATE ROLE` comment above. | 2020-07-18T00:43:40 | 2020-07-18T00:43:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 660,924,212 | 12,573 | 4ertus2 | Hi,
I.e. we have recent stable version 'v20.5.2.7-stable'. You may read it as as 'the 5th version in 20 year, patchset 2'. '7' is some internal number. Do not look at it. `stable` here means we have looked at it for some time and it's expected to be stable enough.
The Major in `Year.Major.Minor.SomeNumber` means ... | 2020-07-20T09:46:34 | 2020-07-20T09:46:34 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 661,245,181 | 12,600 | vitlibar | The `SYSTEM` privilige can be granted on global, database, or table level. It cannot be granted on column level because it doesn't make sense.
The following should work:
```
GRANT SYSTEM ON db0.* TO user0
```
The command `REVOKE SYSTEM(x) ON db0.table0 FROM user0` doesn't throw an exception but it does nothing... | 2020-07-20T18:01:25 | 2020-07-20T18:01:53 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 661,250,829 | 12,588 | den-crane | duplicate https://github.com/ClickHouse/ClickHouse/issues/5125 | 2020-07-20T18:12:36 | 2020-07-20T18:12:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 661,313,511 | 11,933 | den-crane | >Plus querying the database for data for some hour should be faster if I understand correctly how it works.
Only if your query processes data for only single digit number of hours.
>But can you please tell me what are the disadvantages
Big number of partitions -> parts ->
* slow CH start (CH reads every par... | 2020-07-20T20:23:05 | 2020-07-20T20:23:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 661,573,601 | 12,596 | Johnhz | > What CH version do you use?
> Can you provide table DDL? Do you have data sample?
ClickHouse client version 20.3.8.53.
DDL detail is :
CREATE TABLE default.p_event_all (distinct_id String, event String, time Int64, staticdate Date, event_date_time DateTime, material_present_price Nullable(Float64), vip_open_p... | 2020-07-21T02:24:54 | 2020-07-21T02:24:54 | {} | NONE |
ClickHouse | ClickHouse | 660,043,717 | 12,541 | alexey-milovidov | I'm Ok with this change.
It effectively removes default format for EXPLAIN query that was set in parser level. And I think we should not have done it this way, because default format can be also overriden in clickhouse-client config, in URL parameter, etc. | 2020-07-17T11:03:02 | 2020-07-17T11:03:02 | {} | MEMBER |
ClickHouse | ClickHouse | 660,156,901 | 12,565 | den-crane | you can apply it to the table
```
alter table foobar modify setting max_parts_in_total = 6000;
``` | 2020-07-17T15:03:04 | 2020-07-17T15:03:04 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 660,271,315 | 12,522 | akuzm | https://clickhouse-test-reports.s3.yandex.net/12522/a4a23673cc640ab77efc9315a919519a9e49ed89/functional_stateless_tests_(debug)/test_run.txt.out.log
`Logical error: 'Sort order of blocks violated for column ``, left: UInt64_255, right: UInt64_0.'.`
Looks like it's not related to the changes. | 2020-07-17T18:28:13 | 2020-07-17T18:28:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 661,319,012 | 12,571 | gkristic | Sure! I've just run that query with the same setup as before. It produced a first batch of results rather quickly. (It returned the first 10K rows.) Then it kept processing but without printing anything else other than the stats, with understandably decreasing rates for rows/s and MB/s. The output is not comfortable to... | 2020-07-20T20:34:36 | 2020-07-20T20:34:36 | {} | NONE |
ClickHouse | ClickHouse | 660,169,658 | 12,563 | den-crane | it could be mark cache (5GB by default). Try `system drop mark cache` instead of a reboot and check memory after it. | 2020-07-17T15:26:47 | 2020-07-17T15:27:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 660,646,807 | 11,933 | keliss | Hello everyone. I'm working on a project that Alexey used to work on. @den-crane here are a few examples of the output:
```
table 1, 2809 partititions
2020-04-20 13:00:00,1587387600_1_1_0
2020-04-20 14:00:00,1587391200_1_1_0
2020-04-20 15:00:00,1587394800_1_1_0
2020-04-20 16:00:00,1587398400_1_1_0
2020-04-20 17:... | 2020-07-19T13:53:24 | 2020-07-19T13:53:24 | {} | NONE |
ClickHouse | ClickHouse | 660,925,953 | 5,591 | simPod | https://clickhouse.tech/docs/en/sql-reference/data-types/lowcardinality/ | 2020-07-20T09:50:10 | 2020-07-20T09:50:10 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 661,023,763 | 12,540 | KochetovNicolai | > "Pipeline stuck" is still possible from time to time
@azat do we have opened issues now?
All the cases I remember were logical and reproducable without fuzzer. I suppose if tests could find such a problem, we would see it randomly, as tests are executed all the time.
I see that tests time for current pr is 0... | 2020-07-20T12:59:43 | 2020-07-20T12:59:43 | {} | MEMBER |
ClickHouse | ClickHouse | 661,280,870 | 12,571 | 4ertus2 | My bad. I've made something wrong with population scripts and get empy strings `name` inside the table. It should fail on my laptop with 2kb strings.
> 10 mil rows * 2kb = raw data 20 GB, it will take 100GB RAM in a hash table
It does not need hash table for CROSS JOIN, columns only. So probably 32Gb in query log... | 2020-07-20T19:12:33 | 2020-07-20T19:12:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 661,578,637 | 12,596 | Johnhz | > What CH version do you use?
> Can you provide table DDL? Do you have data sample?
Starting time 2020-07-13 Ending time 2020-07-19 The total amount of data between 2020-07-19 is 920240984. If the data for only one day is fixed, the data for multiple days will be approximated. If I want to provide sample data, it i... | 2020-07-21T02:32:13 | 2020-07-21T02:32:13 | {} | NONE |
ClickHouse | ClickHouse | 660,013,632 | 11,546 | RonanMorgan | Hello,
I have a similar issue,
Caused by: java.lang.Throwable: Code: 241, e.displayText() = DB::Exception: Memory limit (total) exceeded: would use 28.06 GiB (attempt to allocate chunk of 4382610 bytes), maximum: 28.06 GiB **(version 20.5.2.7** (official build))
I am using the jdbc official driver to save 1 500K... | 2020-07-17T10:06:17 | 2020-07-17T13:33:41 | {} | NONE |
ClickHouse | ClickHouse | 658,247,316 | 10,310 | den-crane | ```
SELECT
now64() AS x,
toFloat64(x) AS y,
toUInt64(replace(toString(y), '.', '')) AS w,
toDateTime64(w / 1000, 3) AS z
┌───────────────────────x─┬──────────────y─┬─────────────w─┬───────────────────────z─┐
│ 2020-07-14 15:28:55.863 │ 1594740535.863 │ 1594740535863 │ 2020-07-14 15:28:55.863 │
... | 2020-07-14T15:29:17 | 2020-07-17T14:57:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 660,267,415 | 11,546 | RonanMorgan | ok I fixed my issue by using the batch API of the jdbc driver instead of the writer one, I still don't understand the error log though (maybe an overhead when writting rows one by one) | 2020-07-17T18:19:48 | 2020-07-17T18:20:31 | {} | NONE |
ClickHouse | ClickHouse | 661,111,334 | 11,428 | PhantomPhreak | still present in v20.5 | 2020-07-20T15:29:32 | 2020-07-20T15:29:32 | {} | NONE |
ClickHouse | ClickHouse | 660,044,212 | 12,541 | alexey-milovidov | > Yandex synchronization check
Known issue. | 2020-07-17T11:03:57 | 2020-07-17T11:03:57 | {} | MEMBER |
ClickHouse | ClickHouse | 660,174,676 | 12,565 | kekekedeng | @den-crane It works, thanks! | 2020-07-17T15:35:59 | 2020-07-17T15:35:59 | {} | NONE |
ClickHouse | ClickHouse | 660,565,514 | 10,522 | alexey-milovidov | This is how it is documented:
> When ClickHouse merges data parts, it deletes each pair of rows that have the same primary key and version and **different Sign**. The order of rows does not matter.
When you SELECT with FINAL, the logic is the same that for background merge. | 2020-07-19T00:56:27 | 2020-07-19T00:56:27 | {} | MEMBER |
ClickHouse | ClickHouse | 660,800,509 | 11,564 | bobrik | In #12589 I attempt to parallelize skipping indices and PK selection stages between parts. | 2020-07-20T04:52:18 | 2020-07-20T04:52:18 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 660,955,937 | 12,572 | 4ertus2 | Hi,
In short, we are interested in combined CPU+GPU solution.
For now it's a research task. #10232 is an attempt to use OpenCL and #11083 is attempt to use CUDA.
Theoretically the main trouble with GPU is to move data to its memory and back. In practive there're some other issues to solve.
1. CUDA needs its o... | 2020-07-20T10:54:14 | 2020-07-20T10:56:08 | {
"+1": 3
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 661,195,718 | 12,587 | KochetovNicolai | Another approach is to implement it via new type column + expression analyzer.
We may add new type of column: `ColumnProjection(inner_column, mask)`, where `inner_column` may be column of any type, and `mask` is a boolean mask. `ColumnProjection.size() == mask.size()` and `inner_column.size() == sum(mask)`.
Also ... | 2020-07-20T17:04:42 | 2020-07-20T17:04:42 | {
"+1": 2
} | MEMBER |
ClickHouse | ClickHouse | 661,322,618 | 12,571 | gkristic | I ran it again with `format Null` so that I could capture the trace without all the interference from the annoying X's, in case it gives you more information. Available [here](https://github.com/ClickHouse/ClickHouse/files/4949983/ch-trace-1thread.txt). | 2020-07-20T20:42:35 | 2020-07-20T20:42:35 | {} | NONE |
ClickHouse | ClickHouse | 661,642,092 | 12,586 | fpacifici | Indeed, this way it works. So I guess it is expected, right ? | 2020-07-21T05:28:44 | 2020-07-21T05:28:44 | {} | NONE |
ClickHouse | ClickHouse | 660,566,255 | 7,660 | alexey-milovidov | @excitoon No answer? | 2020-07-19T01:02:13 | 2020-07-19T01:02:13 | {} | MEMBER |
ClickHouse | ClickHouse | 661,605,660 | 12,583 | den-crane | @azat
Repro:
20.7.1.4127
http
```
clickhouse-client -q "select arrayStringConcat(arrayMap(x->toString (cityHash64(x)) , range(1000)),' ') from numbers(100000);" >test.json
clickhouse-client -q "CREATE TABLE log (row String) ENGINE = Null"
for i in `seq 1 50`; do curl -s "http://localhost:8123/?query=IN... | 2020-07-21T03:20:02 | 2020-07-21T03:24:11 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 660,530,528 | 12,570 | alexey-milovidov | But if I paste into `bash` in `Konsole` it also works.
Maybe because we issue a single `write` syscall with data of about 8 MB size and it triggers a bug in `Konsole`. | 2020-07-18T19:28:48 | 2020-07-18T19:28:48 | {} | MEMBER |
ClickHouse | ClickHouse | 660,535,357 | 12,570 | alexey-milovidov | Actually it is a bug in "midnight commander" (`mc`).
It only reproduces if I paste query inside `clickhouse-client` from `bash` that is run by `mc`. | 2020-07-18T20:07:34 | 2020-07-18T20:07:34 | {} | MEMBER |
ClickHouse | ClickHouse | 660,747,366 | 12,539 | spihiker | 你好。升级后，观察了几天。没有再出现闪退的情况。 | 2020-07-20T01:21:22 | 2020-07-20T01:22:05 | {} | NONE |
ClickHouse | ClickHouse | 657,430,737 | 12,423 | vitlibar | Waiting for https://github.com/simdjson/simdjson/pull/1037 - done
Waiting for https://github.com/simdjson/simdjson/pull/1042 - done
Waiting for https://github.com/simdjson/simdjson/pull/1044 - done
| 2020-07-13T09:12:20 | 2020-07-20T07:15:52 | {} | MEMBER |
ClickHouse | ClickHouse | 660,925,098 | 12,485 | Jokser | Btw, all of these options are not used in the current HTTP client yet.
I think it makes sense to add real usage of them as well. | 2020-07-20T09:48:24 | 2020-07-20T09:48:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 660,952,065 | 10,672 | Dexig | Hello.
You can add support for the **toStartOf** family of functions for **DateTime64**?
Because if you use these functions with **DateTime64** now, then there is an implicit conversion to just **DateTime**.
| 2020-07-20T10:45:41 | 2020-07-20T10:45:50 | {} | NONE |
ClickHouse | ClickHouse | 661,103,454 | 12,571 | 4ertus2 | @den-crane
There's a message that right table has been loaded into memory and we've got out of memory in join phase not in right table insertion.
```
[8c0c3408045c] 2020.07.17 20:13:06.040223 [ 105 ] {927d1881-3b29-4a59-8725-92f1f8a453ac} <Debug> CreatingSetsBlockInputStream: Created Join with 10000000 entries ... | 2020-07-20T15:16:30 | 2020-07-20T15:16:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 661,264,647 | 12,540 | KochetovNicolai | I've built graph from exception message and suppose understand now why this happens.
It's pretty rare case for external sorting when intermediate data was just written into file buffer, and no more new chunks were added. Will try to reproduce it manually. | 2020-07-20T18:39:54 | 2020-07-20T18:39:54 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 661,291,696 | 12,589 | CurtizJ | It will affect performance not so much. I think parallelizing will compensate it. | 2020-07-20T19:35:56 | 2020-07-20T19:35:56 | {} | MEMBER |
ClickHouse | ClickHouse | 661,484,687 | 12,598 | akuzm | Is it expected that the setting does not affect the queries with `length(URL)`? | 2020-07-21T00:25:55 | 2020-07-21T00:25:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 660,551,913 | 12,575 | ritaank | using `CREATE {USER, PROFILE, QUOTA} default` may create similar behavior. (2 objects named `default`). I have not tried this.
On a related note, is there an easier way to undo this (instead of wiping and reinstalling)? | 2020-07-18T22:32:36 | 2020-07-18T22:32:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 660,831,015 | 7,660 | excitoon | > @filimonov Isn't it fixed?
It was fixed in #8402 | 2020-07-20T06:28:46 | 2020-07-20T06:28:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 660,950,448 | 12,594 | qoega | @vitlibar I remember you faced this problem with dictionaries. Will you fix it for remote also? | 2020-07-20T10:41:58 | 2020-07-20T10:41:58 | {} | MEMBER |
ClickHouse | ClickHouse | 661,155,295 | 12,599 | zhang2014 | This [PR](https://github.com/ClickHouse/ClickHouse/pull/10851) has synchronized the database. Of course the table level approach is also very useful, it's more flexible.
But maybe there's another way to do it(that's only my idea, and it worked in #10851):
``` sql
CREATE DATABASE test ENGINE = MaterializeMySQL('loc... | 2020-07-20T16:27:03 | 2020-07-20T16:38:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 661,227,410 | 12,587 | 4ertus2 | @KochetovNicolai The idea of masked execution is cool. But I cannot understand why we need a new column type here. We colud use ColumnFunction.
Expression: `if(x > 0, a + b, a + c)`
Normal case:
Block before `if()`: `a, b, c, add(a, b), add(a, c)`
Short circuit case:
Block before `if()`: `a, b, c, func('add'... | 2020-07-20T17:36:32 | 2020-07-20T17:50:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 661,245,215 | 12,587 | KochetovNicolai | @4ertus2 Approach with `ColumnFunction` is also good.
I cannot say which one is better by now. I think projections are more explicit. And if you create `ColumnFunction`, you also create separate subexpression for it. It may make some possible, imaginary expression-level optimisations more difficult (or not).
Also, ... | 2020-07-20T18:01:31 | 2020-07-20T18:01:31 | {} | MEMBER |
ClickHouse | ClickHouse | 661,255,693 | 12,509 | den-crane | Check CH logs at this moment in time.
The issue is unclear.
No proof that the reason is CH not a network or a firewall.
No steps to reproduce. | 2020-07-20T18:22:11 | 2020-07-20T18:22:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 661,595,707 | 12,583 | den-crane | @JTCunning I guess you insert several long text columns into a table with a small number of columns? | 2020-07-21T02:56:51 | 2020-07-21T02:56:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 660,324,265 | 6,312 | iameugenejo | I just had the same issue with `20.3.12.112` | 2020-07-17T20:31:20 | 2020-07-17T20:31:20 | {} | NONE |
ClickHouse | ClickHouse | 660,343,333 | 12,566 | alesapin | Traces from gdb when server hangs https://gist.github.com/alesapin/46a52d33d89f100c67cff3c4cff4bea8
https://gist.github.com/alesapin/df27abfa0db390a35cab43ecd66ee838
I don't understand why the server is not able to drop the view or underlying table. Nobody holds table's drop lock, new select queries also wait for t... | 2020-07-17T21:22:11 | 2020-07-17T21:23:38 | {} | MEMBER |
ClickHouse | ClickHouse | 660,534,125 | 12,570 | alexey-milovidov | When I kill process, the terminal remains hung and does not respond to any commands. | 2020-07-18T19:56:35 | 2020-07-18T19:56:35 | {} | MEMBER |
ClickHouse | ClickHouse | 660,564,811 | 11,167 | alexey-milovidov | > Buffer table is not getting cleared, it still has a count of 5 days data.
Could you please show the configuration of your Buffer table?
> Looks like the merge is ever ending . pls have a look at the attached logs.
merge_logs.txt
I see that the table `system.metric_log` is periodically merged in background. ... | 2020-07-19T00:48:42 | 2020-07-19T00:48:42 | {} | MEMBER |
ClickHouse | ClickHouse | 660,937,140 | 12,509 | toby080508050805 | Following is the scenario I encounter today. I cannot access into clickhouse via tcp port 9000 but i can access it by curl via 8123. There has no query in DATA BASE. I try apply telnet to access 9000 and it disconnect immediately.
` curl 'http://10.2.3.152:8123/?query=SELECT%20*%20from%20system.processes;'`
> 1 ... | 2020-07-20T10:11:48 | 2020-07-20T10:11:48 | {} | NONE |
ClickHouse | ClickHouse | 660,976,882 | 11,942 | bobelev | Maybe you can implement remote execution with the same user it was initially started?
You could introduce new setting like "remote_user". This would guarantee, that only authorized users can perform DDL, and other users can continue to use `default` user in cli.
With this approach we could manage resources when ... | 2020-07-20T11:42:36 | 2020-07-20T11:42:36 | {} | NONE |
ClickHouse | ClickHouse | 661,251,370 | 12,604 | 4ertus2 | Reproduced. `GROUP BY` eats column.
```
SELECT source.key, target.key FROM (SELECT 1 key, 'x' name) source
INNER JOIN (SELECT 2 key, 'x' name) target
ON source.name = target.name
ORDER BY source.key; -- OK
```
```
SELECT source.key, max(target.key) FROM (SELECT 1 key, 'x' name) source
INNER JOIN (SELECT 2 ke... | 2020-07-20T18:13:33 | 2020-07-20T18:16:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 661,320,168 | 12,571 | gkristic | Let me quantify the "I let it ran for minutes"... It was ~20 minutes. | 2020-07-20T20:37:03 | 2020-07-20T20:37:03 | {} | NONE |
ClickHouse | ClickHouse | 660,660,002 | 11,933 | keliss | Yes, parts, I guess: I got these numbers with your query. Sorry, got confused by the naming.
I didn't try increasing that yet, will let you know after I do. | 2020-07-19T15:08:33 | 2020-07-19T15:11:23 | {} | NONE |
ClickHouse | ClickHouse | 661,257,461 | 12,597 | 4ertus2 | The error means: ClickHouse expects to get data with type`UInt64` value for `select count(*) from something` but `mysql()` table function returns `String` instead. It could be an error in `mysql()` table function or some unexpected behaviour we could detect.
What result do you have in mysql console for `select count... | 2020-07-20T18:26:03 | 2020-07-20T18:26:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 661,417,236 | 12,596 | den-crane | What CH version do you use?
Can you provide table DDL? Do you have data sample?
| 2020-07-20T23:05:01 | 2020-07-20T23:05:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 661,581,387 | 12,596 | Johnhz | > What CH version do you use?
> Can you provide table DDL? Do you have data sample?
I can assist in recording the video and send it to you to see the execution process | 2020-07-21T02:36:13 | 2020-07-21T02:36:13 | {} | NONE |
ClickHouse | ClickHouse | 661,703,995 | 12,612 | alexey-milovidov | I don't see IA64 in the list of targets for LLVM: https://github.com/llvm/llvm-project/tree/master/llvm/lib/Target
But gcc is available (still). https://www.phoronix.com/scan.php?page=news_item&px=Intel-IA-64-GCC-Deprecation | 2020-07-21T08:07:00 | 2020-07-21T08:07:00 | {} | MEMBER |
ClickHouse | ClickHouse | 660,566,423 | 12,542 | doseeing | It seems `ColumnAggregateFunction` doesn't have `cloneResized` method.
Will it be implemented in future? | 2020-07-19T01:03:12 | 2020-07-19T01:03:12 | {} | NONE |
ClickHouse | ClickHouse | 661,029,511 | 12,601 | CurtizJ | I found a workaround:
```
SELECT
materialize(_shard_num) AS shard_num,
count()
FROM remote('127.0.0.{1,2}', system.one)
GROUP BY shard_num
┌─shard_num─┬─count()─┐
│ 2 │ 1 │
│ 1 │ 1 │
└───────────┴─────────┘
``` | 2020-07-20T13:10:23 | 2020-07-20T13:12:23 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 661,127,114 | 12,508 | qoega | https://clickhouse-test-reports.s3.yandex.net/0/21d3a797944066cb0b520e372570717ab165f3c4/unit_tests_msan_clang-10.html
zkutil.MultiAsync
```
[ RUN ] zkutil.MultiAsync
../src/Common/ZooKeeper/tests/gtest_zkutil_test_multi_exception.cpp:101: Failure
Expected equality of these values:
res.error
Which i... | 2020-07-20T15:55:56 | 2020-07-20T15:55:56 | {} | MEMBER |
ClickHouse | ClickHouse | 661,194,232 | 12,571 | den-crane | 10 mil rows * 2kb = raw data 20 GB, it will take 100GB RAM in a hash table
10 000
select q.id, p.id from (select * from xs limit 1) q cross join (select * from xs limit 10000) p where startsWith(q.name, p.name) format Null;
Peak memory usage (for query): 125.42 MiB.
100 000
select q.id, p.id from (select ... | 2020-07-20T17:03:16 | 2020-07-20T17:03:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 661,247,206 | 12,587 | KochetovNicolai | The approach with `ColumnFunction` is probably the faster to implement. | 2020-07-20T18:05:18 | 2020-07-20T18:05:18 | {} | MEMBER |
ClickHouse | ClickHouse | 661,421,745 | 12,586 | den-crane | Nested is just a syntax sugar for `CREATE TABLE`.
WA (provides the same result as Nested):
```
CREATE TABLE tab1 (
`col1.key` Array(String),
`col.value` Array(String),
col2 Array(String) MATERIALIZED `col1.key`)
ENGINE MergeTree() ORDER BY col1.key
```
| 2020-07-20T23:10:26 | 2020-07-20T23:10:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 661,426,945 | 12,559 | den-crane | CH version?
Table DDL ?
StackTrace ?
Logs? | 2020-07-20T23:16:41 | 2020-07-20T23:16:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 661,707,497 | 12,566 | alesapin | All tests passed, I have just renamed method. | 2020-07-21T08:14:06 | 2020-07-21T08:14:06 | {} | MEMBER |
ClickHouse | ClickHouse | 657,285,867 | 12,424 | alexey-milovidov | > AFAIU this is the intention, but looks tricky
I don't remember any intentions about that. | 2020-07-12T22:41:08 | 2020-07-12T22:41:08 | {} | MEMBER |
ClickHouse | ClickHouse | 657,434,262 | 12,453 | alexey-milovidov | Build check:
> Build failed without build log. It's inner CI problem. Maybe something wrong with docker or host. Just wait for build restart. | 2020-07-13T09:17:51 | 2020-07-13T09:17:51 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.