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 | 627,059,035 | 10,830 | den-crane | already implemented
--default_max_bytes_in_join arg Maximum size of right-side table if limit's required but max_bytes_in_join is not set.
```
--join_algorithm arg Specify join algorithm: 'auto', 'hash', 'partial_merge', 'prefer_par... | 2020-05-12T01:47:34 | 2020-05-12T01:54:00 | {
"-1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 627,091,883 | 10,076 | toby080508050805 | We found there has another query which have no sub query in where condition which can killed by kill query command but took 3000+ seconds.
select *
from source_table
where
WHERE DATA_DATE = '2020-05-09'
AND DATA_HOUR = 12
-- AND PU_ID IN (149,142,120,152,156,148,155,145,144,119,141,154,150,118,146,151,116,1... | 2020-05-12T03:50:14 | 2020-05-12T04:04:39 | {} | NONE |
ClickHouse | ClickHouse | 627,569,966 | 10,405 | 4ertus2 | There's the same error on another installation in case of `min_bytes_to_use_direct_io=1` and `merge_tree_uniform_read_distribution=0`.
```
Code: 271. DB::Exception: Received from localhost:9000. DB::Exception: Data compressed with different methods, given method byte ff, previous method byte 82: (while reading column... | 2020-05-12T20:16:16 | 2020-05-12T20:16:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 627,003,241 | 10,818 | genzgd | Is this issue present on the current 2.3 release with experimental_processors turned off? Is there a current workaround? | 2020-05-11T22:39:21 | 2020-05-11T22:39:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 627,242,565 | 10,841 | alesapin | I don't know how to create a test. | 2020-05-12T10:02:56 | 2020-05-12T10:02:56 | {} | MEMBER |
ClickHouse | ClickHouse | 627,290,992 | 10,050 | KochetovNicolai | The way how it may happen (as I have understood it):
0. There are several replicas, one replica is hung (e.g. have problems with zookeeper). `ON CLUSTER` query is sent.
1. All replicas except one execute `ON CLUSTER` query. `cleanup_delay_period` passes (1 minute), and several other `ON CLUSTER` queries are created t... | 2020-05-12T11:41:48 | 2020-05-12T11:41:48 | {} | MEMBER |
ClickHouse | ClickHouse | 627,654,403 | 10,804 | alexey-milovidov | The issue is clearly visible here: https://clickhouse-test-reports.s3.yandex.net/0/6053f012b84a4174b95c2b021ece40f9a59ad99c/functional_stateless_tests_(memory).html | 2020-05-12T23:30:16 | 2020-05-12T23:30:16 | {} | MEMBER |
ClickHouse | ClickHouse | 627,043,946 | 10,826 | azat | >you can sort iterators
Done | 2020-05-12T00:54:03 | 2020-05-12T00:54:03 | {} | MEMBER |
ClickHouse | ClickHouse | 627,055,091 | 10,797 | alexey-milovidov | It's only possible to transfer less amount of data from remote servers if we can assume that data is sharded by aggregation key, thus data for every key can be aggregated only on a single server.
FYI @azat | 2020-05-12T01:32:23 | 2020-05-12T01:32:23 | {} | MEMBER |
ClickHouse | ClickHouse | 627,113,575 | 10,797 | alexey-milovidov | Yes, it is possible to return approximate result with these settings:
`max_rows_to_group_by = 1000000` and `group_by_overflow_mode = 'any'` | 2020-05-12T05:19:02 | 2020-05-12T05:19:02 | {} | MEMBER |
ClickHouse | ClickHouse | 627,208,792 | 10,770 | azat | >this can be archived by introducing separate min_insert_block_size_rows/min_insert_block_size_bytes for pushing to materialized views (and setting them to 1 will drop extra memory usage, since this will just proxy the data)
Or by adding something like `insert_materialized_view_no_squash`, but this is less flexible | 2020-05-12T08:56:57 | 2020-05-12T08:56:57 | {} | MEMBER |
ClickHouse | ClickHouse | 627,349,402 | 10,578 | KochetovNicolai | Also 01246_insert_into_watch_live_view was failed in debug: https://clickhouse-test-reports.s3.yandex.net/10578/0618c806b77bc0682d1c43066d6d6a9c93d7faed/functional_stateless_tests_(debug)/test_run.txt.out.log | 2020-05-12T13:35:02 | 2020-05-12T13:35:02 | {} | MEMBER |
ClickHouse | ClickHouse | 627,458,536 | 10,845 | shizhenchina | 1.Sorry, domain is a later column that I forgot to write down.
2. I will send the datasample/insert queries later | 2020-05-12T16:40:04 | 2020-05-12T16:40:04 | {} | NONE |
ClickHouse | ClickHouse | 627,536,189 | 10,852 | nvartolomei | A test would be useful, see https://github.com/ClickHouse/ClickHouse/blob/97f2a2213e754ba25dabba4bc8ddf507cd67660c/tests/queries/0_stateless/00948_format_in_with_single_element.sh together with https://github.com/ClickHouse/ClickHouse/blob/97f2a2213e754ba25dabba4bc8ddf507cd67660c/tests/queries/0_stateless/00948_format_... | 2020-05-12T19:09:06 | 2020-05-12T19:09:06 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 627,582,306 | 10,158 | genzgd | We are getting this error as well, and I'm not seeing how it's related to #9940 | 2020-05-12T20:41:44 | 2020-05-12T20:41:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 626,888,293 | 10,797 | alexey-milovidov | `max_result_rows` is about final query result, not intermediate parts. | 2020-05-11T18:49:52 | 2020-05-11T18:49:52 | {} | MEMBER |
ClickHouse | ClickHouse | 626,912,677 | 4,190 | alexey-milovidov | We will try to find someone from community who are interested to implement this feature. | 2020-05-11T19:34:16 | 2020-05-11T19:34:16 | {} | MEMBER |
ClickHouse | ClickHouse | 627,016,220 | 157 | alexey-milovidov | The library for User-Agent parsing was published in #10822
But the task is in progress (now you can see just the exported sources of the library and it cannot be used). | 2020-05-11T23:19:27 | 2020-05-11T23:22:15 | {} | MEMBER |
ClickHouse | ClickHouse | 627,037,003 | 10,791 | CurtizJ | > It might be worth investigating why it was working for `>` and not working for `<` condition
It was, because in case of `>` `KeyCondition` decides to read granule, because it thinks that its any value is greater than date value and then data is filtered properly, but in case of `<` granule was skipped at all. | 2020-05-12T00:28:09 | 2020-05-12T00:32:45 | {} | MEMBER |
ClickHouse | ClickHouse | 627,040,518 | 10,825 | alexey-milovidov | No way. Better to submit PR even if it is in experimental stage or PoC (just make it a draft). | 2020-05-12T00:41:37 | 2020-05-12T00:41:37 | {} | MEMBER |
ClickHouse | ClickHouse | 627,059,572 | 10,830 | jakesylvestre | ah okay @den-crane, from my reading of #7186 it looked like it might've been. I just wasn't sure if my test was just using a silent throw mode. Thanks for the help | 2020-05-12T01:49:38 | 2020-05-12T01:49:38 | {} | NONE |
ClickHouse | ClickHouse | 627,195,741 | 10,593 | KochetovNicolai | #10603 | 2020-05-12T08:31:17 | 2020-05-12T08:31:17 | {} | MEMBER |
ClickHouse | ClickHouse | 627,212,943 | 10,460 | KochetovNicolai | I've lost this issue.
How to reproduce it? | 2020-05-12T09:04:49 | 2020-05-12T09:04:49 | {} | MEMBER |
ClickHouse | ClickHouse | 627,215,652 | 10,829 | qoega | Unwanted diff in tests
``` sql
SELECT varSamp(ResolutionWidth) FROM (SELECT ResolutionWidth FROM test.hits LIMIT 1);
```
0 instead of NaN
PS. We can move test scenarios to run over system.numbers(_mt) in stateless tests. | 2020-05-12T09:10:10 | 2020-05-12T09:10:10 | {} | MEMBER |
ClickHouse | ClickHouse | 627,218,133 | 10,834 | nikitamikhaylov | > The code of cache dictionary is very complicated and has to be thrown away.
I was thinking about simplifying cache-dictionary just yesterday evening.
So, if we will use naive LRU with hash-table and double linked list, then when we discover that some key is too old we can delete all nodes righter this one (or req... | 2020-05-12T09:15:05 | 2020-05-12T09:15:05 | {} | MEMBER |
ClickHouse | ClickHouse | 627,244,579 | 8,159 | yehiaelbehery | If you're using docker add to your docker run commnad:
-e "LIBHDFS3_CONF=/etc/clickhouse-server/hdfs-client.xml"
tested on yandex/clickhouse-server:20.3.8.53 image | 2020-05-12T10:06:38 | 2020-05-12T10:06:38 | {} | NONE |
ClickHouse | ClickHouse | 627,537,160 | 10,383 | akuzm | Given the above discussion, I am inclined to think that the finalized aggregation function state is sufficiently distinct to warrant having a separate data type for it. Also, `finalizeAggregation(state)` should be rewritten by the planner to something like `getResult(finalize(state))`, which would enable proper CSE tha... | 2020-05-12T19:11:12 | 2020-05-12T19:11:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 626,893,546 | 10,818 | azat | @alexey-milovidov Yes, will take a look | 2020-05-11T18:58:53 | 2020-05-11T18:58:53 | {} | MEMBER |
ClickHouse | ClickHouse | 626,897,164 | 10,813 | alexey-milovidov | It's an issue of Poco library, we have to patch it:
https://clickhouse.tech/codebrowser/html_report/ClickHouse/contrib/poco/Net/src/DNS.cpp.html#392
And it's rather easy to do.
It should use the `gai_strerror` function.
| 2020-05-11T19:05:12 | 2020-05-11T19:05:12 | {} | MEMBER |
ClickHouse | ClickHouse | 626,898,421 | 10,486 | alexey-milovidov | Unfortunately, data from this task has gone. | 2020-05-11T19:07:35 | 2020-05-11T19:07:35 | {} | MEMBER |
ClickHouse | ClickHouse | 626,990,309 | 10,775 | alexey-milovidov | Ok, I've disabled this test for Unbundled build. | 2020-05-11T22:04:51 | 2020-05-11T22:04:51 | {} | MEMBER |
ClickHouse | ClickHouse | 627,049,694 | 10,797 | weiyongyuan | 1.I'm sorry,i paste a inapprooriate test result.Here is the bew test result.
I start the distribute query.
Here is the distribute query result:
.
You can change it with the `wait_end_of_query` URL parameter, see #614 | 2020-05-11T22:08:53 | 2020-05-11T22:08:53 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 627,033,051 | 10,818 | alexey-milovidov | @azat It will slowdown group by approx 10..20 times. | 2020-05-12T00:14:47 | 2020-05-12T00:14:47 | {} | MEMBER |
ClickHouse | ClickHouse | 627,193,534 | 10,797 | azat | >2020.05.12 16:09:24.537392 [ 113 ] {} BaseDaemon: Address: 0x10 Access: read. Address not mapped to object.
Can you rerun the same with debug symbols installed? (`apt install clickhouse-common-static-dbg` and restart `clickhouse-server`)
And provide the structure of the distributed and local tables?
P.S. consid... | 2020-05-12T08:27:08 | 2020-05-12T08:27:08 | {} | MEMBER |
ClickHouse | ClickHouse | 627,540,846 | 10,853 | azat | So `sipHash64(cast('foo', 'LowCardinality(String)'))`returns `LowCardinality(UInt64)` and `optimize_skip_unused_shards` for `LowCardinality` has been fixed in #10611 | 2020-05-12T19:18:14 | 2020-05-12T19:19:00 | {
"+1": 3
} | MEMBER |
ClickHouse | ClickHouse | 626,910,358 | 5,681 | alexey-milovidov | > Just thoughts, looks like it can be possible to implement auto increment with the same manner as MySQL does
Yes, it's totally possible.
The only drawback is that even on single replica INSERTs may go out of order (first data is written to filesystem, then it is committed and multiple INSERTs can process in para... | 2020-05-11T19:29:46 | 2020-05-11T19:29:46 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 626,921,204 | 8,840 | alexey-milovidov | They did not. | 2020-05-11T19:50:49 | 2020-05-11T19:50:49 | {} | MEMBER |
ClickHouse | ClickHouse | 626,987,976 | 10,486 | alexey-milovidov | Ok, I will disable this test for unbundled build.
> and myabe you even tried to do this
I don't remember. | 2020-05-11T21:59:06 | 2020-05-11T21:59:06 | {} | MEMBER |
ClickHouse | ClickHouse | 627,032,580 | 10,828 | alexey-milovidov | No test will be provided for this change. | 2020-05-12T00:13:23 | 2020-05-12T00:13:23 | {} | MEMBER |
ClickHouse | ClickHouse | 627,050,005 | 10,797 | weiyongyuan | > `max_result_rows` is about final query result, not intermediate parts.
I paste my new test result.limit and max_result_rows all don't work on local node! | 2020-05-12T01:14:16 | 2020-05-12T01:32:18 | {} | NONE |
ClickHouse | ClickHouse | 627,206,107 | 10,797 | weiyongyuan | > > 2020.05.12 16:09:24.537392 [ 113 ] {} BaseDaemon: Address: 0x10 Access: read. Address not mapped to object.
>
> Can you rerun the same with debug symbols installed? (`apt install clickhouse-common-static-dbg` and restart `clickhouse-server`)
> And provide the structure of the distributed and local tables?
>
... | 2020-05-12T08:51:49 | 2020-05-12T09:01:16 | {} | NONE |
ClickHouse | ClickHouse | 627,289,500 | 10,831 | nikitamikhaylov | You have to leave only one category and write some description about this PR. | 2020-05-12T11:38:12 | 2020-05-12T11:38:12 | {} | MEMBER |
ClickHouse | ClickHouse | 627,633,235 | 10,621 | azat | Check failures looks like due to problems on CI | 2020-05-12T22:30:15 | 2020-05-12T22:30:15 | {} | MEMBER |
ClickHouse | ClickHouse | 626,920,631 | 10,486 | alexey-milovidov | It's very interesting, how it was possible...
I will try to find similar failures for recent commits. | 2020-05-11T19:49:37 | 2020-05-11T19:49:37 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 627,031,504 | 10,818 | azat | >Is there a current workaround?
@genzgd workaround will be to disable two-level aggregation, with:
- `group_by_two_level_threshold=0`
- `group_by_two_level_threshold_bytes=0`
*(but note, that this means that external GROUP BY will not work, i.e. `max_bytes_before_external_group_by`)* | 2020-05-12T00:10:06 | 2020-05-12T00:11:30 | {} | MEMBER |
ClickHouse | ClickHouse | 627,065,683 | 10,434 | alexey-milovidov | It was made with an error... try №2. | 2020-05-12T02:12:28 | 2020-05-12T02:12:28 | {} | MEMBER |
ClickHouse | ClickHouse | 627,072,190 | 10,806 | toby080508050805 | I found other parameter to adjust max block size and min free thread. It is work now.
https://github.com/ClickHouse/ClickHouse/issues/5500
https://github.com/ClickHouse/ClickHouse/issues/2607 | 2020-05-12T02:35:31 | 2020-05-12T02:35:31 | {} | NONE |
ClickHouse | ClickHouse | 627,081,947 | 10,802 | YiuRULE | Seem review ready now, some check were failed, but I cannot check the `Yandex synchronization check` and other tasks doesn't seem to be related to this PR | 2020-05-12T03:12:39 | 2020-05-12T03:12:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 627,188,479 | 10,797 | weiyongyuan | > Yes, it is possible to return approximate result with these settings:
> `max_rows_to_group_by = 1000000` and `group_by_overflow_mode = 'any'`
alexey,thank you a lot! I've test the tow paramters in two enviroments,but they does'nt work in version 20.1.6.30 and I encounter new problem!By the way,the distribute quer... | 2020-05-12T08:16:57 | 2020-05-12T08:16:57 | {} | NONE |
ClickHouse | ClickHouse | 627,207,519 | 10,770 | azat | >(by some reason it uses a little bit more memory for now, but this is a separate issue)
That was because some engines bufffer the write | 2020-05-12T08:54:34 | 2020-05-12T08:54:34 | {} | MEMBER |
ClickHouse | ClickHouse | 627,211,737 | 10,797 | azat | >do you mean this:set send_logs_level='trace';?
No, I mean the following:
**But note that this will restart the clickhouse-server and maybe this is not suitable for you right now** (but since your server `SIGSEGV`'ed I guess it is ok)
- `apt-get install clickhouse-common-static-dbg=$(dpkg-query -W -f '${Versio... | 2020-05-12T09:02:33 | 2020-05-12T09:03:10 | {} | MEMBER |
ClickHouse | ClickHouse | 627,248,199 | 10,833 | qoega | unrelated fail 01076_parallel_alter_replicated_zookeeper | 2020-05-12T10:13:38 | 2020-05-12T10:13:38 | {} | MEMBER |
ClickHouse | ClickHouse | 627,335,591 | 10,846 | den-crane | WA: SELECT if(materialize(isFinite(toUInt64OrZero(toNullable('123')))), 1, 0) | 2020-05-12T13:13:07 | 2020-05-12T13:13:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 627,410,331 | 8,467 | qoega | Probably off topic:
Have you thought about more complex way to deduplicate inserts rather then data only hash? We can use some client `insertID` to deduplicate if client is able to set it on insert. And there can be several strategies
- `data_hash` - your current way to deduplicate. Good for data with high entropy
... | 2020-05-12T15:16:49 | 2020-05-12T15:16:49 | {} | MEMBER |
ClickHouse | ClickHouse | 627,540,653 | 10,856 | alexey-milovidov | Also need to add `REAL`.
https://www.postgresql.org/docs/9.1/datatype-numeric.html | 2020-05-12T19:17:52 | 2020-05-12T19:17:52 | {} | MEMBER |
ClickHouse | ClickHouse | 627,547,949 | 10,726 | alexey-milovidov | I've created another task: https://github.com/ClickHouse/ClickHouse/issues/10862 | 2020-05-12T19:32:44 | 2020-05-12T19:32:44 | {} | MEMBER |
ClickHouse | ClickHouse | 627,568,559 | 4,190 | manish53 | And https://github.com/ClickHouse/ClickHouse/pull/5388 | 2020-05-12T20:13:21 | 2020-05-12T20:13:21 | {} | NONE |
ClickHouse | ClickHouse | 627,173,176 | 6,936 | LizardWizzard |
Sorry, my statement was not complete. Protobuf can omit default
values where it is possible to unambiguously infer them from schema
without introducing situations like one you pointed to.
May be this explanation turned out to be too verbose and contains extra info but for the sake of completeness let it all be ... | 2020-05-12T07:45:37 | 2020-05-12T07:45:37 | {} | NONE |
ClickHouse | ClickHouse | 627,292,642 | 10,832 | akuzm | > Cool and quite old feature. Why we didn't use it before?
Personally, I used thread number from `received signal from thread ...` to find the failing query, but this is more convenient.
Something is broken though:
```
Functional stateless tests (memory) — Some queries hung,
``` | 2020-05-12T11:45:38 | 2020-05-12T11:45:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 627,035,009 | 10,826 | azat | >Was broken here: #9950
Didn't even though that there was std::map (hence didn't even looked into the `git blame`)
>Note that filling std::vector and calling std::sort is more efficient than creating std::map (both time and memory). And also you can sort iterators or even value_type pointers.
Indeed, nice catc... | 2020-05-12T00:21:15 | 2020-05-12T00:21:15 | {} | MEMBER |
ClickHouse | ClickHouse | 627,065,381 | 10,797 | weiyongyuan | > It's only possible to transfer less amount of data from remote servers if we can assume that data is sharded by aggregation key, thus data for every key can be aggregated only on a single server.
>
> FYI @azat
My data is randomly distribute on every server,the same key may be on every server.so,if i have a big cl... | 2020-05-12T02:11:26 | 2020-05-12T02:11:26 | {} | NONE |
ClickHouse | ClickHouse | 627,420,090 | 10,747 | alesapin | Cannot reproduce, this test works if I comment this line. | 2020-05-12T15:33:04 | 2020-05-12T15:33:04 | {} | MEMBER |
ClickHouse | ClickHouse | 627,455,792 | 7,612 | manojfim | Hello @alexey-milovidov,
Thanks for the information. I have got the insert command working for single file. Is there a way to insert or select all the files in the s3 directory using wildcard?
Thanks | 2020-05-12T16:35:08 | 2020-05-12T16:35:08 | {} | NONE |
ClickHouse | ClickHouse | 627,023,699 | 10,825 | alexey-milovidov | Note: it's already possible to run almost the same jobs on local machine.
Just look at the logs (they contain the info what docker images were run with what arguments) and run exactly the same commands. | 2020-05-11T23:43:33 | 2020-05-11T23:43:33 | {} | MEMBER |
ClickHouse | ClickHouse | 627,059,926 | 10,815 | hackertw | We have tried to replace the hardware server and still have the same problem.
steps:
1.Create table with ReplicatedMerge table engine.
2.Insert 27 million data
3.After executing the check table query statement, it returns 0.
4.After executing select count, it was found that the data was reduced to 3 million, and t... | 2020-05-12T01:51:03 | 2020-05-12T01:51:03 | {} | NONE |
ClickHouse | ClickHouse | 627,124,147 | 10,834 | alexey-milovidov | The code of cache dictionary is very complicated and has to be thrown away.
Most likely the bug is still there in String case.
To reprocude, run `01054_cache_dictionary_bunch_update` test.
The issue is reproducing on every commit in master. | 2020-05-12T05:47:41 | 2020-05-12T05:47:41 | {} | MEMBER |
ClickHouse | ClickHouse | 627,270,574 | 8,976 | blinkov | @sundarv85, do you have any further questions? | 2020-05-12T11:00:50 | 2020-05-12T11:00:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 627,281,009 | 10,797 | weiyongyuan | > > do you mean this:set send_logs_level='trace';?
>
> No, I mean the following:
>
> **But note that this will restart the clickhouse-server and maybe this is not suitable for you right now** (but since your server `SIGSEGV`'ed I guess it is ok)
>
> * `apt-get install clickhouse-common-static-dbg=$(dpkg-query ... | 2020-05-12T11:24:12 | 2020-05-12T11:24:12 | {} | NONE |
ClickHouse | ClickHouse | 627,567,174 | 4,190 | manish53 | Hi alexey,
Sometime ago i faced same issue and successfully fixed it by updating poco version in clickhouse.
You can see PR https://github.com/ClickHouse/ClickHouse/pull/5387 | 2020-05-12T20:10:28 | 2020-05-12T20:10:28 | {} | NONE |
ClickHouse | ClickHouse | 627,340,591 | 10,383 | akuzm | Some other broken aggregate functions that modify data in insertResultInto:
https://github.com/ClickHouse/ClickHouse/blob/master/src/AggregateFunctions/AggregateFunctionMaxIntersections.h#L138
https://github.com/ClickHouse/ClickHouse/blob/master/src/AggregateFunctions/AggregateFunctionQuantile.h#L143
https://g... | 2020-05-12T13:20:48 | 2020-05-12T13:20:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 627,375,229 | 10,845 | den-crane | 1. I don't see `domain` column in your tables.
2. Need datasample/insert queries to reproduce
| 2020-05-12T14:18:47 | 2020-05-12T14:18:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 627,400,241 | 10,833 | alexey-milovidov | > unrelated fail 01076_parallel_alter_replicated_zookeeper
I've added it to CI fest. | 2020-05-12T15:00:43 | 2020-05-12T15:00:43 | {} | MEMBER |
ClickHouse | ClickHouse | 627,492,584 | 10,789 | inakivb | Solved!
Thank you @den-crane ! | 2020-05-12T17:44:09 | 2020-05-12T17:44:09 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 627,540,732 | 10,856 | alexey-milovidov | Please provide some tests... | 2020-05-12T19:18:01 | 2020-05-12T19:18:01 | {} | MEMBER |
ClickHouse | ClickHouse | 627,034,309 | 10,825 | azat | Any chance that it will be possible to run the tests on **our internal yandex CI** but **without PR**?
Let's say I want to run them on some branch in my fork (or/and some private repository involved, i.e. checks of some internal libraries or similar), but don't want to pollute PRs list | 2020-05-12T00:18:41 | 2020-05-12T00:41:07 | {} | MEMBER |
ClickHouse | ClickHouse | 627,100,888 | 10,725 | wenwen2019 | Can the dictionary be updated only after restarting clickhouse？ | 2020-05-12T04:28:54 | 2020-05-12T04:28:54 | {} | NONE |
ClickHouse | ClickHouse | 627,267,512 | 10,726 | hhell | > The issue is not about HTTP protocol but about text format (TSV, CSV, etc).
Yes, but this is still an issue. | 2020-05-12T10:53:37 | 2020-05-12T10:53:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 627,285,626 | 10,847 | nikitamikhaylov | after -> alter ?
I see: after ALTER RENAME COLUMN | 2020-05-12T11:33:50 | 2020-05-12T11:35:45 | {} | MEMBER |
ClickHouse | ClickHouse | 627,347,884 | 10,771 | BayoNet | Thanks for your participation. I'll add it to the https://clickhouse.tech/docs/en/engines/table-engines/mergetree-family/replication/ page because the order of the parameters is a common rule for all the Replicated*MergeTree engines. | 2020-05-12T13:32:38 | 2020-05-12T13:32:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 627,492,664 | 10,797 | azat | >do i need to uninstall the clickhouse-common-static-20.1.6.30-2.x86_64 when i've intsalled the clickhouse-common-static-dbg-20.1.6.30-2.x86_64?
Not necessarily (this wont hurt)
>local server.log
Thanks!
Is this development env? If so, maybe you can try latest testing - v20.5.1.3327-testing (there were some f... | 2020-05-12T17:44:19 | 2020-05-12T17:44:19 | {} | MEMBER |
ClickHouse | ClickHouse | 626,973,948 | 6,936 | LizardWizzard | If I understand Protobuf logic correctly default value can be fetched from entity schema, and because of that protocol doesnt have to supply known default value in transmitted data.
| 2020-05-11T21:25:56 | 2020-05-11T21:25:56 | {} | NONE |
ClickHouse | ClickHouse | 626,986,138 | 10,486 | azat | >It's very interesting, how it was possible...
Actually it is not, in a nut shell the problem is in racy check in `Poco::File::createDirectory` in the upstream poco, here is details:
- that build is unbundled so it uses [poco 1.9.2-3ubuntu2](https://packages.ubuntu.com/eoan/libpoco-dev)
- that version has the foll... | 2020-05-11T21:54:28 | 2020-05-11T21:54:28 | {} | MEMBER |
ClickHouse | ClickHouse | 627,250,256 | 10,142 | dmitry3325 | Hi, maybe this is different problems. In my case query
`SELECT campaign_id,stream_id,sub_stream_id
FROM `hits`
GROUP BY campaign_id, stream_id, sub_stream_id
HAVING SUM(event_date >= toDate('2020-05-05') ? toFloat32(debit) : 0) > 0`
Work in version 19.17.5 but didn't work in last one. Here debit is decimal. With ... | 2020-05-12T10:17:36 | 2020-05-12T10:23:12 | {} | NONE |
ClickHouse | ClickHouse | 627,353,104 | 10,818 | genzgd | Thanks for the details. | 2020-05-12T13:41:04 | 2020-05-12T13:41:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 627,406,191 | 10,852 | qoega | Looks OK. Lets look on build results. | 2020-05-12T15:10:04 | 2020-05-12T15:10:04 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 627,456,778 | 10,458 | alexey-milovidov | Why close? | 2020-05-12T16:36:55 | 2020-05-12T16:36:55 | {} | MEMBER |
ClickHouse | ClickHouse | 627,539,520 | 10,856 | alexey-milovidov | Ok, but we have to remove at least INT8 because it will be ambiguous with Int8.
And also please check whether these types
```
"INT1", "Int8"
"INT2", "Int16"
"INT4", "Int32"
```
are standard or not. | 2020-05-12T19:15:47 | 2020-05-12T19:15:47 | {} | MEMBER |
ClickHouse | ClickHouse | 626,942,980 | 9,001 | alexey-milovidov | It can be implemented for Log and TinyLog in trivial way.
And also for StripeLog - because it has an index to seek for specific column. | 2020-05-11T20:28:47 | 2020-05-11T20:28:47 | {} | MEMBER |
ClickHouse | ClickHouse | 627,014,793 | 10,818 | alexey-milovidov | This is expected to present in old releases and to be independent of processors. | 2020-05-11T23:14:51 | 2020-05-11T23:14:51 | {} | MEMBER |
ClickHouse | ClickHouse | 627,034,313 | 10,826 | alexey-milovidov | Was broken here: https://github.com/ClickHouse/ClickHouse/pull/9950 | 2020-05-12T00:18:42 | 2020-05-12T00:18:42 | {} | MEMBER |
ClickHouse | ClickHouse | 627,408,583 | 10,383 | KochetovNicolai | Cool! I will add changes to other functions. | 2020-05-12T15:13:56 | 2020-05-12T15:13:56 | {} | MEMBER |
ClickHouse | ClickHouse | 627,457,567 | 7,612 | den-crane | @manojfim
Yes,
`select * from s3('http://host/bucket/*.csv/...` | 2020-05-12T16:38:20 | 2020-05-12T16:38:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 614,898,212 | 332 | azat | >If we do GROUP BY injective function of sharding key
Although brief look at `isInjective` shows that it is not set correctly for some useful stuff (`module`, consistent hashing), so this also should not be forgotten. | 2020-04-16T21:10:49 | 2020-04-16T21:11:00 | {} | MEMBER |
ClickHouse | ClickHouse | 614,910,572 | 3,889 | alexey-milovidov | No, it won't work for NaNs as expected.
```
SELECT
reinterpretAsFloat64(unhex('000000000000F87F')) AS x,
reinterpretAsFloat64(unhex('100000000000F87F')) AS y
```
These two different nans should not be bitEqual. | 2020-04-16T21:39:16 | 2020-04-16T21:39:16 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.