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 | 346,226,016 | 1,516 | alexey-milovidov | The code is totally and obviously wrong:
In each thread of execution you call `readPrefix` for each input.
In total you do num inputs * num threads calls.
`readPrefix` cannot be called in parallel,
same as `read` method.
This is stated in comments:
```
/** Read next block.
* If there are no more b... | 2017-11-22T02:38:59 | 2017-11-22T02:38:59 | {} | MEMBER |
ClickHouse | ClickHouse | 346,222,887 | 1,516 | alexey-milovidov | Functional tests under ASan fail. | 2017-11-22T02:18:02 | 2017-11-22T02:43:41 | {} | MEMBER |
ClickHouse | ClickHouse | 346,516,361 | 1,533 | alexey-milovidov | .+ arraySort, arrayReverseSort, arrayReduce. | 2017-11-23T01:49:59 | 2017-11-23T01:50:23 | {} | MEMBER |
ClickHouse | ClickHouse | 347,502,982 | 1,541 | ztlpn | @alexey-milovidov good idea, will do.
@silviucpp yes, it is another bug. Still, it is a good thing that you have reviewed this patch :) | 2017-11-28T12:05:07 | 2017-11-28T12:05:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 347,722,116 | 1,554 | LiuYangkuan | my source_tab is Kafka engine table | 2017-11-29T01:29:00 | 2017-11-29T01:29:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 347,828,258 | 1,513 | M1ha-Shvn | Thanks, I'll have a look later. | 2017-11-29T11:10:05 | 2017-11-29T11:10:05 | {} | NONE |
ClickHouse | ClickHouse | 347,873,343 | 1,281 | silviucpp | I tried to fix it few days ago but I lost myself in the debugging process. Will try my luck again when I have some more time in case nobody from yandex will handle until that time. | 2017-11-29T14:18:29 | 2017-11-29T14:18:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 347,990,894 | 538 | pondix | I seem to be having a similar issue, here are the details:
```
- APPLICATION LOG (Python subprocess loading CSV file via stdin to clickhouse-client):
ERROR:root:2017-11-30 04:38:01: Failed loading file /data/mysql-files/adscore.history_details-20171130-043301-3210999986.csv into Clickhouse table test.historyplay... | 2017-11-29T20:46:30 | 2017-11-29T20:46:30 | {} | NONE |
ClickHouse | ClickHouse | 348,464,314 | 393 | gudmojo | @alexey-milovidov thank you. I'm getting this again on a version that has the fix. Using version 1.1.54304 and getting this in my application:
Nov 30 14:15:27 ALTER table `mydatabase`.`mytable_1day` FREEZE PARTITION '201711'
Nov 30 14:15:37 Unknown error field: Poco::Exception. Code: 1000
Nov 30 14:15:37 Unknown e... | 2017-12-01T10:50:33 | 2017-12-01T10:59:59 | {} | NONE |
ClickHouse | ClickHouse | 345,818,223 | 1,519 | alexey-milovidov | :+1: | 2017-11-20T20:16:51 | 2017-11-20T20:16:51 | {} | MEMBER |
ClickHouse | ClickHouse | 345,875,509 | 1,466 | alexey-milovidov | Now it's clear that the bug is on our side. | 2017-11-21T00:22:47 | 2017-11-21T00:22:47 | {} | MEMBER |
ClickHouse | ClickHouse | 346,028,955 | 1,514 | theseusyang | @alexey-milovidov Yes, we want to do some graph query and graph analysis like Titan or Neo4j, and also especially like enterprise product [GraphSQL](http://www.graphsql.com/) , a table with 2 paradigms , graph and relational database. | 2017-11-21T13:39:00 | 2017-11-21T13:39:00 | {} | NONE |
ClickHouse | ClickHouse | 346,212,214 | 1,466 | code-of-kpp | Build errors with this commit.
```
[ 16%] Building CXX object dbms/CMakeFiles/dbms_common_io.dir/src/Common/BackgroundSchedulePool.cpp.o
In file included from /tmp/build/github.com/yandex/ClickHouse/dbms/src/Common/BackgroundSchedulePool.cpp:1:0:
/tmp/build/github.com/yandex/ClickHouse/dbms/src/Common/BackgroundS... | 2017-11-22T01:11:45 | 2017-11-22T01:11:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 346,400,417 | 1,536 | filimonov | @theseusyang at that moment the only caching which Clickhouse use - is caching and memory-mapping of disk data. So if you have a lot of memory (the best: all data fits in memory) you will have practically no disk reads, only one at the beginning. But results of operation on that data are not cached, so if you will do s... | 2017-11-22T16:21:20 | 2017-11-22T17:28:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 346,786,555 | 1,536 | valyala | As I know, [max_concurrent_queries](https://clickhouse.yandex/docs/ru/operations/server_settings/settings.html#max-concurrent-queries) limits the number of concurrent queries only on a single `ClickHouse` server. It doesn't limit concurrent requests to distributed tables on a cluster.
@theseusyang , take a look at [... | 2017-11-24T09:46:23 | 2017-11-24T09:46:23 | {} | NONE |
ClickHouse | ClickHouse | 347,499,007 | 1,552 | alex-zaitsev | I opened a similar issue long ago, no progress yet: https://github.com/yandex/ClickHouse/issues/863 | 2017-11-28T11:46:19 | 2017-11-28T11:46:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 348,284,414 | 1,568 | filimonov | Clickhouse is best in building reports by grouping and filtering data. If your data is 5-15Mb strings - it's quite hard to imagine scenarios how you can group / filter that huge strings. Probably you need some database with full text index.
BUT: if it's possible to extract from that huge log strings some properties ... | 2017-11-30T18:52:49 | 2017-11-30T18:53:53 | {
"+1": 3
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 346,039,468 | 1,524 | silviucpp | Random error sometimes returns:
```
2017.11.21 14:15:38.940300 [ 8 ] <Error> HTTPHandler: Code: 102, e.displayText() = DB::NetException: Unexpected packet from server ck3.office:9000, 172.17.33.125 (expected TablesStatusResponse, got Hello), e.what() = DB::NetException, Stack trace:
0. clickhouse-server(StackTra... | 2017-11-21T14:17:04 | 2017-11-21T14:17:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 346,372,358 | 1,536 | theseusyang | @filimonov Your suggestion is good, especially Best Practice. but actually it is not very relevant between Key-Value and SQL at High-Concurrency Request. Because ClickHouse has memory table, If understand correctly, has also Cache Mechanism, then ClickHouse can cache an entry in this table like Redis does , to respond ... | 2017-11-22T14:51:47 | 2017-11-22T14:51:47 | {} | NONE |
ClickHouse | ClickHouse | 347,027,502 | 1,443 | wawanawna | I see you are using IP addresses in configs. Try to use DNS names instead (from /etc/hosts), it helped me solved similar issue. | 2017-11-26T18:14:17 | 2017-11-26T18:14:17 | {} | NONE |
ClickHouse | ClickHouse | 347,168,510 | 294 | jason-heo | Hi,
I'm wondering `approximate Top-K` has already been implemented.
Event though it looks like it merged, I couldn't find its functionality on the ClickHouse documentation.
Thanks. | 2017-11-27T12:32:10 | 2017-11-27T12:32:10 | {} | NONE |
ClickHouse | ClickHouse | 347,289,452 | 1,489 | alexey-milovidov | I am interested, what are the results of performance testing on your use cases.
I see the following advantages of the modification:
- gdb was starting too slow when using large number of Replicated tables;
- output of `t apply all bt` from gdb became more readable;
- sometimes the user had to increase ulimit; | 2017-11-27T19:02:54 | 2017-11-27T19:02:54 | {} | MEMBER |
ClickHouse | ClickHouse | 348,387,415 | 1,504 | vavrusa | This has happened again, coincidentally INSERT queries started consuming a lot of memory for no apparent reason exactly at 2017-12-01 00:00:00, that's probably unrelated. | 2017-12-01T02:54:44 | 2017-12-01T02:54:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 346,011,282 | 1,510 | goranc | For the situations like this, we need to have option to rollback upgrade of server version.
In our situation we have previous .deb packages and can install older version, but it should be more convenient to have previous stable version of packages in yandex repository.
So can you keep two latest stable versions i... | 2017-11-21T12:25:19 | 2017-11-21T12:25:19 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 346,056,556 | 829 | dulong | Is this issue resolved ? | 2017-11-21T15:12:28 | 2017-11-21T15:12:28 | {} | NONE |
ClickHouse | ClickHouse | 346,430,123 | 1,513 | ztlpn | Partitioning is built into the table engines of the MergeTree family. Traditionally tables are partitioned by the month of the date column that you supply as the first parameter in the engine definition. Starting from version 54310 you can use custom partitioning keys ([documentation in Russian](https://github.com/yand... | 2017-11-22T18:05:09 | 2017-11-22T18:06:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 346,904,538 | 1,526 | alexey-milovidov | Ok.
And code is better now. | 2017-11-24T23:01:17 | 2017-11-24T23:01:17 | {} | MEMBER |
ClickHouse | ClickHouse | 347,157,250 | 1,541 | ztlpn | @silviucpp yes you are right... The code will become much more convoluted though – we would need to save the fact that we must retry the ZooKeeper operation on the next iteration somewhere in the object state and then have two branches of code based on whether we should retry or not.
In fact the problem is present i... | 2017-11-27T11:39:14 | 2017-11-27T11:39:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 347,437,643 | 1,550 | alniks | We use version 1.1.54292 | 2017-11-28T07:32:41 | 2017-11-28T07:32:41 | {} | NONE |
ClickHouse | ClickHouse | 347,472,558 | 1,553 | filimonov | BTW: there is no documentation for CLEAR COLUMN xxx IN PARTITION yyyymm | 2017-11-28T10:00:46 | 2017-11-28T10:00:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 347,498,248 | 1,541 | silviucpp | Problem https://github.com/yandex/ClickHouse/issues/1546 seems not fixed by this patch | 2017-11-28T11:42:49 | 2017-11-28T11:42:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 347,872,248 | 1,281 | akonyaev90 | Guys, this problem is still relevant.
Any user can kill database process with one http-request
http://<host>:8123/?query=select%20toInt64(null)
I think, this is a fiasco! | 2017-11-29T14:14:31 | 2017-11-29T14:15:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 348,231,903 | 1,568 | theseusyang | I think it is not suitable to store a bigger file (5-15MB), but except HDFS. | 2017-11-30T15:57:53 | 2017-11-30T15:57:53 | {} | NONE |
ClickHouse | ClickHouse | 348,257,144 | 1,560 | ludv1x | Yes, it is a bug. | 2017-11-30T17:16:21 | 2017-11-30T17:16:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 345,763,902 | 1,518 | alexey-milovidov | If you don't mind about correct offsets of intervals in different time zones,
just round time to specified precision to GROUP BY.
Example: `toDateTime(intDiv(toUInt32(time), 21600) * 21600)` will round to 6 hour intervals.
Otherwise do something like this: `toStartOfDay(time) + intDiv(toUInt32(toTime(time)), 21600) ... | 2017-11-20T17:14:56 | 2017-11-20T17:15:13 | {} | MEMBER |
ClickHouse | ClickHouse | 346,359,022 | 1,536 | filimonov | Please, remember that ClickHouse is not a key-value storage, and extracting one row with high level of concurrency is not its' strong point. ClickHouse is DB for analytics, to execute rather small amount of cpu-intensive requests, not a high number of simple data extractions.
Documentations says:
"When NOT to use C... | 2017-11-22T14:04:43 | 2017-11-22T14:10:25 | {
"+1": 11
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 346,507,188 | 1,516 | alexey-milovidov | All `BlockInputStream`s that must do some meaningful work before first `read` in `readPrefix` method
(the most important example is `RemoteBlockInputStream` - it must send query to remote server)
also have the logic "if `readPrefix` was not called, do all required stuff in first `read`".
Example:
```
void Remote... | 2017-11-23T00:20:22 | 2017-11-23T00:20:22 | {} | MEMBER |
ClickHouse | ClickHouse | 346,715,512 | 1,466 | code-of-kpp | Still can't avoid segfault.
Got a few warnings during build, may be related:
```
[100%] Linking CXX executable clickhouse /usr/lib64/libcrypto.so.41: warning: warning: EVP_E... | 2017-11-24T01:02:04 | 2017-11-24T01:02:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 347,532,216 | 1,552 | alniks | I stopped using views in ClickHouse due to this problem - had to switch to long queries. And if you make a view materialized and run the last query the system becomes unresponsive for a couple of hours (depending of data volume in underlying tables) | 2017-11-28T14:02:19 | 2017-11-28T14:02:19 | {} | NONE |
ClickHouse | ClickHouse | 346,361,941 | 1,536 | silviucpp | @filimonov Maximum number of concurrent requests in ClickHouse is set to 100 by default (can be increased in config file). whats the name of that param ?
Silviu | 2017-11-22T14:15:20 | 2017-11-22T14:15:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 346,436,789 | 1,531 | vavrusa | What you're limiting is memory allowance for queries. Best thing you can probably do is to use low granularity primary key, disable uncompressed page cache, and rebuild with libc malloc and use MALLOC_TRIM_THRESHOLD_ to force heap shrinking, but it's a bit silly. | 2017-11-22T18:31:30 | 2017-11-22T18:31:30 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 347,287,212 | 1,541 | alexey-milovidov | Probably the simpliest solution is to mark queue as dirty and re-initialize it in next iteration. | 2017-11-27T18:56:16 | 2017-11-27T18:56:16 | {} | MEMBER |
ClickHouse | ClickHouse | 347,291,603 | 1,548 | alexey-milovidov | Ok, but we should make all existing dictionary sources registered by the same method. | 2017-11-27T19:09:15 | 2017-11-27T19:09:15 | {} | MEMBER |
ClickHouse | ClickHouse | 347,407,774 | 1,543 | theseusyang | @KochetovNicolai welcome give me the precious suggestions. | 2017-11-28T04:07:24 | 2017-11-28T04:07:24 | {} | NONE |
ClickHouse | ClickHouse | 346,227,356 | 1,516 | amosbird | OK. Do we still need to prepare the child streams? From `UnionBlockInputStream.h` it says running preparation in parallel but I couldn't find the implementation.
```
/// Do nothing, to make the preparation for the query execution in parallel, in ParallelInputsProcessor.
void readPrefix() override
{
... | 2017-11-22T02:48:24 | 2017-11-22T02:48:24 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 346,443,907 | 1,514 | alexey-milovidov | It's difficult to integrate complex graph processing directly into ClickHouse,
because ClickHouse gains its performance advantages from flat, uniform, well structured data.
I see only few possibilities:
- using ClickHouse as a storage layer for large graphs, as you can quickly fetch data from ClickHouse based on s... | 2017-11-22T18:59:15 | 2017-11-22T18:59:15 | {
"+1": 7
} | MEMBER |
ClickHouse | ClickHouse | 346,556,171 | 1,537 | jpiper | Ah,
It looks like I found it in the documentation:
```
quantilesIfState(0.25,0.5,0.75)(downloadBytes/downloadTime, downloadBytes > 50000)
``` | 2017-11-23T08:33:05 | 2017-11-23T08:33:05 | {} | NONE |
ClickHouse | ClickHouse | 347,214,262 | 1,541 | silviucpp | @ztlpn what we need to convert to async is the zookeeper interaction, from what I see it's the only one that can block the threads for long time | 2017-11-27T15:20:13 | 2017-11-27T15:20:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 348,197,663 | 1,570 | alexey-milovidov | Отлично, давно хотели сделать такое! | 2017-11-30T14:08:19 | 2017-11-30T14:08:19 | {} | MEMBER |
ClickHouse | ClickHouse | 348,504,033 | 1,563 | filimonov | More about architecture: for example why not to store function signatures? In that case you can check the signatures during parsing (not while execution function). You will also know exact response type. And while writing function you will not need to check & sanitize user input. Bad input just can't get into AST, and... | 2017-12-01T14:11:00 | 2017-12-01T14:21:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 345,823,906 | 508 | rpfennin | I am noticing this as well on the latest stable build (1.1.54310) | 2017-11-20T20:38:02 | 2017-11-20T20:38:02 | {} | NONE |
ClickHouse | ClickHouse | 345,870,949 | 1,514 | alexey-milovidov | It's dependent on what kind of graph computations do you want to execute in ClickHouse.
Let's start with any example of real task and we will look how it will be possible to adapt ClickHouse for it. | 2017-11-20T23:56:17 | 2017-11-20T23:56:17 | {} | MEMBER |
ClickHouse | ClickHouse | 346,279,576 | 1,526 | avasiliev | Пофиксил проблемы | 2017-11-22T08:35:55 | 2017-11-22T08:35:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 346,365,138 | 1,536 | filimonov | @silviucpp it's called max_concurrent_queries. See
https://github.com/yandex/ClickHouse/blob/4fcb081f5b0334663aa5c503682bd30db5da1715/dbms/src/Server/config.xml#L72
Hmm... And looks like it present only in russian documentation:
https://clickhouse.yandex/docs/ru/operations/server_settings/settings.html#max-concu... | 2017-11-22T14:26:58 | 2017-11-22T14:26:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 346,454,864 | 1,535 | alexey-milovidov | Missing functional test.
Look at `dbms/tests/queries` directory.
Please find some time to add a test in separate PR.
Otherwise there is high chance that we will break it sometimes. | 2017-11-22T19:44:34 | 2017-11-22T19:44:34 | {} | MEMBER |
ClickHouse | ClickHouse | 347,962,328 | 1,563 | silviucpp | Nullable support I'm not sure it's considered anyway stable or I personally believe they are not using it yet in production. There are too many bugs with nullable columns: alter not working, some aggregate functions are not working as well, crashes like the one in ticket 1281. | 2017-11-29T19:03:41 | 2017-11-29T19:03:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 345,699,048 | 1,466 | proller | Can you try latest develop or 1.8.0 ? | 2017-11-20T13:43:38 | 2017-11-20T13:43:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 345,718,455 | 1,510 | AlexPosix | goranc thanks, i will try in our case. | 2017-11-20T14:52:14 | 2017-11-20T14:52:14 | {} | NONE |
ClickHouse | ClickHouse | 346,874,077 | 1,546 | proller | Looks like known bug https://github.com/yandex/ClickHouse/pull/1541 | 2017-11-24T17:34:19 | 2017-11-24T17:34:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 346,940,410 | 156 | otisg | It's great you are publishing https://github.com/yandex/ClickHouse/blob/master/CHANGELOG.md !
One small suggestion: include the release date. It is really easy to add and very valuable. | 2017-11-25T13:24:11 | 2017-11-25T13:24:11 | {} | NONE |
ClickHouse | ClickHouse | 347,498,151 | 1,546 | silviucpp | @proller I tested now using the branch you mentions and problem is not fixed. | 2017-11-28T11:42:22 | 2017-11-28T11:42:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 348,081,515 | 1,464 | AlexMAS | Hello guys! Is there any release plan to fix Dockerfile or where can I see the schedule or roadmap? P.s. I apologize I'm so importunate but this issue is quite important to me. | 2017-11-30T04:47:17 | 2017-11-30T04:47:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 345,714,214 | 1,466 | code-of-kpp | latest -testing tag fails to compile
```
[ 1%] Built target cityhash
[ 1%] Built target farmhash
[ 1%] Built target metrohash
[ 2%] Built target cctz
[ 3%] Built target cpuid
[ 3%] Built target pocoext
[ 4%] Built target common
[ 4%] Built target our_mysql_client
[ 6%] Built target mysqlxx
[ 7%] ... | 2017-11-20T14:38:22 | 2017-11-20T14:38:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 345,973,873 | 1,059 | jaykelin | @proller How is the case test going?
we has merge request #1233 to 1.1.54310-stable in our production cluster, i hope this request will be merged in the next release. | 2017-11-21T09:51:56 | 2017-11-21T09:51:56 | {} | NONE |
ClickHouse | ClickHouse | 346,146,758 | 1,527 | alexey-milovidov | What about removing old library? | 2017-11-21T20:11:33 | 2017-11-21T20:11:33 | {} | MEMBER |
ClickHouse | ClickHouse | 346,302,554 | 1,534 | msoltyspl | This looks quite similar to what I'm getting (#1522) - from common parts between our cases, we both use distributed tables (in my case 3 shards x 2 replicas). ('Code' and exact wording is different though 99 vs 101). | 2017-11-22T10:04:45 | 2017-11-22T10:05:23 | {} | NONE |
ClickHouse | ClickHouse | 347,932,042 | 1,130 | filimonov | Check #75 | 2017-11-29T17:19:50 | 2017-11-29T17:19:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 348,493,388 | 1,575 | proller | One boost patch lost, but pr sent.
https://github.com/boostorg/filesystem/pull/64 | 2017-12-01T13:22:42 | 2017-12-01T13:22:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 348,516,803 | 1,568 | daledude | I store some log data in Clickhouse but I don't use CH to search the MSG field. I pre-split what interests me from the MSG into fields and search on those fields. I found Elasticsearch is faster and more convenient for searching logs due to the way it parses/indexes the MSG. I just wish it had as good of compression as... | 2017-12-01T15:01:37 | 2017-12-01T15:01:37 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 345,872,164 | 1,466 | code-of-kpp | CMake output
```
+ USE_STATIC_LIBRARIES=1
+ DISABLE_MONGODB=1
+ ENAB... | 2017-11-21T00:02:53 | 2017-11-21T00:02:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 346,041,024 | 1,524 | silviucpp | I can confirm that's the commit that breaks everithing | 2017-11-21T14:22:28 | 2017-11-21T14:22:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 346,581,328 | 1,538 | alexey-milovidov | 👍 | 2017-11-23T10:30:10 | 2017-11-23T10:30:10 | {} | MEMBER |
ClickHouse | ClickHouse | 346,875,429 | 1,546 | silviucpp | @proller thanks a lot ! I'll update to that branch and retest. will let you know about results | 2017-11-24T17:45:25 | 2017-11-24T17:45:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 348,397,515 | 1,573 | bobrik | Number of parts per table also increased, top 4 lines are the table in question and its 3 MVs. Other abruptly elevated lines include inner tables for MVs.
There's also another group with lower elevation, th... | 2017-12-01T04:21:19 | 2017-12-01T04:21:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 348,500,109 | 1,563 | filimonov | @silviucpp unfortunately it's not only about NULLs.
For example try `SELECT caseWithExpr();` Mine v1.1.54304 dies.
I hope that @alexey-milovidov and all clickhouse team will find some general solution to prevent server crashes caused by user input in the future, but probably architecture improvements can take so... | 2017-12-01T13:53:53 | 2017-12-01T13:55:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 345,839,291 | 1,510 | alexey-milovidov | We confirm, that version 54310 has the regression.
Fix is on the way. | 2017-11-20T21:34:49 | 2017-11-20T21:34:49 | {} | MEMBER |
ClickHouse | ClickHouse | 345,877,421 | 1,466 | alexey-milovidov | Try this: https://github.com/yandex/ClickHouse/commit/e4d354e39af0db677716b2b8c24a25ea8b176e15 | 2017-11-21T00:34:00 | 2017-11-21T00:34:00 | {} | MEMBER |
ClickHouse | ClickHouse | 346,224,572 | 1,516 | amosbird | Hmm, does that mean there are other hidden bugs or the `readPrefix` simply cannot be done in parallel ? | 2017-11-22T02:29:05 | 2017-11-22T02:29:05 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 346,545,708 | 1,510 | AlexPosix | Before, we splited batch of inserts on 1000, this means that if we get for example 80000 batch and then split it on 1000, we inserted 8 batch of 1000 inserts and got an error, now we do not split, we insert full batch and now all it's ok. Thanks. | 2017-11-23T07:23:34 | 2017-11-24T08:50:31 | {} | NONE |
ClickHouse | ClickHouse | 347,321,572 | 294 | blinkov | 1. The reference_en.html and reference_ru.html are the original version of documentation and are currenly deprecated. It's actually published as https://clickhouse.yandex/deprecated/reference_en.html and already has your changes.
2. The current documentation lives here https://github.com/yandex/ClickHouse/tree/master/... | 2017-11-27T20:43:00 | 2017-11-27T20:43:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 348,134,848 | 1,464 | Slach | Guys if you want use Docker VOLUMES just need set right UID:GID to /var/lib/clickhouse/data
see answer https://stackoverflow.com/a/46967182/1204665
as i know currently docker-compose don't support UID:GID setting in volumes: syntax
| 2017-11-30T09:42:26 | 2017-11-30T09:42:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 345,719,959 | 1,466 | proller | Are you sure about actual 'testing' ?
dbms/include was removed half year ago.
Can you chow cmake output ? | 2017-11-20T14:56:48 | 2017-11-20T14:56:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 345,764,499 | 1,517 | alexey-milovidov | Yes, `erase(it++)` is common practice. | 2017-11-20T17:16:55 | 2017-11-20T17:16:55 | {} | MEMBER |
ClickHouse | ClickHouse | 346,729,014 | 1,514 | theseusyang | @alexey-milovidov Yes, We are researching how to use ClickHouse as an graph storage layer, maybe We need extend the underlying Storage API to complete it. | 2017-11-24T02:42:01 | 2017-11-24T02:42:01 | {} | NONE |
ClickHouse | ClickHouse | 346,786,926 | 1,540 | filimonov | Found a workaround for getting min/max of partitioning key without fullscans (not sure how it will work with new custom partitioning keys):
```
SELECT
min(min_date),
max(max_date)
FROM system.parts
WHERE table = 'table_name'
``` | 2017-11-24T09:48:02 | 2017-11-24T09:49:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 346,903,527 | 1,545 | alexey-milovidov | > Also - found and fix similar bug in modulo function. The bug was in oferflowing second number while cast to result type.
Thank you for the fix! | 2017-11-24T22:44:49 | 2017-11-24T22:44:49 | {} | MEMBER |
ClickHouse | ClickHouse | 347,323,587 | 294 | alexey-milovidov | It was forgotten while changing documentation engine. I just created a task to restore the documentation of `topK` and add `anyHeavy`. | 2017-11-27T20:50:29 | 2017-11-27T20:50:29 | {} | MEMBER |
ClickHouse | ClickHouse | 347,330,615 | 294 | vavrusa | Ah okay, thanks! | 2017-11-27T21:16:25 | 2017-11-27T21:16:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 347,827,188 | 1,415 | StarWix | Function `intDiv` has the same problems | 2017-11-29T11:05:27 | 2017-11-29T11:05:27 | {} | NONE |
ClickHouse | ClickHouse | 347,875,075 | 1,130 | Desst | bump!
_date DEFAULT toDate(Epoch)
INSERTing without specifying columns sets _date to 0000-00-00 regardless of Epoch | 2017-11-29T14:24:25 | 2017-11-29T14:24:25 | {} | NONE |
ClickHouse | ClickHouse | 347,319,298 | 294 | vavrusa | I added a basic documentation in the PR https://github.com/yandex/ClickHouse/commit/5f1e65b2523d7fee66404fc764a4c5d281c989fa#diff-f186d69ae6d035b5aecbfa9ccf644ba1 not sure why it's not showing up in documentation pages? cc @blinkov | 2017-11-27T20:34:33 | 2017-11-27T20:34:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 348,519,900 | 1,563 | filimonov | Another reason for making function signatures which should be checked at parse time:
`select asflwernwer( (SELECT max(number) from some_huge_table ) );
`
It will calclulate whole select to tell at the end that there is no such a function asflwernwer.
| 2017-12-01T15:12:55 | 2017-12-01T15:40:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 345,721,954 | 1,466 | code-of-kpp | haha, sorting
```
# git status
HEAD detached at v1.1.6-testing
``` | 2017-11-20T15:03:07 | 2017-11-20T15:03:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 346,454,383 | 1,535 | alexey-milovidov | Looks good, thanks! | 2017-11-22T19:42:31 | 2017-11-22T19:42:31 | {} | MEMBER |
ClickHouse | ClickHouse | 346,848,397 | 1,524 | silviucpp | Problem was that even if I dropped the tables the metadata file didn't cleared and for this reason somehow I got this errors. | 2017-11-24T15:08:29 | 2017-11-24T15:08:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 348,395,888 | 1,573 | bobrik | Looks like memory usage doubled:
```
:) select toStartOfHour(event_time) as hour, user, count() as count, formatReadableSize(avg(memory_usage)) as avgMem, formatReadableSize(max(memory_usage)) as maxMem, formatReadableSize(sum(memory_usage)) as sumMem from system.query_log where user = 'writer_secure' and event_dat... | 2017-12-01T04:05:16 | 2017-12-01T04:05:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 346,783,995 | 1,540 | valyala | See the related issue #530 . | 2017-11-24T09:34:58 | 2017-11-24T09:34:58 | {} | NONE |
ClickHouse | ClickHouse | 346,876,660 | 1,541 | silviucpp | Hello,
Looking to this PR doesn't seems 100 % correct. Now we moved the ReplicationMergeTree to not use any longer dedicated threads per table and instead a pool of threads.
And in this PR if I understand correctly it's inserted an infinite loop:
```
for (size_t num_tries = 0; ; ++num_tries)
```
Which bl... | 2017-11-24T17:55:42 | 2017-11-24T17:55:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 348,253,771 | 1,562 | ludv1x | At the moment lazy calculation works only for pure constants, not for constant expressions.
So, queries like `SELECT 1 ? 0 : 1 % 0` work. | 2017-11-30T17:04:41 | 2017-11-30T17:04:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 348,495,092 | 11 | hannes | Since CH already bundles Clang, how about UDFs in C/CPP? Scalar functions could be a good start... | 2017-12-01T13:30:59 | 2017-12-01T13:30:59 | {} | NONE |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.