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
564,365,138
8,122
alexey-milovidov
@amosbird Possible solutions: - rename GRANULARITY to GRANULARITY FACTOR and support old syntax only for ATTACH queries; - remove GRANULARITY from documentation example (so it will be 1 by default) and only briefly mention it;
2019-12-11T03:36:10
2019-12-11T03:36:10
{}
MEMBER
ClickHouse
ClickHouse
564,368,638
8,132
alexey-milovidov
Could you please rebase these changes on top of master?
2019-12-11T03:54:02
2019-12-11T03:54:02
{}
MEMBER
ClickHouse
ClickHouse
564,371,456
8,122
kaijianding
@den-crane thanks, it works after change GRANULARITY 8192 to GRANULARITY 1. close this issue
2019-12-11T04:09:58
2019-12-11T04:09:58
{}
NONE
ClickHouse
ClickHouse
564,531,938
7,170
KochetovNicolai
Some perf compares: https://gistcdn.githack.com/KochetovNicolai/18c6a4b54cfcc8e59856305391504348/raw/c84a17107950dd0c077395efc166aaa39567264a/dt64_comp.html
2019-12-11T13:09:24
2019-12-11T13:09:24
{ "heart": 2 }
MEMBER
ClickHouse
ClickHouse
564,806,434
8,085
tisonkun
Hi @abyss7 , any news here?
2019-12-12T01:19:39
2019-12-12T01:19:39
{}
CONTRIBUTOR
ClickHouse
ClickHouse
564,999,039
8,056
alexey-milovidov
We should enable all format-related settings for Kafka.
2019-12-12T13:06:27
2019-12-12T13:06:27
{}
MEMBER
ClickHouse
ClickHouse
565,022,459
8,172
filimonov
> Query: `exists some.some` Doesn't look like SQL. Can you check what exactly SQL was sent to ClickHouse ? You can find it in /var/log/clickhouse-server/clickhouse-server.log before segfault report.
2019-12-12T14:11:22
2019-12-12T14:11:22
{}
CONTRIBUTOR
ClickHouse
ClickHouse
564,459,667
8,121
Crazylearner30
@alexey-milovidov I have partitioned the table on state code which has some 60 values So if I don't use the PARTITION BY it doesn't have impact on queries??
2019-12-11T09:40:06
2019-12-11T09:40:06
{}
NONE
ClickHouse
ClickHouse
564,509,847
2,542
nikitamikhaylov
Fixed.
2019-12-11T12:02:29
2019-12-11T12:02:29
{ "hooray": 1 }
MEMBER
ClickHouse
ClickHouse
564,574,075
8,141
nikitamikhaylov
Need to test it on real Mac.
2019-12-11T14:39:28
2019-12-11T14:39:28
{}
MEMBER
ClickHouse
ClickHouse
564,796,033
8,129
filimonov
Maybe related to #6781 /cc @tavplubix
2019-12-12T00:34:07
2019-12-12T00:34:07
{}
CONTRIBUTOR
ClickHouse
ClickHouse
564,799,278
8,151
alexey-milovidov
Maybe it's enough to only build some libraries with this option. libcxxabi maybe?
2019-12-12T00:48:54
2019-12-12T00:48:54
{}
MEMBER
ClickHouse
ClickHouse
564,927,664
8,011
Jokser
@alexey-milovidov Everything is done. Could you please look at changes again?
2019-12-12T09:39:21
2019-12-12T09:39:21
{}
CONTRIBUTOR
ClickHouse
ClickHouse
564,336,689
7,946
alexey-milovidov
We can store directories and files on local filesystem, but instead of file contents, store S3 URL along with refcount inside them, and store actual content in S3 (we can use arbitrary path in S3, even generated UUID as file name). Advantages: - we will be able to do atomic renames of files and directories; - we w...
2019-12-11T01:31:56
2019-12-11T01:40:46
{ "+1": 1 }
MEMBER
ClickHouse
ClickHouse
564,369,144
8,132
alexey-milovidov
Or I can simply cherry-pick these two commits: https://github.com/ClickHouse/ClickHouse/pull/8132/commits/1c3210c15c67b161a0d8925c00392764bd1b0142 https://github.com/ClickHouse/ClickHouse/pull/8132/commits/74eb874b04e66a75201c7b93228345abeddf12cc
2019-12-11T03:56:40
2019-12-11T03:56:40
{}
MEMBER
ClickHouse
ClickHouse
564,469,492
8,034
hombit
I've rebased it on current master
2019-12-11T10:05:20
2019-12-11T10:05:20
{}
CONTRIBUTOR
ClickHouse
ClickHouse
564,617,367
8,145
samuelblawrence
Ah - I've found it - it's the extra argument on AggregateFunction - posting to help others! Thanks, S ``` CREATE TABLE from_table ( id UInt32, version UInt32, value UInt32 ) ENGINE = ReplacingMergeTree(version) PARTITION BY tuple() order by id; CREATE TABLE to_table ( id UInt32, value Aggregat...
2019-12-11T16:13:50
2019-12-11T16:14:14
{ "+1": 2 }
NONE
ClickHouse
ClickHouse
564,920,622
7,170
KochetovNicolai
Will do it in other pr. This one is huge already.
2019-12-12T09:20:15
2019-12-12T09:20:15
{}
MEMBER
ClickHouse
ClickHouse
564,987,880
8,165
alexey-milovidov
@KochetovNicolai It looks like the Merge table is created on top of Distributed tables, and one of the shards of Distributed table is local; and constants are materialized for remote shards but not for local shards?
2019-12-12T12:31:35
2019-12-12T12:31:35
{}
MEMBER
ClickHouse
ClickHouse
565,032,682
8,162
akuzm
Oops, seems like we don't have a gcc 8 build in CI. I'll check if we can add it.
2019-12-12T14:37:28
2019-12-12T14:37:28
{}
CONTRIBUTOR
ClickHouse
ClickHouse
564,398,387
7,812
yuzhichang
@Akazz @akuzm The failed `Stress test` doesn't relate to my change.
2019-12-11T06:22:11
2019-12-11T06:22:11
{}
CONTRIBUTOR
ClickHouse
ClickHouse
564,448,580
8,061
ghuname
It is quite pity how clickhouse handles null values, and that using null values causes performance issues. Is there a possibility that handling null values will change in the future?
2019-12-11T09:11:17
2019-12-11T09:11:17
{}
NONE
ClickHouse
ClickHouse
564,603,461
8,132
kmatt
Replaced with #8148
2019-12-11T15:43:36
2019-12-11T15:43:36
{}
NONE
ClickHouse
ClickHouse
564,777,089
7,170
alexey-milovidov
We decided to wait for support for arithmetic operations on INTERVALs and dateDiff function.
2019-12-11T23:19:04
2019-12-11T23:19:04
{}
MEMBER
ClickHouse
ClickHouse
564,828,605
8,039
toby080508050805
After updating to version 19.17.15, this issues already be fixed.
2019-12-12T03:00:52
2019-12-12T03:00:52
{ "+1": 1 }
NONE
ClickHouse
ClickHouse
564,828,941
8,120
den-crane
> Hi, den, > > We do not want to recreate the existing MVs, We want to use these MVs but do not want clickhouse checking new data and insert to these MVs. > > Thanks Do you mean you need a data in MV ? You can detach MV and rename .inner table to a table with normal name and remove MV.sql file.
2019-12-12T03:02:37
2019-12-12T03:02:37
{}
CONTRIBUTOR
ClickHouse
ClickHouse
564,333,913
8,112
alexey-milovidov
Ok. Could you please add performance test and check how slower the function is comparing to the existing round? If it's no slower, we can make it the default. PS. Probably the better name is `bankersRound` instead of `bankerRound`?
2019-12-11T01:19:26
2019-12-11T01:19:26
{}
MEMBER
ClickHouse
ClickHouse
565,008,229
8,169
4ertus2
Yes, we added a check for columns with same names cause there're some queries with wrong column selection in this case. For now it throws an error if it's not sure what column to select: from the first table or from the second one. In particular in your query implementation needs to crop `t1.c1` into `c1` then it de...
2019-12-12T13:32:46
2019-12-12T13:36:57
{}
CONTRIBUTOR
ClickHouse
ClickHouse
565,046,035
8,172
artyomkorzun
Java code snippet: ``` java try (Connection connection = DriverManager.getConnection("jdbc:clickhouse://10.10.87.163:8123?password=----"); Statement statement = connection.createStatement()) { statement.execute("exists somedatabase.sometable"); } ``` The full log: ``` Include not found: clickhous...
2019-12-12T15:08:31
2019-12-12T15:08:31
{}
NONE
ClickHouse
ClickHouse
564,245,355
8,128
alexey-milovidov
`,\1` is 300 little endian - as expected.
2019-12-10T20:33:05
2019-12-10T20:33:05
{}
MEMBER
ClickHouse
ClickHouse
564,301,942
8,011
alexey-milovidov
Almost ready, only minor changes remain.
2019-12-10T23:11:40
2019-12-10T23:11:40
{}
MEMBER
ClickHouse
ClickHouse
564,553,668
8,139
alesapin
This PR affects build, so we can merge it without tests.
2019-12-11T13:58:27
2019-12-11T13:58:27
{}
MEMBER
ClickHouse
ClickHouse
564,722,119
8,136
volfco
The docs give a great overview of partitioning: https://clickhouse.yandex/docs/en/operations/table_engines/mergetree/#use-of-indexes-and-partitions-in-queries
2019-12-11T20:37:20
2019-12-11T20:37:20
{}
NONE
ClickHouse
ClickHouse
564,793,454
8,121
filimonov
> @byx313 > yes, I did ulimit -a to check and yes the change is reflected. Also for clickhouse user? What is your OS? How did you install/run clickhouse? I'm asking because official packages should extend that limit during installation, and 9fficial docker readme mentions how to increase max number of opened fil...
2019-12-12T00:23:19
2019-12-12T00:25:53
{}
CONTRIBUTOR
ClickHouse
ClickHouse
564,949,276
8,120
jsding
The original table is too big and uncontrollable increased, we have to query data daily and combine them in application layer.
2019-12-12T10:34:39
2019-12-12T10:34:39
{}
NONE
ClickHouse
ClickHouse
564,352,770
7,776
alexey-milovidov
Now it is published: https://clickhouse.yandex/docs/ja/
2019-12-11T02:46:09
2019-12-11T02:46:09
{}
MEMBER
ClickHouse
ClickHouse
564,651,270
8,097
alesapin
Unfortunately, without this option, we broke exceptions handling in clang release build. It doesn't test in CI (we will add it), but it's common build type for developers. ``` $ cmake ../ClickHouse -DENABLE_TESTS=0 -DENABLE_UTILS=0 -DCMAKE_CXX_COMPILER=/usr/bin/clang++-8 -DCMAKE_C_COMPILER=/usr/bin/clang-8 -DENABLE...
2019-12-11T17:32:54
2019-12-11T17:33:37
{}
MEMBER
ClickHouse
ClickHouse
564,779,928
7,995
kolen
The same on 10.13.6 (x86_64-apple-darwin17.7.0), with clang 10.0.0 (clang-1000.10.44.4). Latest master requires only clang on darwin.
2019-12-11T23:29:56
2019-12-11T23:29:56
{}
NONE
ClickHouse
ClickHouse
564,822,947
8,120
jsding
I asked this question, because I assumed too many mvs to one table should hurt the clickhouse performance? Is that true?
2019-12-12T02:33:15
2019-12-12T02:33:15
{}
NONE
ClickHouse
ClickHouse
564,988,543
8,129
alexey-milovidov
@tavplubix Are we ready to enable it by default? You can make a PR with this change.
2019-12-12T12:33:46
2019-12-12T12:33:54
{}
MEMBER
ClickHouse
ClickHouse
564,296,581
8,120
jsding
Hi, I create material view by ` create MATERIALIZED view v_event_20191209 ENGINE = MergeTree() PARTITION BY (package_name) ORDER BY (package_name, campaign, toDate(it), event_token) AS select * from rp_event where toDate(created_at/1000) = '2019-12-09' ` All new data coming after 2019-12-09 will not match the ...
2019-12-10T22:53:26
2019-12-10T22:53:41
{}
NONE
ClickHouse
ClickHouse
564,408,461
8,136
blinkov
@byx313 so what's your question?
2019-12-11T07:03:01
2019-12-11T07:03:01
{}
CONTRIBUTOR
ClickHouse
ClickHouse
564,468,952
8,121
Crazylearner30
@alexey-milovidov Yes, I removed the PARTITION BY and without any error I could load the data :) My worry is query returning time. Thank you.
2019-12-11T10:03:48
2019-12-11T10:03:48
{}
NONE
ClickHouse
ClickHouse
564,559,667
8,120
den-crane
@jsding what are you trying to achieve? Do you need to recreate existing MV ?
2019-12-11T14:11:32
2019-12-11T14:11:32
{}
CONTRIBUTOR
ClickHouse
ClickHouse
564,659,658
8,150
amosbird
btw, we can revert this only for clang builds
2019-12-11T17:54:16
2019-12-11T17:54:16
{}
COLLABORATOR
ClickHouse
ClickHouse
564,831,095
8,136
den-crane
Every part has an attributes minmax_partitioncolumn.idx. All parts is in memory with this minmax (CH reads them on start and keeps them in memory). So if your table partitioned by toYYYYMM(date_column) every part will have min and max values of date_column. A query analyzes (partition pruning) an intersection of a...
2019-12-12T03:14:36
2019-12-12T03:15:22
{}
CONTRIBUTOR
ClickHouse
ClickHouse
564,953,433
7,468
alesapin
If you have any questions, just reopen.
2019-12-12T10:44:56
2019-12-12T10:44:56
{}
MEMBER
ClickHouse
ClickHouse
564,356,173
8,122
amosbird
I always find the term "granularity" to be overly used. We have `index_granularity` meaning the max row number of a granule, and we have index granularity meaning the granules one index unit covers.
2019-12-11T02:52:20
2019-12-11T02:52:20
{}
COLLABORATOR
ClickHouse
ClickHouse
564,943,115
8,120
nvartolomei
Why do you need to create a MV per day instead of having a single one partitioned by day?
2019-12-12T10:18:59
2019-12-12T10:18:59
{}
CONTRIBUTOR
ClickHouse
ClickHouse
564,967,032
8,143
Akazz
The only failed test is `00995_exception_while_insert` that has been noticed to be flapping recently
2019-12-12T11:24:21
2019-12-12T11:24:21
{}
CONTRIBUTOR
ClickHouse
ClickHouse
565,060,841
7,997
majedrze
I have the same issue. Since upgrading Clickhouse from 18.x series. After issuing optimize table on cluster process hangs on master until server restart, while all the other nodes respond with "Cannot execute replicated DDL query on leader". ``` ch-04 :) select query_id, query from system.processes SELECT ...
2019-12-12T15:41:43
2019-12-12T15:41:43
{}
NONE
ClickHouse
ClickHouse
564,346,863
8,091
onine007
@den-crane if A and B are Nullable already and some rows are NULL ?
2019-12-11T02:19:24
2019-12-11T03:01:37
{}
NONE
ClickHouse
ClickHouse
564,509,318
5,767
nikitamikhaylov
Yes, #2542 is exactly the same. And all test from that issue are working now without any exceptions.
2019-12-11T12:00:41
2019-12-11T12:00:41
{}
MEMBER
ClickHouse
ClickHouse
564,790,866
8,120
jsding
Hi, den, We do not want to recreate the existing MVs, We want to use these MVs but do not want clickhouse checking new data and insert to these MVs. Thanks
2019-12-12T00:12:45
2019-12-12T00:12:45
{}
NONE
ClickHouse
ClickHouse
564,902,315
8,014
ghost
The bloom filter feature of ClickHouse is awesome and very fast to filter rows. But, I think there are two issues in bloom filter. 1. `GRANULARITY` of index doesn't applied. 2. When merging two bloom filters, it takes too long time. The time complexity of this job may be O(n^3). `making hashes for the elements of ...
2019-12-12T08:27:34
2019-12-12T08:39:05
{ "+1": 1 }
NONE
ClickHouse
ClickHouse
564,988,908
8,165
KochetovNicolai
Indeed. We need to properly calculate common header for streams.
2019-12-12T12:34:56
2019-12-12T12:34:56
{}
MEMBER
ClickHouse
ClickHouse
565,033,684
7,910
filimonov
> Hi, > > Thanks for your help. > > We installed clickhouse-server refer this page: > https://clickhouse.yandex/docs/en/getting_started/ > `sudo yum install clickhouse-server clickhouse-client ` > How to upgrade to 19.17.4 safely? 1) run the same command again to update the packages. 2) restart clickhouse-...
2019-12-12T14:39:56
2019-12-12T14:40:11
{}
CONTRIBUTOR
ClickHouse
ClickHouse
564,225,760
8,115
alexey-milovidov
Missing comments.
2019-12-10T20:15:10
2019-12-10T20:15:10
{}
MEMBER
ClickHouse
ClickHouse
564,453,713
7,170
KochetovNicolai
Surely, I don't suppose to add all features to current pr :) Just have written some ideas which are good to do later.
2019-12-11T09:25:14
2019-12-11T09:25:14
{}
MEMBER
ClickHouse
ClickHouse
564,553,298
8,035
filimonov
That changes fail that tests: https://github.com/filimonov/ch-kafka-consume-perftest/blob/master/test_formats.sh issue with eof - fix was here https://github.com/ClickHouse/ClickHouse/pull/8016/commits/6e426592a7599a373ce19baae50b34ded9bee737 Now should be fixed another way. Sample stacktrace: ``` 2019.12.1...
2019-12-11T13:57:38
2019-12-11T14:53:21
{}
CONTRIBUTOR
ClickHouse
ClickHouse
564,590,673
8,146
Felixoid
Now it's broken in an ugly way https://clickhouse.yandex/docs/ru/operations/system_tables/#system-graphite-retentions
2019-12-11T15:14:51
2019-12-11T15:14:51
{}
MEMBER
ClickHouse
ClickHouse
564,829,148
8,120
den-crane
> I asked this question, because I assumed too many mvs to one table should hurt the clickhouse performance? Is that true? How many? 10 ? 100? Of course any number impact on performance.
2019-12-12T03:03:50
2019-12-12T03:03:50
{}
CONTRIBUTOR
ClickHouse
ClickHouse
564,860,457
8,120
jsding
We roll data daily, so there will be 300+ mv for the table.
2019-12-12T05:51:00
2019-12-12T05:51:00
{}
NONE
ClickHouse
ClickHouse
564,998,592
8,160
filimonov
Currentry you can use JSONEachRow for JSONs (but it's much simpler) After fixing https://github.com/ClickHouse/ClickHouse/issues/8056 it can be done via custom template.
2019-12-12T13:05:13
2019-12-12T13:06:13
{}
CONTRIBUTOR
ClickHouse
ClickHouse
565,007,174
7,772
filimonov
Test case illustrating the problem ``` SELECT toUInt8(bitAnd(number, 4)) AS a, toUInt8(bitAnd(number, 2)) AS b, toUInt8(bitAnd(number, 1)) AS c, a AND b AND c AS `a and b and c`, a OR b OR c AS `a or b or c`, ...
2019-12-12T13:30:01
2019-12-12T13:31:39
{}
CONTRIBUTOR
ClickHouse
ClickHouse
565,040,525
8,172
artyomkorzun
This exactly what was sent. No messages in the log since it failed on parsing the query.
2019-12-12T14:55:24
2019-12-12T14:55:24
{}
NONE
ClickHouse
ClickHouse
564,341,466
8,088
alexey-milovidov
I have added a comment. Please check that I understand the changes correctly :)
2019-12-11T01:54:01
2019-12-11T01:54:01
{}
MEMBER
ClickHouse
ClickHouse
564,400,147
5,767
filimonov
Isn't it the same as #2542? @nikitamikhaylov can you check that?
2019-12-11T06:29:25
2019-12-11T06:29:25
{}
CONTRIBUTOR
ClickHouse
ClickHouse
564,600,593
8,132
kmatt
> Could you please rebase these changes on top of master? @alexey-milovidov I will rebase.
2019-12-11T15:37:33
2019-12-11T17:26:59
{}
NONE
ClickHouse
ClickHouse
564,947,113
8,155
alesapin
https://github.com/ClickHouse/ClickHouse/pull/8163
2019-12-12T10:29:14
2019-12-12T10:29:14
{}
MEMBER
ClickHouse
ClickHouse
565,051,632
8,164
alexey-milovidov
We can add a test with `SELECT count(1) FROM remote('127.0.0.{1,1,2}')` or something like that?
2019-12-12T15:21:08
2019-12-12T15:21:42
{}
MEMBER
ClickHouse
ClickHouse
564,792,967
8,154
laingawbl
Looking at the issue list, it seems the second case I encountered reproduces https://github.com/ClickHouse/ClickHouse/issues/7528.
2019-12-12T00:21:27
2019-12-12T00:21:27
{}
NONE
ClickHouse
ClickHouse
564,797,202
8,078
filimonov
Woraround: create dictionary with HTTP source, define required lifetime. Once it will be created usin DDL (not XML) you can it as a normal table for select. For older versions you may need to create table with engine=Dictionary above.
2019-12-12T00:39:25
2019-12-12T00:39:25
{ "+1": 2 }
CONTRIBUTOR
ClickHouse
ClickHouse
564,938,421
8,088
dingxiangfei2009
Oh, sorry. Yes, I should have used "wrapping". Thanks for the help. :+1:
2019-12-12T10:07:09
2019-12-12T10:07:09
{}
CONTRIBUTOR
ClickHouse
ClickHouse
564,942,962
8,088
dingxiangfei2009
A reference to the "ZooKeeper Internals" manual is included in the comment, which is a permanent link to the section on the definitions.
2019-12-12T10:18:31
2019-12-12T10:18:31
{}
CONTRIBUTOR
ClickHouse
ClickHouse
564,948,699
5,335
alesapin
@Melancholic want to take this issue.
2019-12-12T10:33:12
2019-12-12T10:33:12
{ "+1": 1 }
MEMBER
ClickHouse
ClickHouse
564,974,526
5,744
AntonVozdvizhenskiy
> А откуда вообще взялось это ID в команде? В документации есть. https://clickhouse.yandex/docs/ru/query_language/alter/#alter-how-to-specify-part-expr Строковый идентификатор партиции. Идентификатор партиции используется для именования кусков партиции на файловой системе и в ZooKeeper. В запросах ALTER идентифик...
2019-12-12T11:48:18
2019-12-12T11:48:18
{}
NONE
ClickHouse
ClickHouse
565,012,154
8,169
vincenzon
@4ertus2, Thanks for the quick response. Based on your response and testing I see that: ``` select t1.c1 from t1 left any join t2 on t1.c1 = t2.c1 ``` works as expected. As it happens the queries in question are program generated so I can add a condition to do the check in my code. I see you did some ret...
2019-12-12T13:43:49
2019-12-12T13:43:49
{}
NONE
ClickHouse
ClickHouse
565,047,420
7,946
alesapin
This code is quite isolated, so we can merge it. But I think `IDisk` interface will significantly change in future.
2019-12-12T15:11:49
2019-12-12T15:11:49
{}
MEMBER
ClickHouse
ClickHouse
565,053,012
8,177
alexey-milovidov
Test needed.
2019-12-12T15:24:20
2019-12-12T15:24:20
{}
MEMBER
ClickHouse
ClickHouse
564,409,727
8,136
byx313
> @byx313 so what's your question? Sorry I have changed the question. Thanks in advance.
2019-12-11T07:07:47
2019-12-11T07:07:47
{}
NONE
ClickHouse
ClickHouse
564,429,168
7,684
maqroll
¿0 = no hops allowed (instead of unlimited number of hops)? cc/ @BayoNet @blinkov
2019-12-11T08:16:45
2019-12-11T08:53:39
{}
CONTRIBUTOR
ClickHouse
ClickHouse
565,052,789
5,744
den-crane
@AntonVozdvizhenskiy ``` create table Z1(A Int64, B Date) Engine=MergeTree Partition by (B,A) order by A; insert into Z1 values (1,today()); select partition, name,partition_id from system.parts where table='Z1'; ┌─partition────────┬─name─────────────┬─partition_id─┐ │ ('2019-12-12',1) │ 20191212-1_1_1_0 │ ...
2019-12-12T15:23:48
2019-12-12T15:23:48
{}
CONTRIBUTOR
ClickHouse
ClickHouse
564,223,433
8,121
alexey-milovidov
This happens because you are using too granular partition key in a table. Solution: do not use `PARTITION BY`.
2019-12-10T20:09:16
2019-12-10T20:09:16
{}
MEMBER
ClickHouse
ClickHouse
564,535,058
8,130
CurtizJ
For storage View it works if it's saved query is like `SELECT ... FROM merge_tree_table ORDER BY pk`. But it's hard to support for queries like `SELECT ... FROM view ORDER BY pk` if saved query doesn't have `ORDER BY` clause, because View doesn't store storage of source table unlike to Buffer, MaterializedView and Merg...
2019-12-11T13:18:31
2019-12-11T13:18:31
{}
MEMBER
ClickHouse
ClickHouse
564,784,473
8,117
stavrolia
LGTM. Could you please add perf test?
2019-12-11T23:48:16
2019-12-11T23:48:16
{}
CONTRIBUTOR
ClickHouse
ClickHouse
564,791,496
8,152
filimonov
Duplicate of #8003 Fixed in #8016
2019-12-12T00:15:12
2019-12-12T00:15:12
{}
CONTRIBUTOR
ClickHouse
ClickHouse
564,217,188
8,126
alexey-milovidov
https://github.com/ClickHouse-Extras/poco/pull/14
2019-12-10T19:53:44
2019-12-10T19:53:44
{}
MEMBER
ClickHouse
ClickHouse
564,299,624
8,120
den-crane
@jsding OK. What is your question then?
2019-12-10T23:03:32
2019-12-10T23:03:32
{}
CONTRIBUTOR
ClickHouse
ClickHouse
564,331,505
8,120
jsding
Hi, den, Should I ` detach table .inner. v_event_20191209 ` Not `detach table v_event_20191209`
2019-12-11T01:09:14
2019-12-11T01:09:14
{}
NONE
ClickHouse
ClickHouse
564,334,390
8,130
alexey-milovidov
Also need to check for Buffer, View, MaterializedView.
2019-12-11T01:21:41
2019-12-11T01:21:41
{}
MEMBER
ClickHouse
ClickHouse
564,450,726
7,170
Enmk
@KochetovNicolai Thanks for a review! > Also, it would be probably more convenient to add scale argument to toDateTime64 well, it is already there, current function prototype is `toDateTime64(value, [scale, timezone])` and it is properly verified in tests already. > Also, it would be better to split pr to sm...
2019-12-11T09:17:02
2019-12-11T09:19:43
{ "+1": 1 }
CONTRIBUTOR
ClickHouse
ClickHouse
564,569,792
8,035
filimonov
The performance test result: ``` │ 19.19.1.1836 │ 33554432 │ 309 │ 108590.39482200648 │ ``` Prev results here: https://github.com/ClickHouse/ClickHouse/pull/7935#issuecomment-559780728
2019-12-11T14:31:11
2019-12-11T14:31:32
{}
CONTRIBUTOR
ClickHouse
ClickHouse
564,721,821
8,136
volfco
Partitions are how the data is stored on disk, when you run a query you don't need to worry about which partition your data resides in. Clickhouse will figure out everything. If you partition using `toStartOfDay`, all records from that day are grouped together. If you do a query for data falling inside a single day,...
2019-12-11T20:36:23
2019-12-11T20:36:23
{}
NONE
ClickHouse
ClickHouse
564,725,705
8,061
den-crane
The future's not set.
2019-12-11T20:46:59
2019-12-11T20:46:59
{}
CONTRIBUTOR
ClickHouse
ClickHouse
564,726,580
8,091
den-crane
@onine007 then no way.
2019-12-11T20:49:17
2019-12-11T20:49:17
{}
CONTRIBUTOR
ClickHouse
ClickHouse
564,791,801
7,840
alexey-milovidov
https://github.com/ClickHouse/ClickHouse/issues/8157 https://github.com/ClickHouse/ClickHouse/issues/8156
2019-12-12T00:16:35
2019-12-12T00:16:35
{}
MEMBER
ClickHouse
ClickHouse
564,975,244
8,129
tavplubix
It should work fine with enabled `input_format_values_deduce_templates_of_expressions` setting. ``` $ time cat test1m.value |clickhouse-client -q "INSERT INTO vtest Values" --input_format_values_deduce_templates_of_expressions=1 real 0m0.321s user 0m0.247s sys 0m0.051s ``` https://clickhouse.yandex/docs/en/opera...
2019-12-12T11:50:37
2019-12-12T11:57:51
{}
MEMBER
ClickHouse
ClickHouse
565,015,403
5,744
nvartolomei
@AntonVozdvizhenskiy just unescape returned before running freeze partition. In your case you just need to drop backslashes.
2019-12-12T13:52:18
2019-12-12T13:52:18
{}
CONTRIBUTOR
ClickHouse
ClickHouse
560,946,030
7,810
alexey-milovidov
All `test_multiple_disks` tests have failed.
2019-12-03T00:43:32
2019-12-03T00:43:38
{}
MEMBER