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 | 301,184,577 | 778 | alexey-milovidov | ```
$ diff <(clickhouse-client -n < 00453_top_k.sql) 00453_top_k.reference
1c1
< [0,1,2,3,4,5,6,7,8,9]
---
> [0,1,2,3,4,5,6,7,8,9]
\ No newline at end of file
```
Fixed. | 2017-05-12T20:59:21 | 2017-05-12T20:59:21 | {} | MEMBER |
ClickHouse | ClickHouse | 301,390,853 | 785 | alexey-milovidov | I just quickly scanned the code and it looks fine. | 2017-05-15T06:50:56 | 2017-05-15T06:51:12 | {} | MEMBER |
ClickHouse | ClickHouse | 301,393,072 | 785 | alexey-milovidov | FYI.
In Yandex Metrica, we are simply using sorted `std::vector` and function `std::lower_bound` to do lookups*, for similar task: mapping IP ranges to geographical locations (geoip). (This is implemented in application code outside of ClickHouse).
You are using trie - it is more powerful, because it support inse... | 2017-05-15T07:03:55 | 2017-05-15T07:03:55 | {} | MEMBER |
ClickHouse | ClickHouse | 301,501,633 | 792 | robot-metrika-test | Can one of the admins verify this patch? | 2017-05-15T14:59:04 | 2017-05-15T14:59:04 | {} | NONE |
ClickHouse | ClickHouse | 302,171,135 | 803 | alexey-milovidov | :+1:
What about adding `actual_new_part_name` to `system.replication_queue`? | 2017-05-17T17:45:06 | 2017-05-17T17:45:06 | {} | MEMBER |
ClickHouse | ClickHouse | 303,551,876 | 821 | robot-metrika-test | Can one of the admins verify this patch? | 2017-05-23T22:34:02 | 2017-05-23T22:34:02 | {} | NONE |
ClickHouse | ClickHouse | 304,421,774 | 827 | jaykelin | @alexey-milovidov but, before insert same block, Has been successfully executed **alter table x drop partition 201705** | 2017-05-27T02:43:37 | 2017-05-27T02:44:22 | {} | NONE |
ClickHouse | ClickHouse | 305,841,737 | 839 | ztlpn | Thanks @jgoizueta ! Confirmed. | 2017-06-02T16:37:35 | 2017-06-02T16:37:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 300,434,069 | 775 | robot-metrika-test | Can one of the admins verify this patch? | 2017-05-10T09:54:06 | 2017-05-10T09:54:06 | {} | NONE |
ClickHouse | ClickHouse | 301,249,616 | 720 | kszucs | Are there any batch processing frameworks already integrated with clickhouse? | 2017-05-13T13:49:40 | 2017-05-13T13:49:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 302,033,278 | 124 | khamin | Hello,
any update on it?
Same problem on 17.04 :) | 2017-05-17T09:14:50 | 2017-05-17T09:14:50 | {} | NONE |
ClickHouse | ClickHouse | 302,141,550 | 124 | proller | We have no builds fot 17.04
But you can build manually as described here:
https://github.com/yandex/ClickHouse/blob/master/doc/build_debian.sh
Or make package by running this script (dont forget install all depends):
https://github.com/yandex/ClickHouse/blob/master/release | 2017-05-17T16:15:03 | 2017-05-17T16:15:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 302,872,326 | 809 | alexey-milovidov | Missing test (in `tests` directory). | 2017-05-20T13:10:02 | 2017-05-20T13:10:02 | {} | MEMBER |
ClickHouse | ClickHouse | 303,081,518 | 818 | mxzwrnz | Had to update clikckhouse to newest version. Works now. | 2017-05-22T12:07:19 | 2017-05-22T12:07:19 | {} | NONE |
ClickHouse | ClickHouse | 303,554,000 | 821 | alexey-milovidov | :+1: | 2017-05-23T22:45:54 | 2017-05-23T22:45:54 | {} | MEMBER |
ClickHouse | ClickHouse | 304,460,034 | 830 | robot-metrika-test | Can one of the admins verify this patch? | 2017-05-27T15:49:01 | 2017-05-27T15:49:01 | {} | NONE |
ClickHouse | ClickHouse | 300,522,764 | 779 | robot-metrika-test | Can one of the admins verify this patch? | 2017-05-10T15:39:02 | 2017-05-10T15:39:02 | {} | NONE |
ClickHouse | ClickHouse | 300,812,181 | 781 | yaroslav-korotaev | ```
ls -l *.mrk
-rw-rw-rw- 1 clickhouse clickhouse 16 май 11 11:12 activation_stamp.mrk
-rw-rw-rw- 1 clickhouse clickhouse 16 май 11 11:12 ad_id.mrk
-rw-rw-rw- 1 clickhouse clickhouse 16 май 11 11:12 age.mrk
-rw-rw-rw- 1 clickhouse clickhouse 16 май 11 11:12 amount.mrk
-rw-rw-rw- 1 clickhouse clickhouse 16 май 11... | 2017-05-11T14:46:12 | 2017-05-11T14:46:12 | {} | NONE |
ClickHouse | ClickHouse | 301,204,151 | 784 | robot-metrika-test | Can one of the admins verify this patch? | 2017-05-12T22:59:04 | 2017-05-12T22:59:04 | {} | NONE |
ClickHouse | ClickHouse | 301,567,689 | 793 | alexey-milovidov | `clickhouse-client` is a symlink to `clickhouse` binary.
You could create symlink with name `clickhouse-client` an run it.
`clickhouse` will decide by program name, what do you want to run.
Also you could run ClickHouse client as `clickhouse --client`. | 2017-05-15T18:46:35 | 2017-05-15T18:46:35 | {} | MEMBER |
ClickHouse | ClickHouse | 303,100,831 | 694 | ludv1x | autotests | 2017-05-22T13:32:09 | 2017-05-22T13:32:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 304,682,652 | 635 | ishirav | +1
This is a bug that should be fixed IMO. It was reported also in the infi.clickhouse_orm project - see **https://github.com/Infinidat/infi.clickhouse_orm/issues/37** | 2017-05-29T15:04:19 | 2017-05-29T15:04:19 | {} | NONE |
ClickHouse | ClickHouse | 302,384,870 | 796 | sergeyignatov | grep -oP 'DB::Exception: Sizes of columns.*' /var/log/clickhouse-server/clickhouse-server.log |sort |uniq -c
59457 DB::Exception: Sizes of columns doesn't match: Path: 0, Value: 1, e.what() = DB::Exception, Stack trace:
| 2017-05-18T12:06:16 | 2017-05-18T12:06:16 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 302,935,378 | 246 | emakarov | получается вот такое; после разлогина опять ошибка:
`:) SELECT * FROM system.settings WHERE name = 'max_threads'
SELECT *
FROM system.settings
WHERE name = 'max_threads'
Received exception from server:
Code: 49. DB::Exception: Received from localhost:9000, 127.0.0.1. DB::Exception: Logical error: zero number ... | 2017-05-21T13:02:03 | 2017-05-21T13:02:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 303,223,897 | 793 | alexey-milovidov | Recently clickhouse-client was separate binary. But it links most of ClickHouse codebase and size of that binary was almost the same as size of clickhouse-server.
Then we needed to add few more binaries - clickhouse-benchmark, clickhouse-local, and decided to make single binary to avoid duplication.
Now we are go... | 2017-05-22T21:23:24 | 2017-05-22T21:23:24 | {} | MEMBER |
ClickHouse | ClickHouse | 304,465,689 | 831 | robot-metrika-test | Can one of the admins verify this patch? | 2017-05-27T17:29:02 | 2017-05-27T17:29:02 | {} | NONE |
ClickHouse | ClickHouse | 300,817,146 | 780 | alexey-milovidov | "source repli**b**a" looks like random bit flip in memory.
Do you have any related messages in `dmesg` on servers with ClickHouse or on servers with ZooKeeper? Do you have ECC memory on your servers?
What is the exact bit representation of what you have pasted as � in ZooKeeper? | 2017-05-11T15:01:42 | 2017-05-11T15:03:10 | {} | MEMBER |
ClickHouse | ClickHouse | 300,884,984 | 780 | alexey-milovidov | It is very unlikely (near impossible), that single bit flips are related to software bug in ClickHouse itself.
You should suspect your hardware (example: memory on network controller).
It is also possible (but very unlikely), that you have accidentially have run zookeeper under some fuzzer, like `zzuf`. | 2017-05-11T18:57:55 | 2017-05-11T18:57:55 | {} | MEMBER |
ClickHouse | ClickHouse | 301,461,243 | 791 | robot-metrika-test | Can one of the admins verify this patch? | 2017-05-15T12:34:03 | 2017-05-15T12:34:03 | {} | NONE |
ClickHouse | ClickHouse | 302,171,611 | 785 | alexey-milovidov | Do I understand correctly, that we don't really need a trie? | 2017-05-17T17:46:50 | 2017-05-17T17:46:50 | {} | MEMBER |
ClickHouse | ClickHouse | 303,071,856 | 346 | alexey-milovidov | It is implemented.
Behaviour is controlled by `join_use_nulls` setting.
If set to 1, NULLs will be substituted instead of default values.
You could adjust this setting like any other - per query, per session, per user... | 2017-05-22T11:18:37 | 2017-05-22T11:18:37 | {
"hooray": 1
} | MEMBER |
ClickHouse | ClickHouse | 303,190,529 | 799 | alexey-milovidov | https://github.com/yandex/ClickHouse/commit/7269f3a23b2e4fb4836ccce81be334cd3b2c3231 | 2017-05-22T19:01:31 | 2017-05-22T19:01:31 | {} | MEMBER |
ClickHouse | ClickHouse | 303,207,522 | 798 | alexey-milovidov | Already in master. | 2017-05-22T20:13:29 | 2017-05-22T20:13:29 | {} | MEMBER |
ClickHouse | ClickHouse | 304,234,207 | 826 | qietingfengling | i fix my problem .the internal_replication must by set "false" | 2017-05-26T09:18:51 | 2017-05-26T09:18:51 | {} | NONE |
ClickHouse | ClickHouse | 304,471,104 | 827 | alexey-milovidov | Yes, this is the issue. `DROP PARTITION` doesn't clear set of blocks for deduplication.
Currently, the only solution, is to clear all nodes in ZooKeeper inside `/blocks/` in table path. | 2017-05-27T19:15:31 | 2017-05-27T19:15:31 | {} | MEMBER |
ClickHouse | ClickHouse | 300,821,098 | 777 | ztlpn | Abort is caused by a double free bug in libzookeeper (haven't investigated deeply). But the root cause seems to be that zetcd does not yet support multi-operations: https://github.com/coreos/zetcd/issues/38
So I guess the takeaway is that zetcd currently cannot be used with ClickHouse. | 2017-05-11T15:14:03 | 2017-05-11T15:14:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 300,882,292 | 780 | SaltTan | I didn't save those records, so I cannot tell you what exactly was there. It definitely looks like corruption. But there are no messages in dmesg, and we have ECC memory in our servers. | 2017-05-11T18:47:52 | 2017-05-11T18:47:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 301,204,149 | 785 | robot-metrika-test | Can one of the admins verify this patch? | 2017-05-12T22:59:02 | 2017-05-12T22:59:02 | {} | NONE |
ClickHouse | ClickHouse | 305,931,145 | 849 | bobrik | Another funny example that I can just fix myself:
* https://clickhouse.yandex/docs/en/interfaces/cli.html
| 2017-06-02T23:27:56 | 2017-06-02T23:27:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 301,877,762 | 800 | alexey-milovidov | Ok. | 2017-05-16T18:46:06 | 2017-05-16T18:46:06 | {} | MEMBER |
ClickHouse | ClickHouse | 301,884,726 | 804 | robot-metrika-test | Can one of the admins verify this patch? | 2017-05-16T19:09:03 | 2017-05-16T19:09:03 | {} | NONE |
ClickHouse | ClickHouse | 301,893,317 | 804 | ludv1x | Looks like `logo.svg` is forgotten:
`file:///home/vludv/ClickHouse/docs/build/html/_static/logo.svg net::ERR_FILE_NOT_FOUND` | 2017-05-16T19:41:43 | 2017-05-16T19:42:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 302,674,738 | 790 | mgilbir | There is this other closed issue that may be relevant here: https://github.com/yandex/ClickHouse/issues/721 | 2017-05-19T11:05:17 | 2017-05-19T11:05:17 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 303,828,067 | 246 | alexey-milovidov | https://github.com/yandex/ClickHouse/commit/2e47c35fa904ec0a78c4087714437fd79dc6b7da
https://github.com/yandex/ClickHouse/commit/0a553100760207d24a0b90f52d05a263520442ec | 2017-05-24T19:32:37 | 2017-05-24T19:32:37 | {} | MEMBER |
ClickHouse | ClickHouse | 304,372,962 | 785 | vavrusa | Yes, that's about expected (see benchmarks in [1], it's also dominated by memory stalls).
I get around 40M rows/s for IPv4 on much smaller dataset. I still didn't get the time for reimplementing poptrie or std::lower_bound comparator, hopefully this weekend. The performance is not that a big deal currently, as I use t... | 2017-05-26T19:47:17 | 2017-05-26T19:47:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 305,628,396 | 842 | robot-metrika-test | Can one of the admins verify this patch? | 2017-06-01T21:39:02 | 2017-06-01T21:39:02 | {} | NONE |
ClickHouse | ClickHouse | 306,013,716 | 847 | alexey-milovidov | https://github.com/BayoNet is going to add the documentation.
> What's the meaning of type column? — that was my initial question, though.
1 - successful start of query execution
2 - successful end of query execution
3 - exception before start of query execution
4 - exception while query execution
Each quer... | 2017-06-04T02:34:50 | 2017-06-04T02:35:12 | {
"+1": 6
} | MEMBER |
ClickHouse | ClickHouse | 303,169,113 | 805 | alexey-milovidov | Wrong documentation. Fixed. | 2017-05-22T17:37:18 | 2017-05-22T17:37:18 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 305,959,513 | 849 | blinkov | Good idea, pulled this into our internal task tracker to later take a look on how this's supposed to be done.
The last screenshot has already been [fixed few days ago](https://github.com/yandex/ClickHouse/commit/dc3a618d7ab506ffd9601571c5a23d8e22be7020), but it still haven't been published because unfortunately this... | 2017-06-03T08:02:53 | 2017-06-03T08:02:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 300,894,099 | 770 | alexey-milovidov | Please, provide results of the following query:
```
SELECT * FROM system.parts WHERE active AND table = '_short'
```
Before and after OPTIMIZE.
You will see, what parts and how many of them was merged.
Also note, that FINAL in SELECT query works in a slightly different way than collapsing while merge.
If f... | 2017-05-11T19:33:20 | 2017-05-11T19:33:20 | {} | MEMBER |
ClickHouse | ClickHouse | 300,931,665 | 770 | magnum-ferox | ```
SELECT count()
FROM tbl
WHERE uid = 168
┌─count()─┐
│ 15299 │
└─────────┘
1 rows in set. Elapsed: 0.001 sec. Processed 24.58 thousand rows, 196.61 KB (22.73 million rows/s., 181.83 MB/s.)
```
```
SELECT count()
FROM tbl
FINAL
WHERE uid = 168
┌─count()─┐
│ 7267 │
└─────────┘
1 rows in... | 2017-05-11T22:14:12 | 2017-05-11T22:14:12 | {} | NONE |
ClickHouse | ClickHouse | 301,776,722 | 800 | robot-metrika-test | Can one of the admins verify this patch? | 2017-05-16T13:09:03 | 2017-05-16T13:09:03 | {} | NONE |
ClickHouse | ClickHouse | 301,827,612 | 802 | robot-metrika-test | Can one of the admins verify this patch? | 2017-05-16T15:54:03 | 2017-05-16T15:54:03 | {} | NONE |
ClickHouse | ClickHouse | 303,039,551 | 346 | mxzwrnz | Is there any news on that? Thanks a lot! | 2017-05-22T08:52:35 | 2017-05-22T08:52:35 | {} | NONE |
ClickHouse | ClickHouse | 301,015,772 | 781 | yaroslav-korotaev | Yes!
Old tables was created manually and has different index rather than new one, which created automatically by our tool.
Thank you so much! | 2017-05-12T08:24:18 | 2017-05-12T08:24:18 | {} | NONE |
ClickHouse | ClickHouse | 301,687,400 | 795 | alexey-milovidov | Thanks! Much better. | 2017-05-16T06:25:56 | 2017-05-16T06:25:56 | {} | MEMBER |
ClickHouse | ClickHouse | 302,094,387 | 796 | ztlpn | Hi @sergeyignatov ! This is a known bug, but we have not been able to pinpoint the cause of it yet. Here is how you can help: please provide all different strings of the form `Path: x, Value: y` that appear in the exception messages (I am interested in the values of x and y and how often they appear). Thanks! | 2017-05-17T13:44:13 | 2017-05-17T13:45:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 304,371,007 | 828 | alexey-milovidov | It was not supposed.
But we could add padding.
PS. Sometimes I specify max width greater than needed,
but default (PrettyCompact) format displays everything aligned by actual maximum,
so lack of padding has its use. | 2017-05-26T19:37:21 | 2017-05-26T19:37:21 | {} | MEMBER |
ClickHouse | ClickHouse | 305,837,193 | 844 | ztlpn | Strange, because this bug was fixed in v1.1.54236 which is the current version at hub.docker.com. Could you check the server version and re-execute `docker pull yandex/clickhouse-server` to be sure? | 2017-06-02T16:22:03 | 2017-06-02T16:22:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 301,432,362 | 789 | ztlpn | Caused by: https://bugs.launchpad.net/cloud-images/+bug/1682622
| 2017-05-15T10:05:14 | 2017-05-15T10:05:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 301,877,740 | 789 | ztlpn | We've pushed fixed images to docker hub under the `latest` tag. @zloidemon Thanks for the PR! | 2017-05-16T18:46:02 | 2017-05-16T18:46:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 302,450,147 | 809 | alexey-milovidov | This is not zero-cost modification. (Anyway, test for performance degradation is missing.)
But for this feature, I expect zero cost in all cases except JSONEachRow format (even more - probably only in UInt8 case).
It means, that `if` should be somewhere in `DataTypeNumber::readJSONText`. | 2017-05-18T16:01:17 | 2017-05-18T16:02:02 | {} | MEMBER |
ClickHouse | ClickHouse | 302,512,944 | 796 | ztlpn | #810 | 2017-05-18T19:10:23 | 2017-05-18T19:10:23 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 305,581,005 | 836 | ztlpn | Agreed, the error is very confusing. | 2017-06-01T18:33:47 | 2017-06-01T18:33:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 305,782,150 | 843 | bgaifullin | Sorry, there is no bug. | 2017-06-02T13:07:36 | 2017-06-02T13:07:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 306,020,722 | 847 | hagen1778 | Are there any way to avoid writing `INSERT` queries into query_log ? | 2017-06-04T06:09:07 | 2017-06-04T06:09:07 | {} | NONE |
ClickHouse | ClickHouse | 301,278,498 | 787 | robot-metrika-test | Can one of the admins verify this patch? | 2017-05-13T22:24:03 | 2017-05-13T22:24:03 | {} | NONE |
ClickHouse | ClickHouse | 301,825,214 | 790 | ludv1x | What is your docker version?
I faced with the same problem using docker `17.04.1~ce-0~ubuntu-trusty`.
But, downgrade to `17.03` helped me to avoid the problem.
So, I guess that it could be docker issue. | 2017-05-16T15:46:22 | 2017-05-16T15:46:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 302,174,249 | 803 | ludv1x | And `system.part_log` also. | 2017-05-17T17:56:32 | 2017-05-17T17:56:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 302,934,918 | 246 | emakarov | have same issue,trying to solve it.
is this setting (max_threads) applies to all users and sessions of clickhouse?
I'm fixing it, but in another session (from python client) still have same error | 2017-05-21T12:52:44 | 2017-05-21T12:52:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 303,197,970 | 815 | proller | maybe you have disabled ipv6.
try edit config: remove <listen_host>::1</listen_host> and leave only <listen_host>127.0.0.1</listen_host> | 2017-05-22T19:33:12 | 2017-05-22T19:33:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 303,205,368 | 785 | alexey-milovidov | > The main reason I started with trie interface is that I want to reuse it for cache-like dictionaries, where resorting after each cache miss would be cost prohibitive.
Ok, got it.
> The benefit of trie is that the search time is always bounded by the length of the key (32 - 128 bits), while binary search is boun... | 2017-05-22T20:04:55 | 2017-05-22T20:04:55 | {} | MEMBER |
ClickHouse | ClickHouse | 303,207,634 | 798 | alexey-milovidov | https://github.com/yandex/ClickHouse/commit/0e77575a76a425cbf04c84203731fa1d9bfbf011 | 2017-05-22T20:13:52 | 2017-05-22T20:13:52 | {} | MEMBER |
ClickHouse | ClickHouse | 304,371,729 | 827 | alexey-milovidov | It happens, when you insert duplicate block - with exactly the same content as before.
Documentation:
> Blocks of data are duplicated. For multiple writes of the same data block (data blocks of the same size containing the same rows in the same order), the block is only written once. The reason for this is in cas... | 2017-05-26T19:40:58 | 2017-05-26T19:41:12 | {} | MEMBER |
ClickHouse | ClickHouse | 304,894,413 | 833 | robot-metrika-test | Can one of the admins verify this patch? | 2017-05-30T14:24:03 | 2017-05-30T14:24:03 | {} | NONE |
ClickHouse | ClickHouse | 300,815,549 | 781 | alexey-milovidov | Is it possible, that you have different primary keys in old MergeTree and new ReplicatedMergeTree table (where you trying to do attach)? | 2017-05-11T14:56:49 | 2017-05-11T14:56:49 | {} | MEMBER |
ClickHouse | ClickHouse | 301,583,839 | 795 | robot-metrika-test | Can one of the admins verify this patch? | 2017-05-15T19:49:03 | 2017-05-15T19:49:03 | {} | NONE |
ClickHouse | ClickHouse | 301,705,379 | 793 | daskol | @alexey-milovidov Wow! This is very unexpectedly and is not obvious. Thanks.
Is it possible to split `clickhouse` binary into `clickhouse-client` and `clickhouse-server`. The reason is sometimes one needs either client or server but both. | 2017-05-16T07:57:44 | 2017-05-16T07:57:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 302,935,264 | 246 | emakarov | cat output of VM (google cloud compute engine instance - g1-small (1 виртуальный ЦП, 1,7 ГБ памяти), Ubuntu 17
cat /proc/cpuinfo
processor : 0
vendor_id : GenuineIntel
cpu family : 6
model : 62
model name : Intel(R) Xeon(R) CPU @ 2.50GHz
stepping : 4
microcode : 0x... | 2017-05-21T12:59:41 | 2017-05-21T12:59:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 303,076,812 | 817 | robot-metrika-test | Can one of the admins verify this patch? | 2017-05-22T11:44:02 | 2017-05-22T11:44:02 | {} | NONE |
ClickHouse | ClickHouse | 304,161,840 | 825 | alexey-milovidov | https://github.com/yandex/ClickHouse/commit/ed44304c49c23d4fa2c0f5d2134a57ebdd47945d
https://github.com/yandex/ClickHouse/commit/fa64d14f26367eb2579a6bcccb7ff85030c55eea | 2017-05-26T00:47:45 | 2017-05-26T00:47:45 | {} | MEMBER |
ClickHouse | ClickHouse | 304,847,114 | 569 | silviucpp | Based on documentation to delete a replica: To delete a replica, run DROP TABLE. However, only one replica is deleted - the one that resides on the server where you run the query.
I don't think there is any need to manage yourself Zookeeper from behind.
Silviu | 2017-05-30T11:13:19 | 2017-05-30T11:13:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 305,203,579 | 829 | hagen1778 | Converting both tables to same structure could be possible workaround for this situation.
It is possible to change column order via editing metadata of table by replacing them in the `.sql` file at `/clickhouseWorkDir/metadata/default/foobar.sql`. CH-server must be restarted after changes.
@alexey-milovidov can th... | 2017-05-31T14:27:12 | 2017-05-31T14:27:12 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 305,898,358 | 835 | ztlpn | Yes, this is a known peculiarity (strings are compared up to the first \0 in some cases) allowed in the name of performance. We should fix it, of course. | 2017-06-02T20:15:57 | 2017-06-02T20:18:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 305,915,902 | 847 | bobrik | What's the meaning of `type` column? — that was my initial question, though. | 2017-06-02T21:41:00 | 2017-06-02T21:41:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 305,916,429 | 848 | robot-metrika-test | Can one of the admins verify this patch? | 2017-06-02T21:44:02 | 2017-06-02T21:44:02 | {} | NONE |
ClickHouse | ClickHouse | 300,434,059 | 778 | robot-metrika-test | Can one of the admins verify this patch? | 2017-05-10T09:54:03 | 2017-05-10T09:54:03 | {} | NONE |
ClickHouse | ClickHouse | 303,211,337 | 785 | vavrusa | > Actually, log(set_size) is also bounded by length of the key, because set cannot have more than 2^length entries.
Indeed. While full key comparison will be more expensive than single bit comparisons, sizes where it will matter (e.g. 2^128) are not practical for this type of in-memory dictionary anyway. Both are go... | 2017-05-22T20:29:19 | 2017-05-22T20:29:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 301,883,005 | 802 | blinkov | It's also included in https://github.com/yandex/ClickHouse/pull/804 | 2017-05-16T19:04:47 | 2017-05-16T19:04:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 303,192,475 | 785 | vavrusa | Hi, sorry for the late reply, I just returned from vacation. The benefit of trie is that the search time is always bounded by the length of the key (32 - 128 bits), while binary search is bounded by the set size. I agree though the basic non-compressing trie will be slower for most datasets (32-128 branches for tries, ... | 2017-05-22T19:09:53 | 2017-05-22T19:09:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 303,897,482 | 815 | qietingfengling | Oh yes,you are right ,thanks,I fix it | 2017-05-25T01:30:10 | 2017-05-25T01:30:10 | {} | NONE |
ClickHouse | ClickHouse | 300,804,672 | 781 | ludv1x | It means that the problem is not related with the issue which broken-part-cutter fixes.
Could you provide output of:
```
ls -l /var/lib/clickhouse//data/dbname/actions/tablename/20100101_20100106_4231042_6296536_113/*.mrk
```
? | 2017-05-11T14:21:55 | 2017-05-11T14:22:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 300,937,751 | 770 | alexey-milovidov | You have just one part and OPTIMIZE does nothing (because it's nothing to merge).
For CollapsingMergeTree to work properly, you should insert different changes separately (so, they will be inserted to different parts).
Also you could run `OPTIMIZE TABLE tbl PARTITION 197001 FINAL` - a special way of OPTIMIZE to re-... | 2017-05-11T22:50:02 | 2017-05-11T22:50:02 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 301,189,412 | 167 | alexey-milovidov | This estimate is outdated :( Now we expect it not sooner than Q4 2017. | 2017-05-12T21:23:32 | 2017-05-12T21:23:32 | {
"confused": 8
} | MEMBER |
ClickHouse | ClickHouse | 301,265,426 | 786 | robot-metrika-test | Can one of the admins verify this patch? | 2017-05-13T18:19:02 | 2017-05-13T18:19:02 | {} | NONE |
ClickHouse | ClickHouse | 301,482,232 | 789 | zloidemon | https://github.com/yandex/ClickHouse/pull/791 | 2017-05-15T13:55:35 | 2017-05-15T13:55:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 302,825,787 | 796 | sergeyignatov | with fix-graphite-rollup Exceptions are gone | 2017-05-19T22:31:17 | 2017-05-19T22:31:17 | {} | NONE |
ClickHouse | ClickHouse | 303,180,129 | 799 | alexey-milovidov | Reasonable. Will do. | 2017-05-22T18:20:56 | 2017-05-22T18:20:56 | {} | MEMBER |
ClickHouse | ClickHouse | 304,367,264 | 785 | alexey-milovidov | We just started to use this dictionary in production.
It solved exactly the task we need today!
Performance is following (for IPv6 addresses, on dictionary that maps IPv6 to AS number; dictionary has 405478 entries):
50 MB/sec per single core (about 3 mln lookups/sec.)
This looks quite slow.
In perf top, funct... | 2017-05-26T19:21:03 | 2017-05-26T19:22:03 | {} | MEMBER |
ClickHouse | ClickHouse | 305,226,992 | 829 | ztlpn | This workaround won't work with Replicated tables. | 2017-05-31T15:38:08 | 2017-05-31T15:38:08 | {} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.