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 | 373,895,428 | 1,981 | derekperkins | I am seeing this too in testing. ClickHouse worked with all the defaults after installing with `apt-get` on Debian 9, but then I saw this error when I changed the location of the logs.
```
<log>/mnt/resource/chdata/clickhouse-server.log</log>
<errorlog>/mnt/resource/chdata/clickhouse-server.err.log... | 2018-03-17T05:17:48 | 2018-03-17T05:17:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 373,918,392 | 2,045 | alexey-milovidov | This is very clever idea!
Probably we can start with encryption/decryption functions.
Special encrypted data types and syntax in table definition may go later.
Compression rate will suffer. But not so much if we just encrypt only few certain fields out of many.
CPU cost will be not so much if we will be able to... | 2018-03-17T13:01:20 | 2018-03-17T13:01:20 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 373,934,010 | 838 | zhang2014 | What is the current state of this issue? I plan to implement it. @alexey-milovidov | 2018-03-17T16:39:19 | 2018-03-17T16:39:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 373,938,562 | 1,972 | alexey-milovidov | This is because Distributed table is pointed at itself.
I've added a check for this in recent ClickHouse relese (1.1.54370).
> by the way: if crash select from client, it can restart. but it start from default config.xml. i want know how to change this ，i use different config.xml dir.
The server is restarted by ... | 2018-03-17T17:38:34 | 2018-03-17T17:38:34 | {} | MEMBER |
ClickHouse | ClickHouse | 374,283,452 | 2,015 | alexey-milovidov | Basic idea of implementation:
Look at expression in WHERE and PREWHERE.
(This should be done after expansion of aliases.)
It can be an AND of multiple elements or not. In latter case it is considered as single element.
In first case, you should consider cases when you have single AND function
`and(a, b, c...)`... | 2018-03-19T16:50:30 | 2018-03-19T16:55:31 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 374,238,254 | 2,045 | filimonov | To utilize AES-NI instruction the only algorithm of choice is AES (in one of the modes).
I think that CTR mode of AES should be nice - reasonable good, fast, work in stream mode,
no padding needed (implementations should support SIMD, probably should support parallelization).
For algorithm itself you need to pr... | 2018-03-19T14:47:56 | 2018-03-19T17:34:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 374,620,255 | 2,020 | StarWix | @ludv1x image with latest tag was built only. Other images weren't built, eg: https://hub.docker.com/r/yandex/clickhouse-server/builds/b5uvodq9uwlkdshpeuhi2ka/ | 2018-03-20T14:34:58 | 2018-03-20T14:34:58 | {} | NONE |
ClickHouse | ClickHouse | 375,710,510 | 2,076 | alexey-milovidov | Thanks!
I've additionally tested it with a cluster with two shards on different ports.
The difference is that when your shard is local, the query is processed without network communication, and the issue could be not visible, as in that case query is not formatted/parsed.
| 2018-03-23T15:49:53 | 2018-03-23T15:49:53 | {} | MEMBER |
ClickHouse | ClickHouse | 375,827,503 | 2,070 | jonatasfreitasv | I have the same error. | 2018-03-23T23:54:53 | 2018-03-23T23:54:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 375,848,863 | 2,096 | sundy-li | yeah，I am the author of that article | 2018-03-24T05:41:13 | 2018-03-24T05:41:13 | {
"laugh": 3
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 373,742,592 | 2,067 | kshvakov | `max_parallel_replicas` works only for the tables created with [`SAMPLE`](https://clickhouse.yandex/docs/en/single/#sample-clause) clause. ClickHouse server splits the query by sample section and executes part of the query on N (max_parallel_replicas) replicas. | 2018-03-16T15:08:06 | 2018-03-16T15:08:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 373,933,584 | 2,015 | zhang2014 | I don't know how to write test for this `Pull Request` | 2018-03-17T16:33:32 | 2018-03-17T16:33:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 374,588,549 | 208 | filimonov | I'm wondering - what if some special syntax can appear to allow to monitor running queries 'from outside'. Something like MONITOR in redis.
So you can just open second Clickhouse connection, send some command like ```monitor query where query_id = 123123``` and start getting info like:
timestamp,status,read_rows,r... | 2018-03-20T13:00:31 | 2018-03-20T13:00:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 374,740,221 | 2,084 | rajadayalan | Yes, found "Converting aggregation data to two-level" in the logs.
2018.03.19 22:01:42.510429 [ 637 ] <Trace> Aggregator: Aggregating
2018.03.19 22:01:42.510490 [ 637 ] <Trace> Aggregator: Aggregation method: concat
2018.03.19 22:01:42.588868 [ 637 ] <Trace> Aggregator: Converting aggregation data to two-level.
2... | 2018-03-20T20:12:14 | 2018-03-20T20:12:14 | {} | NONE |
ClickHouse | ClickHouse | 374,914,924 | 2,086 | filimonov | If i understand it correctly:
1) With one insert you create one part.
2) Duplicates are removed at the moment of MERGING different parts
3) OPTIMIZE forces merge process (you already have one part so it do nothing)
4) FINAL apply the same logic as merge process (so replace old version with newer from different pa... | 2018-03-21T12:05:11 | 2018-03-21T12:06:00 | {
"+1": 2,
"heart": 2,
"hooray": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 375,339,663 | 2,087 | QSer29 | Да, это так. Спасибо. | 2018-03-22T15:06:05 | 2018-03-22T15:06:05 | {} | NONE |
ClickHouse | ClickHouse | 375,353,579 | 1,722 | silviucpp | @ztlpn Now the branch has no longer conflicts with master. | 2018-03-22T15:43:17 | 2018-03-22T15:43:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 375,594,027 | 2,095 | sundy-li | https://github.com/yandex/ClickHouse/blob/62492450180fa37d018a0f454066765b125580b3/dbms/src/Storages/StorageKafka.cpp#L579
I just found that, topic is just single str value, not a list. | 2018-03-23T09:33:56 | 2018-03-23T09:33:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 375,826,486 | 1,282 | derekperkins | Is this now fixed? | 2018-03-23T23:46:16 | 2018-03-23T23:46:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 375,848,816 | 2,096 | lamberken | @sundy-li,
here are some examples use udf, can just use clickhouse's native grammar to realize the funnel query with time window? I have no idea.
- https://www.jianshu.com/p/4c86a2478cca
- https://github.com/analysys/olap
| 2018-03-24T05:39:38 | 2018-03-24T05:40:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 375,852,857 | 2,096 | filimonov | > Code: 184. DB::Exception: Received from hadoop:9000, DB::Exception: Aggregate function minIf(createtime, equals(pageid, 2538)) is found inside another aggregate function in query.
Hmmm... Sounds like a bug in ClickHouse. minArray is not 'fair' aggregate function. So it rather should be safe to use other 'real' agg... | 2018-03-24T07:20:37 | 2018-03-24T08:07:33 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 375,855,067 | 2,096 | lamberken | @filimonov，thanks for your help，I use `arrayReduce` also can realize it.
```
SELECT
countIf(1, step1_time != toDateTime(0)) AS home,
countIf(1, step2_time != toDateTime(0)) AS list,
countIf(1, step3_time != toDateTime(0)) AS detail
FROM
(
SELECT
loginkey,
minIf(createtime... | 2018-03-24T08:08:54 | 2018-03-24T08:08:54 | {
"+1": 3,
"laugh": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 374,598,380 | 208 | ztlpn | There is the `system.processes` table that contains this information. You won't be able to receive updates in the streaming fashion though, you would have to poll. | 2018-03-20T13:32:01 | 2018-03-20T13:32:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 374,732,299 | 2,084 | ztlpn | There is a known bug that under certain conditions inserts into aggregating materialized views generate x256 number of original inserts. And this then leads to "too many parts" exception.
So let us first check if this bug is the culprit or not. To do it first track the failed INSERT in your server logs. It will look... | 2018-03-20T19:46:23 | 2018-03-20T19:46:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 375,394,748 | 1,956 | Amper | @daledude, look at [this pull request](https://github.com/yandex/ClickHouse/pull/2089), will it help you?
| 2018-03-22T17:38:05 | 2018-03-22T17:39:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 375,680,382 | 2,100 | tobad357 | This fixes #2095 | 2018-03-23T14:20:40 | 2018-03-23T14:20:40 | {} | NONE |
ClickHouse | ClickHouse | 375,745,396 | 2,098 | rpuch | Let's see how it looks case by case.
1. Someone makes a request from outside and that request goes through a reverse proxy, the reverse proxy will not let the existing X-Forwarded-For header through and instead it will replace it with a valid header. This case is ok.
2. Someone makes a request from outside, but tha... | 2018-03-23T17:39:27 | 2018-03-23T17:39:27 | {} | NONE |
ClickHouse | ClickHouse | 375,852,752 | 2,096 | lamberken | @filimonov, I see, you are too responsible, thank you very match.
But it can't run through, if use `minArray`
```
SELECT
countIf(1, step1_time != toDateTime(0)) AS home,
countIf(1, step2_time != toDateTime(0)) AS list
FROM
(
SELECT
loginkey,
minIf(createtime, pageid = 2538) A... | 2018-03-24T07:18:14 | 2018-03-24T07:18:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 374,030,500 | 838 | otisg | Nice. I know this will vary from situation to situation, but have you done any work to see how savings one could expect from this compression and what the price of that might be in terms of increased CPU usage? | 2018-03-18T18:39:12 | 2018-03-18T18:39:12 | {} | NONE |
ClickHouse | ClickHouse | 374,105,696 | 1,815 | lamberken | :+1: | 2018-03-19T05:32:20 | 2018-03-19T05:32:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 374,804,938 | 2,078 | lamberken | :+1: | 2018-03-21T00:44:07 | 2018-03-21T00:44:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 375,847,764 | 1,956 | daledude | @Amper Thank you! I missed the crucial "see ... for list of other options" so I didn't realize I could use any of the librdkafka options. | 2018-03-24T05:13:56 | 2018-03-24T05:13:56 | {} | NONE |
ClickHouse | ClickHouse | 375,874,579 | 2,084 | ztlpn | You should count active parts only. Insertion stops when the number of active parts in one partition exceeds 300 (and generally if this number is greater than 100 it is a cause for concern). You can track it with the `MaxPartCountForPartition` metric (either in the `system.asynchronous_metrics` table or in graphite if ... | 2018-03-24T11:19:28 | 2018-03-24T11:20:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 373,553,138 | 2,012 | alexey-milovidov | My algorithm will eat more memory if there are many intervals that start/end at the same second.
It's expected, for example, if we accumulating data for one day into single aggregation state, and we have more than about 100 000 intervals per day. That's quite natural.
We can simply improve it later by adding "compr... | 2018-03-15T23:18:51 | 2018-03-15T23:19:24 | {} | MEMBER |
ClickHouse | ClickHouse | 373,817,692 | 1,722 | silviucpp | Hello @ztlpn ,
I'll try to rebase with last master but there are conflicts in several places which are not straight forward to fix. I first need to review what changes were done in the meantime and then work on this to make sure I won't break something.
Last time when I tried I failed :) maybe I was to tired. I... | 2018-03-16T19:16:20 | 2018-03-16T19:16:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 374,088,715 | 1,389 | linmounong | +1 for this feature | 2018-03-19T03:09:01 | 2018-03-19T03:09:01 | {} | NONE |
ClickHouse | ClickHouse | 374,922,624 | 2,083 | ztlpn | Confirmed. | 2018-03-21T12:38:52 | 2018-03-21T12:38:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 375,445,612 | 2,091 | alexey-milovidov | The reason is that `if(is_attributed is null, -1, is_attributed)` has Nullable type.
You can check it with `toTypeName(if(is_attributed is null, -1, is_attributed))`.
You can replace it with `coalesce(is_attributed, -1)` to get non-Nullable type.
The first byte in binary representation of Nullable is "is NULL" f... | 2018-03-22T20:25:36 | 2018-03-22T20:25:43 | {} | MEMBER |
ClickHouse | ClickHouse | 375,480,259 | 2,026 | alexey-milovidov | It's unfinished. This feature has low/medium priority. | 2018-03-22T22:37:26 | 2018-03-22T22:37:26 | {} | MEMBER |
ClickHouse | ClickHouse | 375,738,669 | 1,959 | alexey-milovidov | This is a copy paste:
```
std::map<String, String> environment =
{
{"host", config().getString("host", "localhost")},
{"port", config().getString("port", "9000")},
{"user", config().getString("user", "default")},
{"display_name", ser... | 2018-03-23T17:18:06 | 2018-03-23T17:18:06 | {} | MEMBER |
ClickHouse | ClickHouse | 375,774,611 | 2,098 | rpuch | Thank you! I'll take a look at it | 2018-03-23T19:28:47 | 2018-03-23T19:28:47 | {} | NONE |
ClickHouse | ClickHouse | 375,730,945 | 2,096 | filimonov | Probably you've already seen that video https://www.youtube.com/watch?v=YpurT78U2qA ?
Presentation is here: https://events.yandex.com/lib/talks/247/
Your funnel is not too complicated, you can avoid sorting here, just try to play a bit with array operators, and aggregate function modifiers, for example something li... | 2018-03-23T16:53:29 | 2018-03-23T19:46:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 373,744,196 | 2,067 | kshvakov | From official help:
```sql
SELECT *
FROM system.settings
WHERE name = 'max_parallel_replicas'
Row 1:
──────
name: max_parallel_replicas
value: 1
changed: 0
description: The maximum number of replicas of each shard used when the query is executed. For consistency (to get different parts... | 2018-03-16T15:13:01 | 2018-03-16T15:13:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 374,887,859 | 2,043 | ztlpn | Oops, sorry :) Didn't notice your PR... | 2018-03-21T10:11:04 | 2018-03-21T10:11:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 375,624,579 | 2,091 | mephistopheies | thanks! | 2018-03-23T11:04:53 | 2018-03-23T11:04:53 | {} | NONE |
ClickHouse | ClickHouse | 375,842,247 | 2,076 | zhang2014 | Yes, I tested it locally, but I'm not sure if I want to rename the test to `00609_shard_with_case_when_then.sql` | 2018-03-24T03:07:14 | 2018-03-24T03:07:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 375,850,133 | 2,100 | tobad357 | @sundy-li sorry about that, serves me right for changing variable name after testing late at night... :)
I've pushed an update that I've tested locally against two different topics and with single topic as well | 2018-03-24T06:15:30 | 2018-03-24T06:15:30 | {} | NONE |
ClickHouse | ClickHouse | 373,982,702 | 2,071 | alexey-milovidov | We have fixed it in 1.1.54370. | 2018-03-18T09:01:36 | 2018-03-18T09:01:36 | {
"+1": 3
} | MEMBER |
ClickHouse | ClickHouse | 374,573,325 | 2,008 | ImreSamu | maybe a temporary workaround from carto.com ( based on "quadtree keys", or "quadkeys" )
* "Geospatial processing with Clickhouse" ( May 23, 2017 )
* https://carto.com/blog/inside/geospatial-processing-with-clickhouse/
* "Geospatial processing with Clickhouse, take 2" ( Aug 9, 2017 )
* https://carto... | 2018-03-20T12:03:57 | 2018-03-20T12:03:57 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 374,618,370 | 2,082 | proller | Already fixed in master | 2018-03-20T14:29:55 | 2018-03-20T14:29:55 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 375,760,210 | 1,956 | daledude | @Amper, thanks for the pointer but I don't see anything related to SASL PLAIN or I missed something? | 2018-03-23T18:30:49 | 2018-03-23T18:30:49 | {} | NONE |
ClickHouse | ClickHouse | 373,828,438 | 1,722 | silviucpp | @ztlpn How many threads are created per each insert and select query ? It's a fixed number or depends from query to query ? | 2018-03-16T19:59:32 | 2018-03-16T19:59:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 374,292,806 | 2,075 | ztlpn | Confirmed. It is a deadlock during the shutdown of Kafka tables if they didn't have a chance to start up. | 2018-03-19T17:15:08 | 2018-03-19T17:15:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 374,613,677 | 208 | filimonov | > There is the system.processes table that contains this information.
> You won't be able to receive updates in the streaming fashion though, you would have to poll.
I know that and actually i do that. :) But that doesn't contain final stats - like number of rows extracted, final memory used, etc. That data can be... | 2018-03-20T14:16:49 | 2018-03-20T14:16:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 373,575,463 | 2,015 | zhang2014 | @KochetovNicolai @alexey-milovidov Please give me some advice about this issue. | 2018-03-16T01:33:40 | 2018-03-16T01:40:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 375,226,719 | 1,323 | silviucpp | I'm not able to replicate this any longer. It's true that since then lot of changes happened in clickhouse. | 2018-03-22T09:09:27 | 2018-03-22T09:09:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 375,274,076 | 2,087 | ztlpn | Это тот же баг, если websites_21 это materialized view, а site_id в первичном ключе. @QSer29 это так? | 2018-03-22T11:37:32 | 2018-03-22T11:37:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 375,802,390 | 1,697 | alexey-milovidov | It's not expected. | 2018-03-23T21:24:25 | 2018-03-23T21:24:25 | {} | MEMBER |
ClickHouse | ClickHouse | 375,849,179 | 2,096 | lamberken | :+1: , filimonov's answer is enough for me right now, but I also want to realize the window funnel.
the udf can't use in new version, so it would be better if clickhouse can support use native api to realize it | 2018-03-24T05:49:20 | 2018-03-24T05:49:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 373,814,097 | 1,722 | ztlpn | Please rebase the branch or merge the latest stable into it so that the changes are easily testable. | 2018-03-16T19:02:34 | 2018-03-16T19:02:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 373,845,830 | 1,722 | ztlpn | > How many threads are created per each insert and select query ? It's a fixed number or depends from query to query ?
Depends on the query and settings. Typically, to execute a SELECT query ClickHouse will create around max_threads new threads that will then immediately exit. But if it is a Distributed query, a thr... | 2018-03-16T21:09:57 | 2018-03-16T21:09:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 373,870,198 | 2,066 | harlinb | I was able to reproduce this bug.
- Start with a MergeTree table that has a Nested Array similar to above
- Continue to insert data until new parts is merged at least to level 3
- Merges start to remove the size1.mrk file from the Nested field
- Restart clickhouse
- Clickhouse complains that the size1.mrk doesn't ... | 2018-03-16T23:20:39 | 2018-03-20T18:02:49 | {} | NONE |
ClickHouse | ClickHouse | 374,828,323 | 2,043 | kshvakov | duplicate 86d53b34a55226715edb592070dd6ca52a83a9a3 | 2018-03-21T03:57:26 | 2018-03-21T03:57:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 375,450,732 | 2,026 | silviucpp | @alexey-milovidov I see you worked on this .. is this bug a priority ? | 2018-03-22T20:42:23 | 2018-03-22T20:42:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 375,836,639 | 2,084 | rajadayalan | We are noticing the parts count for ReplicatedAggregatingMergeTree also same around 200. We use custom partition, so we think the nature of data at that time caused more parts. . | 2018-03-24T01:28:38 | 2018-03-24T01:29:25 | {} | NONE |
ClickHouse | ClickHouse | 375,850,114 | 2,096 | filimonov | You have chaged minArray to just first element of result array for step_2 time. That can give wrong result, because clickhouse can find some event which happened after step3, so step3 will not be foubd | 2018-03-24T06:14:50 | 2018-03-24T06:14:50 | {
"eyes": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 373,981,975 | 2,072 | proller | fix: https://github.com/yandex/ClickHouse/pull/2068/files#diff-0393a0d8064ceb32924635d04ebf4e72R377 | 2018-03-18T08:47:19 | 2018-03-18T08:47:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 373,983,483 | 838 | Sindbag | @zhang2014 I am doing the task, plan to pull request soon. | 2018-03-18T09:18:26 | 2018-03-18T09:18:26 | {} | NONE |
ClickHouse | ClickHouse | 374,763,801 | 2,084 | ztlpn | Well, we better fix this bug then :)
Meanwhile, here is a workaround. Two-level aggregation is controlled by two settings: `group_by_two_level_threshold` and `group_by_two_level_threshold_bytes` - if an aggregated result size (in rows or bytes correspondingly) exceeds any of these thresholds, it is turned on. Your g... | 2018-03-20T21:31:03 | 2018-03-20T21:31:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 375,447,470 | 1,722 | silviucpp | @alexey-milovidov @ztlpn : Excepting 2 points where I added some comments because is not clear enough for me how you except to implement them all should be fixed and conflict free with master. Once you clarify the other 2 I'll fix them and run a load test and some functionality tests again. | 2018-03-22T20:31:29 | 2018-03-22T20:31:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 375,662,616 | 2,098 | rpuch | I've used a simple echo server https://gist.github.com/solusipse/6419144 to make sure X-Forwarded-For is actually passed. For the following request
curl http://clickhouse.mydomain/aaa
I got the following:
GET /aaa HTTP/1.0
Host: clickhouse.mydomain
X-Real-IP: <my.ip>
X-Forwarded-For: <my.ip>
Connection: cl... | 2018-03-23T13:20:49 | 2018-03-23T13:20:49 | {} | NONE |
ClickHouse | ClickHouse | 375,850,148 | 2,100 | tobad357 | The build failure doesn't look to be related to my change | 2018-03-24T06:15:54 | 2018-03-24T06:15:54 | {} | NONE |
ClickHouse | ClickHouse | 375,852,147 | 2,103 | tobad357 | I think the issue is that you are trying to create a materialized view on top of a distributed table
try
```CREATE MATERIALIZED VIEW activations_view
TO activations
AS SELECT * FROM activations_kafka
```
This will then put data into the local table on each node that you have the Kafka table on
When you wan... | 2018-03-24T07:05:35 | 2018-03-24T07:05:35 | {} | NONE |
ClickHouse | ClickHouse | 374,448,076 | 2,067 | lamberken | Thank you! | 2018-03-20T02:08:23 | 2018-03-20T02:08:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 374,927,027 | 2,086 | StarWix | @filimonov
> If you specify FINAL, optimization will be performed even when all the data is already in one part.
https://clickhouse.yandex/docs/en/single/#optimize | 2018-03-21T12:56:09 | 2018-03-21T12:56:09 | {
"+1": 2,
"hooray": 1
} | NONE |
ClickHouse | ClickHouse | 374,584,848 | 2,078 | ztlpn | Max length of the array stored in the groupArray state is limited to 16777215. This arbitrary limit is checked only on deserialization of a state and acts as a guard against corrupted data. | 2018-03-20T12:47:42 | 2018-03-20T12:47:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 375,046,982 | 2,084 | rajadayalan | Thank You Alex.
Changing the values for the settings "group_by_two_level_threshold" and "group_by_two_level_threshold_bytes" helped. Dont see the "Converting aggregation data to two-level" any more. It looks good :).
Total number of parts(including the inactive parts) with ReplicatedAggregatingMergeTree for the ... | 2018-03-21T18:21:07 | 2018-03-21T18:21:07 | {} | NONE |
ClickHouse | ClickHouse | 375,291,029 | 1,566 | valyala | FYI, I tried implementing `O_DIRECT` support for another project and failed to do so, since it has been appeared that certain filesystems and/or configurations don't support `O_DIRECT`. Additionally, `O_DIRECT` usually implies synchronous I/O, which may lead to slow read/write operations on network filesystems with hig... | 2018-03-22T12:44:41 | 2018-03-22T12:44:41 | {} | NONE |
ClickHouse | ClickHouse | 375,707,377 | 2,089 | alexey-milovidov | Thank you! | 2018-03-23T15:40:31 | 2018-03-23T15:40:31 | {} | MEMBER |
ClickHouse | ClickHouse | 375,769,021 | 2,098 | filimonov | > It seems that this would reduce the flexibility of access control. I would like to have two users to only be able to connect from 2 IPs, one for each. Would it be possible?
Of course. If you will use Basic HTTP Authentication then you have a lot of configuration options.
Also, please take a look on that projec... | 2018-03-23T19:05:20 | 2018-03-23T19:08:06 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 375,851,765 | 2,096 | filimonov | With timeframes, once i've tryed, but it was a scary looking query with about 100 lines which worked for one particular case. Generally it was about spliting one chain of user events to multiple chains when some specific event appear so that was not fair time framing but event-based framing. Not sure if can be generali... | 2018-03-24T06:57:32 | 2018-03-24T06:59:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 373,506,774 | 1,323 | alexey-milovidov | Cannot reproduce on another query:
```
SELECT NULL as n, count() FROM visits_all WHERE CounterID = 34 AND n = 1 FORMAT JSONCompact
``` | 2018-03-15T20:09:06 | 2018-03-15T20:09:06 | {} | MEMBER |
ClickHouse | ClickHouse | 374,108,697 | 2,072 | millin | The same issue occurs during build v1.1.54370 on Centos 7. | 2018-03-19T05:55:31 | 2018-03-19T05:55:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 374,292,354 | 2,077 | ztlpn | No, currently capnproto format doesn't support deserialization of packed messages. | 2018-03-19T17:13:51 | 2018-03-19T17:13:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 375,723,679 | 1,959 | alexey-milovidov | Ok, but I don't like
```
localhost :)
```
by default... I'll modify a little. | 2018-03-23T16:30:39 | 2018-03-23T16:30:53 | {} | MEMBER |
ClickHouse | ClickHouse | 375,845,509 | 2,096 | lamberken | @filimonov :+1: :+1: :+1: it's very useful for me, thank you very match
```
SELECT
countIf(1,step1_time != toDateTime(0)) AS home,
countIf(1,step2_time != toDateTime(0)) AS list,
countIf(1,step3_time != toDateTime(0)) AS detail
FROM
(
SELECT
loginkey,
minIf(createti... | 2018-03-24T04:17:09 | 2018-03-24T04:17:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 375,847,563 | 2,096 | lamberken | @filimonov, hi
Do you have some complicated funnel examples?
for exapmle, the funnel with `time window`?
| 2018-03-24T05:09:05 | 2018-03-24T05:09:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 375,848,032 | 2,096 | sundy-li | @Jaxvk could u publish your dataset? In last year we solved the same window funnel problem. Maybe now I can do some help with it. | 2018-03-24T05:20:42 | 2018-03-24T05:20:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 373,791,592 | 1,722 | ztlpn | > For example, if you run quite loaded server under ASan, it will complain about "thread limit exceeded" after few days of run. And your modification will fix it.
In fact this patch will not help because the root cause is that several threads are created for executing each query, not the threads doing replication (t... | 2018-03-16T17:44:16 | 2018-03-16T17:44:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 373,983,141 | 1,562 | ilyshnikova | I'll take this issue | 2018-03-18T09:11:10 | 2018-03-18T09:11:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 374,560,868 | 2,077 | tobad357 | Will revert to non packed form
Would be great to have packed format as an option in the future | 2018-03-20T11:15:41 | 2018-03-20T11:15:41 | {} | NONE |
ClickHouse | ClickHouse | 375,712,106 | 2,098 | filimonov | X-Forwarded-For probably can be added to some stats, but using it during authentication sounds highly unsafe.
If Clickhouse will start to trust X-Forwarded-For header (which can be easily added to any request), it will give a lot of options to abuse that.
So for example you will connect from every ip, and will ... | 2018-03-23T15:55:06 | 2018-03-23T15:55:26 | {
"+1": 2,
"-1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 374,034,902 | 838 | Sindbag | @otisg well, currently I did not precisely benchmark the performance, but it is planned and, I expect, will be packed with other tests. | 2018-03-18T19:09:55 | 2018-03-18T19:09:55 | {} | NONE |
ClickHouse | ClickHouse | 374,115,071 | 2,061 | alniks | The problem may be not in ALIAS column. Dropping it did not help:
`SELECT domain_name
FROM events
WHERE (event_date >= '2017-08-01') AND (event_date <= '2017-08-31') AND (user_id = 1) AND (domain_name = 'gmail.com')
Received exception from server (version 1.1.54362):
Code: 9. DB::Exception: Received from local... | 2018-03-19T06:40:33 | 2018-03-19T06:40:33 | {} | NONE |
ClickHouse | ClickHouse | 374,306,055 | 2,075 | ztlpn | fixed in master | 2018-03-19T17:53:49 | 2018-03-19T17:53:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 374,955,754 | 2,087 | alex-zaitsev | Could be related to: https://github.com/yandex/ClickHouse/commit/590feb02b5907aad4715227e64efc4b773c2a72f | 2018-03-21T14:26:47 | 2018-03-21T14:26:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 375,725,879 | 1,959 | alexey-milovidov | ```
std::map<String, String> terminalCharacters = {
{ "\\e[", "\e[" },
{ "\\33[", "\33[" },
{ "\\033[", "\033["},
{ "\\x1B[", "\x1B["},
};
```
That's weird, because:
- it's partially duplicated by `readEscapedString` (but `readEscapedString` doesn't... | 2018-03-23T16:37:34 | 2018-03-23T16:37:34 | {} | MEMBER |
ClickHouse | ClickHouse | 375,728,276 | 1,959 | alexey-milovidov | I would like to:
1. Add support for `\e` to `readEscapedString`.
2. Use `readEscapedString` for parsing prompt.
3. Don't support octal escape sequences. | 2018-03-23T16:44:46 | 2018-03-23T16:44:46 | {} | MEMBER |
ClickHouse | ClickHouse | 375,843,938 | 1,956 | Amper | @daledude, section "Configuration":
<img width="1797" alt="kafka" src="https://user-images.githubusercontent.com/495795/37860018-0bab2ea2-2f50-11e8-99ce-3bca54209bd6.png">
| 2018-03-24T03:43:25 | 2018-03-24T03:43:25 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 375,848,498 | 2,096 | lamberken | @sundy-li, thank you.
yes, here's the table structure and some datas.
```
CREATE TABLE trajectory_trackless_complex(
loginkey UInt64,
provinceid Int32,
cityid Int32,
browser String,
pageid UInt32,
visittime DateTime,
productcode UInt32,
elepath String,
createtime DateTim... | 2018-03-24T05:31:52 | 2018-03-24T05:31:52 | {} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.