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 | 569,174,707 | 8,370 | zhang2014 | > BTW, could you please share your use case - where it is needed for you?
We have some users who are not very proficient with ClickHouse, and some of their code depend `DESC TABLE` result
> Suspicious places in code:
Sorry, I only focused on the test result and missed them. I will look into the code immediatel... | 2019-12-27T03:06:37 | 2019-12-27T03:06:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,323,919 | 6,725 | alexey-milovidov | Deadlock on TRUNCATE of Replicated tables was fixed in #6695 version 19.14.3.3. | 2019-12-27T18:42:41 | 2019-12-27T18:42:58 | {} | MEMBER |
ClickHouse | ClickHouse | 569,344,908 | 8,441 | den-crane | not sure how to do it.
perf record -F 99 -p $(pidof clickhouse-server) -g -- sleep 30
[perf_data.zip](https://github.com/ClickHouse/ClickHouse/files/4006242/perf_data.zip)
perf diff perf.data.19.13.4.32 perf.data.19.13.5.44
```
[bridge] with build id dccadf6fa578c2aae846a7cedfc95c00ba258abb not found, cont... | 2019-12-27T21:00:54 | 2019-12-27T21:00:54 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,481,874 | 8,447 | amosbird | Cannot reproduce the heavy regressions locally.
1. bounding_ratio.xml
2. system_numbers.xml
3. select_format.xml ( added another pr to achieve more reliable test result without disk io https://github.com/ClickHouse/ClickHouse/pull/8455) | 2019-12-29T07:23:38 | 2019-12-29T07:23:38 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 569,508,192 | 8,455 | alexey-milovidov | Ok. | 2019-12-29T13:59:21 | 2019-12-29T13:59:21 | {} | MEMBER |
ClickHouse | ClickHouse | 569,080,084 | 8,394 | BayoNet | I have a global task on the configuration part of documentation CLICKHOUSEDOCS-292. I'll do it there. | 2019-12-26T15:55:28 | 2019-12-26T15:55:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,083,947 | 6,767 | 4ertus2 | One more place there we should care about push down enabled #8417 | 2019-12-26T16:20:31 | 2019-12-26T16:20:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,194,307 | 8,159 | weiqxu | @gubinjie Your uri seems has a space between "hdfs:" and "//ns/hive/". | 2019-12-27T05:45:02 | 2019-12-27T05:45:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,204,867 | 8,159 | gubinjie | @weiqxu
这应该是复制的问题，实际情况中不存在空格的。我看了一下libhdfs3的源码，的确是找LIBHDFS3_CONF的地址，但是我在hdfs-client.xml文件中配置了dfs.ha.namenodes.ns，但是貌似找不到这个hdfs-client.xml文件，所以是否遗漏了什么？还是我现在用的clickhouse版本是19.16的版本。
能否加个微信：16621128482，
`class DefaultConfig {
public:
DefaultConfig() : conf(new Hdfs::Config) {
bool reportError = false;
... | 2019-12-27T06:52:27 | 2019-12-27T06:52:27 | {} | NONE |
ClickHouse | ClickHouse | 569,283,107 | 8,369 | ghuname | ...but I have just found a new "feature". Please look at this:
```
SELECT
nullIf(nan, nan),
0 / 0,
nullIf(0 / 0, nan),
nullIf(nan, 0 / 0)
┌─nullIf(nan, nan)─┬─divide(0, 0)─┬─nullIf(divide(0, 0), nan)─┬─nullIf(nan, divide(0, 0))─┐
│ ᴺᵁᴸᴸ │ nan │ ... | 2019-12-27T14:42:01 | 2019-12-27T14:42:01 | {
"+1": 2
} | NONE |
ClickHouse | ClickHouse | 569,305,981 | 8,429 | alexey-milovidov | > because the client might not know the alias of types
It's about all the clients that use Native format?
> We should probably add a setting to keep it compatible?
Maybe we can do it in a different way:
- do not use custom names inside DataType objects; use canonical names as before;
- save user-specified na... | 2019-12-27T16:54:11 | 2019-12-27T16:54:11 | {} | MEMBER |
ClickHouse | ClickHouse | 569,502,985 | 8,219 | alexey-milovidov | ```
~/work/ClickHouse$ find . -name 'CMakeLists.txt' -or -name '*.cmake' | xargs grep 'unistring'
~/work/ClickHouse$
```
We don't depend on unistring, only on icu18n, icuuc, icudata.
And we provide our own CMakeLists that are rather clean.
And also we check for dependencies in CI... | 2019-12-29T12:56:25 | 2019-12-29T12:56:47 | {} | MEMBER |
ClickHouse | ClickHouse | 569,324,595 | 8,437 | den-crane | ```
SELECT
nan = nan,
0 / 0,
nan = (0 / 0),
isNaN(0 / 0)
┌─equals(nan, nan)─┬─divide(0, 0)─┬─equals(nan, divide(0, 0))─┬─isNaN(divide(0, 0))─┐
│ 1 │ nan │ 0 │ 1 │
└──────────────────┴──────────────┴───────────────────────────┴──... | 2019-12-27T18:47:11 | 2019-12-27T18:50:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,329,118 | 8,437 | alexey-milovidov | ```
SELECT
0 AS a,
0 + 0 AS b,
a = b,
a / a AS n1,
b / b AS n2,
n1 = n2,
nan = nan
┌─a─┬─b─┬─equals(0, plus(0, 0))─┬──n1─┬──n2─┬─equals(divide(0, 0), divide(plus(0, 0), plus(0, 0)))─┬─equals(nan, nan)─┐
│ 0 │ 0 │ 1 │ nan │ nan │ ... | 2019-12-27T19:15:28 | 2019-12-27T19:19:53 | {} | MEMBER |
ClickHouse | ClickHouse | 569,330,391 | 8,441 | alexey-milovidov | Is it attributed to https://github.com/ClickHouse/ClickHouse/pull/4454 ? | 2019-12-27T19:24:07 | 2019-12-27T19:24:07 | {} | MEMBER |
ClickHouse | ClickHouse | 569,471,127 | 8,452 | wangzhen11aaa | > Please check and provide an output of `select * from system.mutations where not is_done;`
One row likes this:
select * from system.mutations where not is_done limit 1 FORMAT JSON;
SELECT *
FROM system.mutations
WHERE NOT is_done
LIMIT 1
FORMAT JSON
{
"meta":
[
{
"name": "database",
"type": ... | 2019-12-29T03:39:48 | 2019-12-29T03:39:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,502,792 | 8,301 | alexey-milovidov | https://clickhouse.yandex/docs/en/development/developer_instruction/ | 2019-12-29T12:54:10 | 2019-12-29T12:54:10 | {} | MEMBER |
ClickHouse | ClickHouse | 569,522,602 | 2,250 | alexey-milovidov | Not relevant anymore.
Added a test just in case: 966f46cfab | 2019-12-29T16:52:49 | 2019-12-29T16:52:49 | {} | MEMBER |
ClickHouse | ClickHouse | 569,076,194 | 8,412 | den-crane | lamda arguments should not be named as columns, use i,j,x,y something
```
SELECT
sum(payment_count_delta_1_1_1_2) as payment_delta_1_1_1_2_count1,
sum(payment_amount_delta_1_1_1_2) as payment_delta_1_1_1_2_amount1
FROM(
WITH groupArray(idx) as index,
groupArray(timestamp) as event_time,
groupArray(paymen... | 2019-12-26T15:27:29 | 2019-12-26T15:27:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,086,879 | 8,369 | ghuname | In my case, all aggregates can be of finite data, but if (sum(sum_weight) * sum(di * workingSec24h)) = 0, it will become NaN, so I have to check the whole expression whether it is NaN or not, or check whether (sum(sum_weight) * sum(di * workingSec24h)) is zero or not.
Anyway thanks for avgIf suggestion.
| 2019-12-26T16:37:35 | 2019-12-26T16:37:35 | {} | NONE |
ClickHouse | ClickHouse | 569,149,282 | 8,419 | alexey-milovidov | Test is missing but Ok. | 2019-12-26T23:25:53 | 2019-12-26T23:25:53 | {} | MEMBER |
ClickHouse | ClickHouse | 569,324,007 | 6,725 | alexey-milovidov | > UPD: is this issue already fixed here ? #6695
Yes. | 2019-12-27T18:43:20 | 2019-12-27T18:43:20 | {} | MEMBER |
ClickHouse | ClickHouse | 569,378,074 | 8,303 | nikitamikhaylov | I have problems with intergration tests. Our PartitionManager() does not restore connection between two nodes. I don't know why it works in a such way (It is rather strange) and written code looks like it seems to be no bug here, but I have some examples. That's why there are many test files in directory but not many ... | 2019-12-28T02:40:45 | 2019-12-28T02:40:45 | {} | MEMBER |
ClickHouse | ClickHouse | 569,470,640 | 8,452 | den-crane | Please check and provide an output of `select * from system.mutations where not is_done;` | 2019-12-29T03:29:23 | 2019-12-29T03:29:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,471,384 | 8,452 | den-crane | OK. I believe there is an issue `There is no query` with one of these of 8183 mutations.
Check `select * from system.mutations where not is_done and latest_fail_reason <> '';`
I would cancel all of them (8183) https://clickhouse.yandex/docs/en/query_language/misc/#kill-mutation starting from the failed ones (with ... | 2019-12-29T03:45:25 | 2019-12-29T03:47:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,473,113 | 8,452 | den-crane | Maybe these 3168 are new ones? After you killed 8133 ?
Try kill again. | 2019-12-29T04:17:44 | 2019-12-29T04:17:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,474,125 | 8,452 | wangzhen11aaa | > Mutations is not supposed to used on daily basis. It should be rare admin operations.
> Check that a mutation is_done before run a next one.
Thanks very much , you guys. It is healthy now. | 2019-12-29T04:38:13 | 2019-12-29T04:40:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,507,386 | 8,455 | amosbird | > Another option is to simply allow /dev/null path.
`/dev/null` is 200% slower than what this pr proposed in `select-format` perf test.
1. `CREATE TABLE IF NOT EXISTS table_{format} ENGINE = File({format}, null) AS test.hits`
```
"min_time": 0.206000,
"quantiles": {
... | 2019-12-29T13:49:21 | 2019-12-29T13:49:21 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 569,113,818 | 8,329 | amosbird | How about a set of aliases to dictGet* functions? Like dictASTGet (smelly name...) however this does serve a purpose of inferring default database names (when used in background jobs, views) | 2019-12-26T18:48:22 | 2019-12-26T18:48:22 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 569,311,532 | 8,428 | amosbird | it's gcc (Gentoo 9.2.0-r2 p3) 9.2.0 and it happens after I rebased to latest master. The fix is similar to what we did in https://github.com/yandex/ClickHouse/blob/master/dbms/src/Functions/GeoUtils.h#L227 | 2019-12-27T17:26:04 | 2019-12-27T17:26:04 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 569,500,967 | 8,369 | ghuname | @vpanfilov
After clarification in https://github.com/ClickHouse/ClickHouse/issues/8437, the best solution is as you have suggested in https://github.com/ClickHouse/ClickHouse/issues/8369#issuecomment-569211904.
Thanks and best regards. | 2019-12-29T12:30:41 | 2019-12-29T12:30:41 | {} | NONE |
ClickHouse | ClickHouse | 569,076,320 | 8,412 | den-crane | duplicate for https://github.com/ClickHouse/ClickHouse/issues/5046 | 2019-12-26T15:28:17 | 2019-12-26T15:28:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,470,909 | 8,452 | wangzhen11aaa | > select * from system.mutations where not is_done;
There are a lot of stuff .
One node outputs:
> select count(0) from system.mutations where not is_done limit 1;
>
> SELECT count(0)
> FROM system.mutations
> WHERE NOT is_done
> LIMIT 1
>
> ┌─count(0)─┐
> │ 8183 │
> └──────────┘
>
> 1 rows in ... | 2019-12-29T03:35:03 | 2019-12-29T03:37:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,472,507 | 8,452 | den-crane | >DELETE WHERE oiid IN (SELECT toInt32(order_item_id) FROM remote
CH does not support such alters. Kill them all.
You should create a dictionary or a table with Engine JOIN on all replicas and populate it with required order_item_id, then run ` alter delete where dictHas...` | 2019-12-29T04:06:54 | 2019-12-29T04:06:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,472,770 | 8,452 | wangzhen11aaa | > > DELETE WHERE oiid IN (SELECT toInt32(order_item_id) FROM remote
>
> CH does not support such alters. Kill them all.
> You should create a dictionary or a table with Engine JOIN on all replicas and populate it with required order_item_id, then run ` alter delete where dictHas...`
```
KILL MUTATION WHERE (dat... | 2019-12-29T04:12:17 | 2019-12-29T04:12:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,473,515 | 8,452 | den-crane | Mutations is not supposed to used on daily basis. It should be rare admin operations.
Check that a mutation is_done before run a next one. | 2019-12-29T04:25:09 | 2019-12-29T04:25:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,520,070 | 8,455 | alexey-milovidov | Now INSERT with all formats look the same: https://clickhouse-test-reports.s3.yandex.net/8455/f0bdcdd677807c039bfbaa3cdd9eca3efaef1bc0/performance_test/comparison_min_time_gcc_9.html
It doesn't look true. | 2019-12-29T16:20:43 | 2019-12-29T16:20:43 | {} | MEMBER |
ClickHouse | ClickHouse | 569,520,310 | 8,455 | amosbird | Hmm, will inspect now | 2019-12-29T16:23:16 | 2019-12-29T16:23:16 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 569,333,865 | 8,441 | den-crane | ```
clickhouse-client -q "CREATE TABLE log (row String) ENGINE = Null"
19.13.4.32
for i in `seq 1 50`; do curl -s -w 'Total: %{time_total}s\n' "http://localhost:8123/?query=INSERT%20INTO%20log%20format%20TSV" --data-binary @test.json ; done;
Total: 1.487038s
Total: 1.462736s
Total: 1.482538s
Total: 1.465072s... | 2019-12-27T19:46:17 | 2019-12-27T19:53:08 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,426,282 | 8,326 | alexey-milovidov | Almost. Because all tests have failed and server has crashed :)
I will look. | 2019-12-28T15:32:36 | 2019-12-28T15:32:36 | {} | MEMBER |
ClickHouse | ClickHouse | 569,444,809 | 8,013 | nvartolomei | #8439 | 2019-12-28T19:28:47 | 2019-12-28T19:28:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,447,752 | 8,013 | alexey-milovidov | @nvartolomei These are two independent refactorings :) | 2019-12-28T20:13:24 | 2019-12-28T20:13:24 | {} | MEMBER |
ClickHouse | ClickHouse | 569,473,297 | 8,452 | wangzhen11aaa | > OK. I believe there is an issue `There is no query` with one of these of 8183 mutations.
> Check `select * from system.mutations where not is_done and latest_fail_reason <> '';`
>
> I would cancel all of them (8183) https://clickhouse.yandex/docs/en/query_language/misc/#kill-mutation starting from the failed ones... | 2019-12-29T04:20:50 | 2019-12-29T04:20:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,392,650 | 8,329 | amosbird | > Maybe we can just enable it regardless to minor compatibility issue.
+1, it has already required a constant string. | 2019-12-28T06:54:09 | 2019-12-28T06:54:09 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 569,121,058 | 5,765 | qoega | Currently it works for me with
`-D USE_INTERNAL_MYSQL_LIBRARY=1 -DENABLE_ICU=0` and #8297 | 2019-12-26T19:35:04 | 2019-12-26T19:35:04 | {} | MEMBER |
ClickHouse | ClickHouse | 569,327,974 | 8,356 | alesapin | Build failed, but it's CI error. | 2019-12-27T19:08:24 | 2019-12-27T19:08:24 | {} | MEMBER |
ClickHouse | ClickHouse | 569,335,899 | 8,441 | alexey-milovidov | @den-crane Could you please also provide `perf report` comparisons? | 2019-12-27T19:59:40 | 2019-12-27T19:59:40 | {} | MEMBER |
ClickHouse | ClickHouse | 569,435,521 | 8,445 | simPod | Definitely, can close, IMO it will get resolved with those. | 2019-12-28T17:24:33 | 2019-12-28T17:24:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,500,735 | 8,447 | alexey-milovidov | We already use tmpfs for performance tests in CI and current regressions in perf tests are well known and to be fixed in another PRs. | 2019-12-29T12:27:22 | 2019-12-29T12:27:22 | {} | MEMBER |
ClickHouse | ClickHouse | 569,556,056 | 8,285 | alexey-milovidov | Yes, the first LTS version is 19.14. | 2019-12-30T00:29:26 | 2019-12-30T00:29:26 | {} | MEMBER |
ClickHouse | ClickHouse | 569,080,503 | 8,415 | 4ertus2 | #7955 | 2019-12-26T15:58:04 | 2019-12-26T15:58:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,093,989 | 8,370 | alexey-milovidov | Suspicious places in code:
1. FunctionsComparison:
```
|| (left.isEnum() && right.isEnum() && arguments[0]->getName() == arguments[1]->getName()) /// only equivalent enum type values can be compared against
```
2. RowInputFormatWithDiagnosticInfo:
```
if (exception)
{
if (type->getName() ==... | 2019-12-26T17:12:08 | 2019-12-26T17:12:28 | {} | MEMBER |
ClickHouse | ClickHouse | 569,095,813 | 8,363 | alexey-milovidov | It was merged to 19.18 but it isn't relevant anymore for master. | 2019-12-26T17:21:23 | 2019-12-26T17:21:23 | {} | MEMBER |
ClickHouse | ClickHouse | 569,116,160 | 8,297 | alexey-milovidov | @abyss7 The build in CI still fails. | 2019-12-26T19:02:15 | 2019-12-26T19:02:15 | {} | MEMBER |
ClickHouse | ClickHouse | 569,116,844 | 8,248 | alexey-milovidov | Filimonov is on vacation until Jan 6th. | 2019-12-26T19:06:31 | 2019-12-26T19:06:31 | {} | MEMBER |
ClickHouse | ClickHouse | 569,122,599 | 5,765 | alexey-milovidov | > Currently it works for me with
-D USE_INTERNAL_MYSQL_LIBRARY=1
It means that we disabling it for MacOS, but there is no reason to do it. | 2019-12-26T19:45:55 | 2019-12-26T19:45:55 | {} | MEMBER |
ClickHouse | ClickHouse | 569,287,887 | 8,438 | hustnn | Does anybody have idea on it? Thx.
I am running ClickHouse on Mac. | 2019-12-27T15:08:26 | 2019-12-27T15:18:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,317,602 | 8,438 | alexey-milovidov | It is related to Mac OS build. I will assign to @qoega. He is PM, not in dev. team, but he has Mac and can reproduce and investigate this issue. | 2019-12-27T18:02:38 | 2019-12-27T18:02:38 | {
"heart": 1
} | MEMBER |
ClickHouse | ClickHouse | 569,323,581 | 7,005 | alexey-milovidov | This issue will be solved in fundamental fashion in https://github.com/ClickHouse/ClickHouse/issues/6787
But simpler solution is possible. We can check that directory already exists on CREATE TABLE and throw an exception. | 2019-12-27T18:40:34 | 2019-12-27T18:40:34 | {} | MEMBER |
ClickHouse | ClickHouse | 569,326,943 | 8,439 | alesapin | Very strange, but tests are OK... I didn't break anything... | 2019-12-27T19:02:08 | 2019-12-27T19:02:08 | {} | MEMBER |
ClickHouse | ClickHouse | 569,344,627 | 8,326 | kononencheg | > `SELECT runningAccumulate()` leads to crash.
Now arguments count are tested before invocation. | 2019-12-27T20:58:46 | 2019-12-27T20:58:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,501,385 | 8,329 | alexey-milovidov | @alesapin @4ertus2 and @KochetovNicolai must write what they think here. | 2019-12-29T12:36:24 | 2019-12-29T12:36:24 | {} | MEMBER |
ClickHouse | ClickHouse | 569,503,183 | 8,011 | alexey-milovidov | @amosbird
Let's cut all the references to idn2 from CMakeLists.
Nothing can be more stupid that depending on IDN for using Amazon SDK. | 2019-12-29T12:59:14 | 2019-12-29T12:59:14 | {} | MEMBER |
ClickHouse | ClickHouse | 569,523,291 | 6,269 | alexey-milovidov | CC @alesapin
We should remove all the logic from `checkDataPart` except reading files and validating checksums. | 2019-12-29T17:00:31 | 2019-12-29T17:00:31 | {} | MEMBER |
ClickHouse | ClickHouse | 569,194,501 | 8,159 | weiqxu | @filimonov I'd like to enhance the document this week. | 2019-12-27T05:46:26 | 2019-12-27T05:46:26 | {
"eyes": 1,
"heart": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,279,967 | 8,369 | ghuname | Yes, this is much better, but I found another one:
`select nullIf(sum(weight * di * workingSec24h * errorsUpXday) / (sum(sum_weight) * sum(di * workingSec24h)), NaN)
`
I just hope that clickhouse developers will not change nullIf(NaN, NaN) = Null in the future. | 2019-12-27T14:23:39 | 2019-12-27T14:34:20 | {} | NONE |
ClickHouse | ClickHouse | 569,333,126 | 8,329 | alexey-milovidov | > Like dictASTGet (smelly name...)
Too much burden. | 2019-12-27T19:41:33 | 2019-12-27T19:41:33 | {} | MEMBER |
ClickHouse | ClickHouse | 569,435,437 | 6,104 | simPod | Same with MV creation. Seems like everything with `AS` is not resolved properly. Both `a` and `b`.
```sql
CREATE MATERIALIZED VIEW mv TO a AS SELECT * FROM b
``` | 2019-12-28T17:23:54 | 2019-12-28T21:44:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,523,316 | 8,455 | amosbird | after fixing the issue, this pr has similar perf result compared to `/dev/null` , will switch to that in favor of simplicity | 2019-12-29T17:00:56 | 2019-12-29T17:00:56 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 569,556,540 | 5,250 | alexey-milovidov | Fixed in https://github.com/ClickHouse/ClickHouse/pull/8311 | 2019-12-30T00:36:46 | 2019-12-30T00:36:46 | {} | MEMBER |
ClickHouse | ClickHouse | 569,069,735 | 8,356 | Alex-Burmak | > Also maybe we need some lightweight class for Path representation? Very often you wrap your String path into Poco::Path.
When dealing with paths and IDisk interface, Poco::Path is currently used only for getting file name from path and getting parent path. So it would probably better to just add two utility functi... | 2019-12-26T14:43:06 | 2019-12-26T14:43:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,378,484 | 8,303 | nikitamikhaylov | Mysterious logic for cell with id=0 without understandable comments.
https://github.com/ClickHouse/ClickHouse/blob/9e8c3acffc79a7b7c3d51c9493c4801c04e477cc/dbms/src/Dictionaries/CacheDictionary.h#L274
https://github.com/ClickHouse/ClickHouse/blob/9e8c3acffc79a7b7c3d51c9493c4801c04e477cc/dbms/src/Dictionaries/CacheD... | 2019-12-28T02:48:47 | 2019-12-28T02:48:47 | {} | MEMBER |
ClickHouse | ClickHouse | 569,427,962 | 8,445 | den-crane | related https://github.com/ClickHouse/ClickHouse/issues/6104 https://github.com/ClickHouse/ClickHouse/issues/6107 | 2019-12-28T15:56:47 | 2019-12-28T15:57:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,502,703 | 8,301 | alexey-milovidov | @mayankkt9 Sure. This is yours! | 2019-12-29T12:53:07 | 2019-12-29T12:53:07 | {} | MEMBER |
ClickHouse | ClickHouse | 569,510,049 | 8,219 | amosbird | fortunately that libunistring.so.2 comes from libidn2.so.0 | 2019-12-29T14:22:32 | 2019-12-29T14:22:32 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 569,087,318 | 8,360 | alexey-milovidov | https://github.com/ClickHouse/ClickHouse/issues/8418 | 2019-12-26T16:39:16 | 2019-12-26T16:39:16 | {} | MEMBER |
ClickHouse | ClickHouse | 569,211,904 | 8,369 | vpanfilov | @ghuname in any case you can use `WITH` clause (https://clickhouse.yandex/docs/en/query_language/select/#with-clause) to reduce code duplication. It will be something like this:
```sql
WITH
sum(weight * di * workingSec24h * errorsUpXday) / (sum(sum_weight) * sum(di * workingSec24h)) AS calculation
SELECT
sumIf... | 2019-12-27T07:36:16 | 2019-12-27T07:36:16 | {
"+1": 3
} | NONE |
ClickHouse | ClickHouse | 569,334,237 | 8,242 | alexey-milovidov | Thank you! Everything is perfect. | 2019-12-27T19:48:55 | 2019-12-27T19:48:55 | {} | MEMBER |
ClickHouse | ClickHouse | 568,924,669 | 8,397 | alexey-milovidov | TurboBase64 library looks half baked and it doesn't support strings longer than 4 GiB.
Update: now it does! | 2019-12-25T19:23:14 | 2019-12-28T00:16:38 | {} | MEMBER |
ClickHouse | ClickHouse | 569,500,800 | 8,437 | ghuname | Thanks for clarification. It will be nice if you put in documentation what you have explained here.
Best regards. | 2019-12-29T12:28:16 | 2019-12-29T12:28:16 | {} | NONE |
ClickHouse | ClickHouse | 569,521,057 | 8,456 | alesapin | We got a race on simultaneous read/write of `system.columns` and alter of `TinyLog`. It's not a direct consequence of this PR and It's not so critical, but fix is simple. | 2019-12-29T16:32:56 | 2019-12-29T16:35:48 | {} | MEMBER |
ClickHouse | ClickHouse | 569,526,379 | 8,426 | alexey-milovidov | It requires iconv that is part of libc, but we don't have it in cross-compile toolchain. | 2019-12-29T17:39:44 | 2019-12-29T17:39:44 | {} | MEMBER |
ClickHouse | ClickHouse | 569,556,073 | 6,501 | alexey-milovidov | Solved #8285 | 2019-12-30T00:29:46 | 2019-12-30T00:29:46 | {
"hooray": 1
} | MEMBER |
ClickHouse | ClickHouse | 569,118,130 | 7,925 | alexey-milovidov | @levysh Should we continue? | 2019-12-26T19:15:18 | 2019-12-26T19:15:18 | {} | MEMBER |
ClickHouse | ClickHouse | 569,356,994 | 8,329 | alexey-milovidov | Maybe we can just enable it regardless to minor compatibility issue. | 2019-12-27T22:28:28 | 2019-12-27T22:28:28 | {} | MEMBER |
ClickHouse | ClickHouse | 569,484,749 | 8,301 | mayankkt9 | Can I work on this ? | 2019-12-29T08:17:21 | 2019-12-29T08:17:21 | {} | NONE |
ClickHouse | ClickHouse | 569,507,570 | 8,221 | alexey-milovidov | We only rolled back `-Wmissing-declarations` and related changes because it requires to collect registration functions in headers, and when we edit these headers adding a new function, it leads to long rebuild time. | 2019-12-29T13:51:22 | 2019-12-29T13:51:39 | {} | MEMBER |
ClickHouse | ClickHouse | 569,522,944 | 1,820 | alexey-milovidov | Just add to `MergeTreeSettings` and pass into `SimpleMergeSelector`. | 2019-12-29T16:56:26 | 2019-12-29T16:56:26 | {} | MEMBER |
ClickHouse | ClickHouse | 569,523,384 | 4,295 | alexey-milovidov | Will be implemented after #6787 | 2019-12-29T17:01:42 | 2019-12-29T17:01:42 | {} | MEMBER |
ClickHouse | ClickHouse | 569,218,549 | 8,429 | zhang2014 | Currently, we have broken all alias type usage, because the client might not know the alias of types.
We should probably add a setting to keep it compatible? for example: `use_canonical_type_name_to_format` ? | 2019-12-27T08:18:46 | 2019-12-27T08:19:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,471,830 | 8,452 | wangzhen11aaa | > OK. I believe there is an issue `There is no query` with one of these of 8183 mutations.
> Check `select * from system.mutations where not is_done and latest_fail_reason <> '';`
>
> I would cancel all of them (8133) https://clickhouse.yandex/docs/en/query_language/misc/#kill-mutation starting from the failed ones... | 2019-12-29T03:54:15 | 2019-12-29T03:54:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,478,934 | 8,219 | amosbird | contrib/icu build somehow introduces libunistring.so.2 dependency with my local toolchain. | 2019-12-29T06:26:44 | 2019-12-29T06:26:44 | {
"eyes": 1
} | COLLABORATOR |
ClickHouse | ClickHouse | 569,500,942 | 8,455 | alexey-milovidov | Ok.
Another option is to simply allow `/dev/null` path. | 2019-12-29T12:30:26 | 2019-12-29T12:30:26 | {} | MEMBER |
ClickHouse | ClickHouse | 569,506,868 | 8,456 | alexey-milovidov | TSan report looks related: https://clickhouse-test-reports.s3.yandex.net/8456/44370467f0bb36471e0508eeae08d65e256a37a1/stress_test_(thread)/stderr.log | 2019-12-29T13:42:59 | 2019-12-29T13:42:59 | {} | MEMBER |
ClickHouse | ClickHouse | 569,552,906 | 8,455 | alexey-milovidov | Ok, but the difference is small and the only query that has degradation still has it:
```
INSERT INTO table_Vertical SELECT * FROM test.hits LIMIT 100000
``` | 2019-12-29T23:44:50 | 2019-12-29T23:44:50 | {} | MEMBER |
ClickHouse | ClickHouse | 569,081,807 | 8,416 | abyss7 | No tests for now. | 2019-12-26T16:06:15 | 2019-12-26T16:06:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,288,028 | 8,438 | hustnn | I just run a server and execute the sql in 00632_get_sample_block_cache.sql. | 2019-12-27T15:09:21 | 2019-12-27T15:18:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 569,478,910 | 8,011 | amosbird | contrib/curl build somehow introduces libidn2.so.0 dependency with my local toolchain. | 2019-12-29T06:26:12 | 2019-12-29T06:26:12 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 569,543,676 | 8,013 | nvartolomei | @alexey-milovidov to me it looks like the other one also covers what this one intended to do with encapsulating alters within StorageInMemoryMetadata, but maybe I didn't pay enough attention. | 2019-12-29T21:30:28 | 2019-12-29T21:30:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 568,698,104 | 7,969 | chukynax | there was a problem with name resolution when clickhouse and kafka's were on the same hosts | 2019-12-24T08:55:16 | 2019-12-24T08:55:16 | {} | NONE |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.