channel、pageId，分桶数是10
问题：命中相同的rollup表，用bitmap聚合的时间是不用bitmap聚合的50-100倍左右。不清楚哪块的问题造成bitmap性能这么差
以下是bitmap聚合查询sql和耗时：
select dayStr dt,hourStr hour,channel,pageId,eventId,sum(pv)
pv,bitmap_count(bitmap_union(uv)) uv from aggr_user_action_event_rt_v2 where
dayStr=20210312 and hourStr=9 and
wangbo commented on issue #5509:
URL:
https://github.com/apache/incubator-doris/issues/5509#issuecomment-797394279
你可以加一个profile的指标看下
即使在命中rollup的情况，在基数比较高时，目前bitmap的主要计算开销在bitmap列的反序列化与基数计算上
还有一部分开销在内存池的分配和释放上
对于普通用户来说，目前bitmap比较快速的的优化方式如下
1 增加并发度，包括单机的查询并发以及增加更多be节点
2 优化输入数
EmmyMiao87 commented on issue #5509:
URL:
https://github.com/apache/incubator-doris/issues/5509#issuecomment-797417753
The bitmap algorithm itself does not have very good computational
performance when the cardinality is large and the distribution is loose.
Performance tuning needs to be
morningman closed issue #5442:
URL: https://github.com/apache/incubator-doris/issues/5442
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to
morningman merged pull request #5504:
URL: https://github.com/apache/incubator-doris/pull/5504
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL abov
This is an automated email from the ASF dual-hosted git repository.
morningman pushed a change to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-doris.git.
from 543ed46 (#5390)fix NPE when replay colocate group (#5391)
add e9a73ee [Bug] Fix the memory expand
xxiao2018 opened a new issue #5510:
URL: https://github.com/apache/incubator-doris/issues/5510
Sometimes we find that the scanner thread itself takes a short time to scan
data,
but the overall OLAP_SCAN_NODE takes a long time.
Analysis found that part of the time was spent waiting for
xxiao2018 opened a new pull request #5511:
URL: https://github.com/apache/incubator-doris/pull/5511
## Proposed changes
1.
Add timer to count the time the transfer thread waits for the scaner thread
to return rowbatch.
2.
Add timer to count the time that the scanner thread w
xxiao2018 opened a new issue #5512:
URL: https://github.com/apache/incubator-doris/issues/5512
**Describe the bug**
Query stmt like:
`SELECT /* SER_VAR(var1=xx, var2=xx) */ ...`
does not support.
This is
xxiao2018 opened a new pull request #5513:
URL: https://github.com/apache/incubator-doris/pull/5513
## Proposed changes
Sometimes we need to set multi variables for a single query stmt. like:
SELECT /* SER_VAR(var1=xx, var2=xx) */ ...
## Types of changes
- [x] Bug
stdpain closed issue #5213:
URL: https://github.com/apache/incubator-doris/issues/5213
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go
stdpain opened a new pull request #5514:
URL: https://github.com/apache/incubator-doris/pull/5514
This Patch work well on tpcds-1T query-24
NOTE: we shouldn't apply this patch to [internel-master] because of
MinMaxFilter
## Proposed changes
will close #5503
## Types
acelyc111 commented on a change in pull request #5514:
URL: https://github.com/apache/incubator-doris/pull/5514#discussion_r593548993
##
File path: be/src/exec/hash_join_node.h
##
@@ -174,6 +175,11 @@ class HashJoinNode : public ExecNode {
// This is only used for debuggin
HappenLee opened a new issue #5515:
URL: https://github.com/apache/incubator-doris/issues/5515
## Motivation
At present, Doris MemTracker are very chaotic. There are a lot of memory
occupation not recorded. As a result, many scenes can not accurately judge the
memory usage, and quickly
HappenLee opened a new pull request #5516:
URL: https://github.com/apache/incubator-doris/pull/5516
This PR main about:
1. Improve the readability of MemTrackers' name
2. Add the MemTracker of:
* Load
* Compaction
* SchemaChange
* StoragePageCache
13671653088 opened a new issue #5517:
URL: https://github.com/apache/incubator-doris/issues/5517
hi all,
I deployed a doris cluster with one fe and three be, version 0.12.
fe and be 10002 deployed on one server,and be 10003,10004 deploy on the
other two servers;
show backends can't
amosbird opened a new pull request #5518:
URL: https://github.com/apache/incubator-doris/pull/5518
## Proposed changes
Some fs might not support O_DIRECT and O_SYNC is semantically the same to be
used for disk checking.
--
17 matches
Mail list logo
