[jira] [Commented] (IOTDB-2549) [Query performance] The query performance of aligned timeseries is slower than that of unaligned timeseries

["Minghui Liu (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/IOTDB-2549?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17495103#comment-17495103 ] 

Minghui Liu commented on IOTDB-2549:
------------------------------------

CPU sampling via JProfiler

[^aligned.jps]

 [^non_aligned.jps]

> [Query performance] The query performance of aligned timeseries is slower than that of unaligned timeseries
> -----------------------------------------------------------------------------------------------------------
>
>                 Key: IOTDB-2549
>                 URL: https://issues.apache.org/jira/browse/IOTDB-2549
>             Project: Apache IoTDB
>          Issue Type: Improvement
>          Components: Core/Query
>    Affects Versions: 0.13.0-SNAPSHOT
>            Reporter: 刘珍
>            Assignee: Minghui Liu
>            Priority: Minor
>         Attachments: aligned.jps, config.properties.no-vector, config.properties.query, config.properties.vector, image-2022-02-16-14-44-40-880.png, image-2022-02-16-14-44-43-366.png, iotdb-engine.properties, iotdb-env.sh, non_aligned.jps
>
>
> master-0214
> commitID: 
> 08c46e3a2a45e1190d57ac151c6bdcb89a7b0107
> 问题描述：
> 对齐序列的查询性能比非对齐序列慢（顺序数据）。
>  !image-2022-02-16-14-44-43-366.png! 
> Q5 sql语句类似如下：
>  SELECT count(s_4), count(s_9), count(s_12), count(s_16), count(s_10), count(s_5), count(s_14), count(s_6), count(s_17), count(s_18) FROM root.test.g_4.d_84 WHERE  root.test.g_4.d_84.s_4 > -5 AND root.test.g_4.d_84.s_9 > -5 AND root.test.g_4.d_84.s_12 > -5 AND root.test.g_4.d_84.s_16 > -5 AND root.test.g_4.d_84.s_10 > -5 AND root.test.g_4.d_84.s_5 > -5 AND root.test.g_4.d_84.s_14 > -5 AND root.test.g_4.d_84.s_6 > -5 AND root.test.g_4.d_84.s_17 > -5 AND root.test.g_4.d_84.s_18 > -5
> 复现步骤：
> 数据库配置文件，benchmark配置文件见附件。
> 1.测试机器
> 192.168.10.68
> 数据库路径：/data/liuzhen_test/0214_iotdb2517
>            数据备份：
>            data_no-vector
>            data_vector
> benchmark：
>          /data/liuzhen_test/aligned_test/benchmark
> 2. 生成数据
> iotdb关闭合并。
> 使用config.properties.vector配置生成对齐序列的数据。
> 使用config.properties.no-vector配置生成非对齐序列的数据。
> 3. 执行查询
> 使用config.properties.query配置对比查询性能。



--
This message was sent by Atlassian Jira
(v8.20.1#820001)