Hi, > 4. the number of chunks, and average size of each chunk.
Thanks, just an explanation, the average size of chunk is the average number of data points in each chunk. For now, the query performance highly depends on the number of chunks need to read. So measuring the chunk size is important for us to locate the problem. Bests, -- Jialin Qiao School of Software, Tsinghua University 乔嘉林 清华大学 软件学院 > -----原始邮件----- > 发件人: "Xiangwei Wei" <[email protected]> > 发送时间: 2020-06-22 00:36:24 (星期一) > 收件人: [email protected] > 抄送: > 主题: I've submitted a PR for issue IOTDB-736 Query performance tracing > > Hi, all > I submitted a pr[1] to support [IOTDB-736] Query performance tracing[2]. > > I try to provide the performance tracing function by adding a performance > log and give a switch statement "Tracing on/off" to control whether to > print this log, which is off by default. > > The log records including: > > 1. the query sql, and the start time of this query > 2. the number of series paths > 3. the number of distinct tsfiles, including sequence and unsequence > 4. the number of chunks, and average size of each chunk. > > This pr concentrates on raw data query, it may modify something for > aggregation query and group by query later. > > > Please leave your opinions, thanks. > > [1] https://github.com/apache/incubator-iotdb/pull/1399 > [2] https://issues.apache.org/jira/projects/IOTDB/issues/IOTDB-736 > > -- > Best, > Xiangwei Wei
