Hi Jialin,
Thanks for the reminder! I will make sure to use Long.MIN_TIME as the lower bound in my code. 在 2020-06-23 15:52:58，"Jialin Qiao" <[email protected]> 写道： >Hi, > >Thanks! One thing we need to pay attention, we support time < 0. >So It's better to record a [Long.MIN_TIME, T] for time <= T > >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: "Wei Shao" <[email protected]> >> 发送时间: 2020-06-23 15:31:45 (星期二) >> 收件人: "[email protected]" <[email protected]> >> 抄送: >> 主题: I've submitted a PR for issue IOTDB-627 >> >> Hi all, >> >> >> I submitted a PR[1] to support range delete for a timeseries. JIRA issue is >> [2]. >> >> >> Now the "delete from ..." SQL support specifying a range in the where >> expression. The where expression may contain Lt/LE, Gt/GE, EQ, AND with two >> Lt/LE, Gt/GE, EQs. >> The deletion range is represented by a [startTime, endTime] interval in >> current implementation. >> >> >> The record format in Mods file has been changed as well. For a deletion SQL: >> "delete from root.sg.d0.s0 where time <= 100 and time >= 50" >> Assuming the version number is 10, then it will leave a line "root.sg.d0.s0, >> 10, 50, 100" in the Mods file. >> For Mods files with only one "endTime" timestamp in previous version, the >> query engine will recognize it as having a [0, endtime] deletion range. >> >> >> Please leave your opinions. >> >> >> [1] https://github.com/apache/incubator-iotdb/pull/1400 >> [2] https://issues.apache.org/jira/browse/IOTDB-627 >> >> >> --------------------------- >> Thanks, >> Wei
