Hi, > The answer is NO. If it calculated twice, it's a bug.
So, it is a typo of the document, the correct version should be [0, 1d), [1d, 2d). > The new group by statement is like this: select count(*) from root where > time >= 1 and time <= 100 group by ([0, 100], 20ms, 20ms) > The 0 in [0, 100] is used as the start point. This statement should satisfy > most scenarios. Ah... so you give up the ability that letting users define the start point? Besides, I feel confusing about: `time >= 1 and time <= 100 group by ([0, 100]`, if `1` is for the time filter, and `0` is for the start time point, so why we have two `100`? What it is mean about `time >= 1 and time <= 99 group by ([0, 100]`, or `time >= 1 and time <= 100 group by ([0, 99]` ? Best, ----------------------------------- Xiangdong Huang School of Software, Tsinghua University 黄向东 清华大学 软件学院 Jialin Qiao <[email protected]> 于2019年12月10日周二 下午6:55写道： > Hi, > > > My question is, is the data point at 1d calculated twice? > > The answer is NO. If it calculated twice, it's a bug. > > > it loses the ability to define the start point of the timestamp for > splitting the time window. > > The new group by statement is like this: select count(*) from root where > time >= 1 and time <= 100 group by ([0, 100], 20ms, 20ms) > The 0 in [0, 100] is used as the start point. This statement should satisfy > most scenarios. > > Thanks, > Jialin Qiao > > xiangdong Huang (Jira) <[email protected]> 于2019年12月10日周二 下午4:50写道： > > > xiangdong Huang created IOTDB-358: > > ------------------------------------- > > > > Summary: Questions about the `group by` down-samplying > clause > > Key: IOTDB-358 > > URL: https://issues.apache.org/jira/browse/IOTDB-358 > > Project: Apache IoTDB > > Issue Type: Bug > > Affects Versions: 0.9.0 > > Reporter: xiangdong Huang > > > > > > When reading issue #331, I read the `group by` document again and find > > some confusing contents: > > > > V0.9: > > > > [http://iotdb.apache.org/#/Documents/0.9.0/chap5/sec2] > > # (minor), typo, Fig 5.2 rather than Fig 3.2 > > # The first parameter of the GROUP BY statement above is the time > > interval for dividing the time axis. Taking this parameter (1d) as time > > interval and the default origin as the dividing origin, the time axis is > > divided into several continuous intervals, which are [0,1d], [1d, 2d], > [2d, > > 3d], etc. > > > > My question is, is the data point at 1d calculated twice? > > > > > > > > V0.10: > > [http://iotdb.apache.org/#/Documents/progress/chap5/sec2] > > > > the syntax is quite different with v0.9.0. I originally thought it was an > > improvement comparing with 0.9, but when I read the document, I find it > > loses the ability to define the start point of the timestamp for > splitting > > the time window. > > > > Can someone tell me why to design like this? > > > > > > > > > > > > > > > > -- > > This message was sent by Atlassian Jira > > (v8.3.4#803005) > > > > > -- > ————————————————— > Jialin Qiao > School of Software, Tsinghua University > > 乔嘉林 > 清华大学 软件学院 >
