Hi Ning, Thanks for your contribution! The PR looks good to me, I have merged it.
Steve Su 苏志 <273670...@qq.com.invalid> 于2022年5月1日周日 23:23写道: > > The following shows the SQL syntax of registering a trigger. > > CREATE TRIGGER <TRIGGER-NAME> > (BEFORE | AFTER) INSERT ON <FULL-PATH> AS <CLASSNAME> WITH ( > <KEY-1>=<VALUE-1>, <KEY-2>=<VALUE-2>, ... )TRIGGER-NAME is > a globally unique ID of the trigger, which is case sensitive. At present, the > trigger can listen to all data insertion operations on the time series. The > hook can be called BEFORE or AFTER the data is inserted. FULL-PATH can be a > time-series (measurement) path such as root.sg1.d1.s1, a device path such as > root.sg1.d1, a storage group path such as root.sg1, or even a non-measurement > path with business semantic such as root.sg1.x. CLASSNAME is the full class > name of the trigger. Note > CLASSNAME, KEY and VALUE in the attributes need to be quoted in single or > double quotes. > > Only one trigger can be registered per full-path. > > When multiple prefix paths of a path are registered with triggers, for > example, trigger trigger-sg1d1s1 is registed on root.sg1.d1, trigger > trigger-sg1d1 is registed on root.sg1.d1, and trigger trigger-sg1 is registed > on root.sg1. When inserting data to root.sg1.d1.s1, triggers will be > triggered in the following order: trigger-sg1d1s1 -> trigger-sg1d1 -> > trigger-sg1. > > Best, > ------------------------------------ > gongning