Hi,
First of all, sorry for the delay doing this function as I got some personal
stuff to do these days. I got some problem while implementing this issue.
It could be implemented in some place by changing the filter operator from
“and” to “or” within the group by device function such as below:
Hi,
I want to share some of my thoughts about the to-be-changed ,
after I took a look at the introduction of GROUP BY in the Microsoft SQL Docs
[1] and
ISO/IEC 9075-2:2003 2 ("SQL/Foundation") .
Basically I agree with the semantic changes of as Tsai
described in the email.
I try to sum
Hi,
The term "group by device" is fine for ok.
Or, we can I think classify the queries into 3 part:
1. return data points time series by time series, while in each
timeseries, the data points are ordered by the timestamp. (Or, totally
ordered by timestamp for all timeseries)
2. Join all time
Hi Tsai,
+1 for your idea. It sounds good to me.
Best,
Lei Rui
On 11/23/2019 09:42，Jack Tsai wrote：
Hi all,
I am recently working on this issue:
https://issues.apache.org/jira/browse/IOTDB-305, which is about resolving the
problem while users want to query something like below as the issue
Hi all,
I am recently working on this issue:
https://issues.apache.org/jira/browse/IOTDB-305, which is about resolving the
problem while users want to query something like below as the issue reporter
Lei Rui said:
select * from root.sg.d1,root.sg.d2 where s1=1 (a demo
conception)
/ select *
