While using Power BI for a few months now, we (the user group) encountered an issue that is not really clear to us.
We use Power-BI with a remote SQL-Server data source, we access the data source through direct query.
Let's pretend we have 2 Tables as below-
Table name: Issue
Column:
ResolutionTime(Date/Time)
IssueID(Unique Numbers)
Table Name: WorkItem
Column:
start (Date/Time)
end (Date/Time)
IssueID (Unique Numbers, Foreign Key to "Issue" table)
Table WorkItem also contain a calculated column "WorkTime" which uses this DAX-expression as below-
WorkTime = WorkItem[end] - WorkItem[start]
The two tables are configured through Power-Bi having a two-way 1:n relationship that can be queried to collect all "WorkItem"(s) assigned to an "Issue" entry, using the "IssueID" as correlation column.
To be able to compute the aggregated "work-time" for each "WorkItem", we use a new/calculated table with the following DAX expression to aggregate the total amount of time invested for a single "Issue":
SumWork =
SUMMARIZE(
WorkItem, WorkItem[IssueID], "All work per item", SUM(WorkItem[WorkTime])
)
The above table computes the total invested work-time for a particular issue, grouping/summarizing results based on the "IssueID" foreign key. This new calculated table is also configured to have a relationship with the "Issue" table, this time a "1:1" relationship, using the IssueID as correlation column.
Now to compute the time that the issue was worked on + the time for Resolution should be summarized in a calculated column inside "Issue", but this does not work:
ResolutionAndWorkTime = Issue[ResolutionTime] + SumWork["All work per item"]
But the above DAX expression fails to compile, as it always reports that it returns "more than one result", thus not being a singular result. But that is suprising, as the two table ("Issue" and "SumWork" are related to each other with a "1:1" relationship).
Tables:
Issues
IssueID ResolutionTime ResolutionAndWorkTime
1 03:20:20 ???
2 01:20:20 ???
3 00:20:20 ???
WorkItem
IssueID start end WorkTime
1 1-2-2020 3:20:20 1-2-2020 3:25:20 00:05:00
1 2-2-2020 6:20:20 2-2-2020 7:20:20 01:00:00
3 1-3-2020 3:20:20 1-3-2020 3:29:20 00:09:00
Any ideas what to look for? Data-types? Table-definition? Table-relationships? We checked other Stackoverflow questions/answers, but no good ideas retrieved so far.
NOTE that a lot of join/merge features of Power BI are not available if direct-query is used and thus joining the tables is not really an option (we think).