Forum Discussion
data lineage
I'm in exactly the same situation than you.
It seems that PowerBI doesnt allow LEFT OUTER JOIN relations tables as SQL.
I just found one solution but it doesnt seduce me: subject here
I don't understand the aim to merge large tables by creating a new one (and getting multiples colums) when we do know that will reduce the performance.
Any help please ?
Sorry for the delay, I have had to create a new account as it wouldn't let me reset the password....
You probably aren't going to like the work around. Basically I'm using R to clean up the data before loading it.
That said the data I have is so large that I haven't been able to load it into memory to do the join in R. So I now have a workaround on the workaround.
There is an R package which allows the data to be used in creating an SQLlite DB on disk just for the duration of a command. It takes an inordinate amount of time to run but has no restriction on the size of the data and allows some complex SQL to manipulate the data.
That requires 'R' to be installed (it's free public domain software and has a Microsoft supported version) and I have some experience with it. Apparently the latest BI has python available as well, but being an R person I wouldn't know much about it.
Did you try to create combinations of columns and use them as an index to link the two ?
Roy