Forum Discussion
Left outer Join - but only selected columns from both tables
Thank you - if performance isn't impacted my question is redundant then.
However, I was aware of only selecting the columns.
THe problem is , in Table A for instance, what if I still need to refer to the full Table for certain visualisations, while the merged table doesn't need those columns? What I mean is, pratically, I'd go to Table A, pick the coumns I need, and then merge with Table B to create a new table. However, Table A now only has the reduced number of columns. What if one of the ones I removed to create the merge is one I actually need in Table A?
I see what you mean.
Make a new query as a reference to TableA, select your columns and do the merge.
This will keep your TableA as it is and will give you a new table with the merge result.In real life,my Table A in extremely large - 500m rows +. Even if I make a reference, it's still hitting the dB twice isn't it? I was trying to avoid doing this by only selecting the columns needed to at least save some performance - but if columns don't matter I guess that's a moot point.
What I did try was to copy the table in DAX and merge it from there. This was successful, although I ran into another issue as I then tried to use this table as part of a scatter chart, and it wouldn't let me drag anything into the Y axis field. Not sure if it's related to how my new table is built, or something completely unrelated... but in theory, is this methond a lot quicker than doing a duplicate/reference in PQ? Cheers, Andrew
