Forum Discussion
PC crashes when doing a left join
- 5 years ago
Ahh okay Merging is not joining. now I understand and for sure its a size issue that you are facing.
Try doing it on Power BI Service, or through Power BI Dataflows in case you have a Premium capacity or Pro License. That may reduce the usage of your memory.
Other options would be filtering the tables by some criteria/range and then appending them. Not efficient, but still a workaround.
The datasource are PowerBI Dataflows in a premium environment. And i have a pro license.
Yeah . . . i already skimmed as much as i could, this crashing is also the reason i am only trying left join 1 columnn instead of 8 (which is the bare minimum i need).
Ill try exporting the dataflows to 2 csv files and see i if can join them that way. If that would work, it should point to an issue within the DataFlow itsself, no?
- aj19735 years agoCommunity Champion
Sorry I didn't get your scenario! You have 2 dataflows, you call these dataflows to your Desktop and you pick different entities from these 2 Dataflows. In your model you try to join (relationship) 2 entities from those 2 dataflows and your system crashes?
Is this your issue?
- decarsul5 years agoHelper V
I do the left join from PowerQuery, not from DAX relationmanager if that's what you mean.
- aj19735 years agoCommunity Champion
Sorry but why using PowerQuery for joining? Power Query is good for ETL (Extraction, transformation and loading). Modeling(joins and relationships) is done in power bi desktop and DAX.
- Anonymous5 years agoNot applicable
If you are trying to join the dataflows in Desktop, it will still consume your system's memory. Try joining them on the Power BI Service on Get Data.
Exporting the CSVs may work but still if they have a considerable size, they'll likely crash again.
Other thing that could work could be creating a new dataflow that joins the other 2 dataflows, you can use the concept of linked entities since you have a Premium environment.
- decarsul5 years agoHelper V
Yeah i'm aware it takes system resources, but i don't see how a 2x 500Mb can eat up 32Gb of ram. I'm really only 'adding' 1 column with a zipcode.
Yeah i can try joining them in the dataflow in the premium environment. See if that works.