Forum Discussion
Basket Size Analysis
- 7 years ago
Ashish, thank you so much! To make sure I understand, you essentially create a "Buckets" dimension table and then do the calcs to fit the fact table measures within each bucket's dimensions?
Trying this on our actual sales table, the concept works, but execution on a large transaction table (billions of rows), doesn't seem to work. I believe it's the addcolumns/calculatetable driving the issue as the error mentioned cannot return results of more than 1000000 rows (1 million). So potentially on a smaller raw table this would work. Unless I missed something else that would result in this issue.
We've resorted to, for now, creating a group by in the edit query. It's a first step, but ultimately to answer the further basket questions of how often does my customer buy X and only X and in what basket size quantities, I need something like what you've done.
Thanks again! I consider this answered.
You are welcome. Your understanding is correct. It's the large number of rows which is causing the slowdown for sure. I cannot seem to suggest an alternative.