That seems odd.
One thing that could cause that is if you had a data source filter. Check by using the drop down menu on the data source and select Edit Data Source Filters. If you have a data source filter that is limiting the number of records to 3.1M, then the extract could pull in 26M, but you'd never see most of them.
I believe that is because it is factoring in columns as well. So if you have 1,000 rows and 20 columns it will count as 20,000 rows in that count up window, or something along those lines.
I use data source filters alot, but now you have me wondering if I included the extract filter. wish Tableau would make extracts automatically check for data source filters & make extract filters = data source filters.
this is something I hadn't considered. I'll check the math for columns x rows. Maybe the dialog window is actually trying to say xxxxxx values retrieved....
Just seems strange cuz I've noticed this on every extract for 4 years since I've been using Tableau.
Tableau does this at the very first extract creation - we promote/copy all data source filters to be extract filters however since those two are pretty different - we don't maintain that parity if you make changes to either.
1 of 1 people found this helpful
You did probably pick incorrect join keys and created a cross-product.
When you say "1000 rows" - do you mean that each out of 5 tables has 1000 rows?
You are correct, my joins are weak and did create a cross-product records.
I had the incorrect perception due to the no or records shown in the Tableau data source view. Apologies n00b error.
Thanks for the reply.
Don't beat yourself up - it's a very common mistake all of us do from time to time.
This is one of the areas we are currently brainstorming on: how to help people choose the right join keys. We already automatically join on PK/FK or matching names but we should do much better like performing column domain analysis and telling users "hey, you are going to get a crossjoin" before they execute it. I dream big