I'm not positive it's called data blending when editing relationships, but I think you can accomplish this using that feature since you said you have 2 different datasources.
I just shared this information in another forum post on this topic here:
Go to Data --> Edit Relationships
should bring up something that looks like this:
Click Add and choose the field from each source that would serve as the join. You can add more than one field as in the example above. In your case, you'd find "OrderID" on both sources.
Create a view with data from both views, but make sure you add the field you're joining on SOMEWHERE on the view. Level of Detail is a good rule of thumb, because if you do not add all the joining fields at least there, the join won't be able to work properly. So in your case, make sure you have OrderID at least in the Level of Detail, if it's not on one of your other shelves actually being used in the view.
Hope this helps!
This is definitely closer to what I was looking for! Though I think it's still showing me all of the items (in this case surveys) instead of doing an inner join and only showing me results for surveys with order IDs available. Also I now have a problem with overlapping data.
Are you able to post a packaged workbook?
Yes - I'm attaching it here. There are two tabs - the first shows what happens if I don't put "Order ID" in the level of detail (the measures from the orders table shows up as just '*'), and the second what happens if I do put "Order ID" in the level of detail (and there is overlapping text, as well as only one order per bucket). I've weeded out the survey data so only surveys with order IDs show up - there are 107 orders total (all represented in the first tab - just not broken down by Groupon/Non-Groupon).
NPS Report tb.twbx.zip 2.3 MB
Wow. I apologize Dalia...
This one slipped off my radar. Got busy at work...
At any rate, I had to brush up on blending vs. joining.
These are the resources I could find:
You were on the right track, but it seems when joining data sources in Tableau, the secondary source always provides an aggregation. And when you attempt to pull a dimension from the secondary data source to "cut" your primary source, it has to use the Attribute tag. So, you aren't actually able to pull the row-level dimension from the primary source and secondary source.
After tinkering for a while, I couldn't get any further. So I had to resort to linking the data on the OrderID using a vlookup in Excel. That way, you can get both the ScoreType and Non/Groupon dimension values for each record at a non-aggregate level and aggregate the data in your view.
See attached workbook and Excel source file.
Again, sorry for the delay!
Not a problem! Thanks for the links - it looks like I will just have to use vlookup. Thanks again for the help!