Showing results for 
Search instead for 
Did you mean: 

Alteryx designer Discussions

Find answers, ask questions, and share expertise about Alteryx Designer.

Join on Multiple Columns Dynamically

I am trying to join two files to bring in certain data to match up with my data set.  My issue is my criteria are not static.  I have a sample map attached. 


The issue:


My map has 20+ criteria but not all will be used every time for the join.  For example to get to a join for a specific piece of data I may only need 5 out of the 20 criteria fields to make the join and other times I may need more.  The map has the necessary combination of criteria to get to the answer but I do not know how to dynamically choose the criteria.




how do you choose the column to use for the join ? What is your criteria ? If I understand all culumns are always in the files, is it correct ? what's happen if you join a column ususally not needed in the join ?



Well that is my issue.  The sample spreadsheet is the map I am trying to use in the join (columns B-G are the join columns plus many more in the live matrix).  Choosing the columns is the issue.  I will be feeding lines of data (left side) that may need to use column B, C, and G where the next line may need column B, G, and E for example.  I am trying to see how join on columns necessary for one line of data while joining on a different set of data for the next line I data.  The map will have the columns I am looking to join on.  


Could you please provide 2 sample lines with 2 different cases from le lest side ?

The left side tab has the two lines.  The green highlighted columns are the join columns.  On the right side tab the green highlighted rows are the rows I want joined to the left side based on column B, C, and D.  Row 3 uses column B and D to join while Row 4 uses column B and C to join.


I understand now, I don't see at this point any easy solution (but I ban be missing something), so I have created a solution but you will explicitely create all possible join and do a union on your data. Attached a sample.





This would work if I had only a few categories.  I have 30+ columns so there are a lot of combinations.