cancel
Showing results for 
Search instead for 
Did you mean: 

Problem Merging Datasets from Excel in Lumira

Former Member
0 Kudos

I was trying to merge two datasets from Excel and kept getting the error "No column in the Lookup Dataset matches...." when I tried tolink by Username.

When I edit the two datasets within Lumira I can see a difference.

One dataset holds the Username as 'SMITHJ'; the other as SMITHJ (no inverted commas); so I assumed this wass the problem.

Thus I thought that in Excel the format might be different.

But even when I copy/pasted between the cells in Excel to ensure they were exactly the same (& saved the Excel file & refreshed the datasets in Lumira) I found that Lumira had again imported one as SMITHJ, the other as 'SMITHJ'.

After much cursing I found a solution of sorts...

It seems that if your column is at the end of the field list in Excel it won't merge.

The order of my columns in one of my Excel sheets was a few Dimension (ABC) columns, then Measure (123) columns and then the last column (which I was trying to merge on) was another Dimension.

Seems like Lumira doesn't like this when attempting to merge. Whatever I put in that last column it wouldn't merge.

In the end I copy/pasted that column so it sat next to the other Dimension columns in Excel; then re-imported the data.

When I tried to merge this time worked fine.

Thought I'd flag this up in case it can be fixed in a future release (or in case I'm doing something wrong...)

Regards

Rob

Accepted Solutions (0)

Answers (1)

Answers (1)

Former Member
0 Kudos

Ah, actually that solution did not work, well not in a repeatable way. Thought I'd post screenshots here in case anyone has any ideas. Even with data cut down to 10 rows I couldn't get it to merge.

I also tried exporting to csv files and loading from there, but that didn't work.

Lumira version 1.15.1; Office 2010.

som
Advisor
Advisor
0 Kudos

Hi Rob

Can you confirm whether the key you are joining on the LOOLUP DATASET is a primary key (unique values)?

Thank you.

Sharon


Former Member
0 Kudos

Hi Sharon

It's a 1:n join; The first file has a unique list of User Names, 1 record per user; the file I am joining to has a list of Queries those users have run, so there are multiple records per user.

Regards

Rob