Table Grouping/Cumulative


(Ethan Blitstein) #1

I am working on a project for my retail team so they can see conversions in less than 3 days vs. conversions >= 3 days by retail store by week. If i were working in SQL I would just group by the store name. but i am using a calculation like: if(${captured_consumer_lead.conversion_whole_days}<3,${captured_consumer_lead.count_conversions},0) to show me the count of consumer leads converted in less than 3 days per week.

I’m hiding the whole day conversions from the visualization but since everyt case with a different number of whole days as the conversion time formulates as a different row, i am seeing the same store listed 20 times.

Long story short, I can’t figure out if I can do something to group by store name. I think it would be easier for me to create a measure for conversions in less than 3 days and conversions >= 3 days. Then when i pivot by week it would be simple. But i don’t have permissions currently to build a measure and have to wait till the next push. Is there an interim fix for this grouping? Hope to hear from someone soon.


Sales by Yesterday and WTD
(David Chiaramonte) #2

Hey there @Ethan_Blitstein!

Based off of what you’ve described above, it looks like doing this in the front end through table calculations isn’t going to function how you’re needing it to. Ideally, we’d want to create filtered measures (assuming your conversion field is a measure) that filtered for >=3 days and <3 days respectively, but seeing as how you don’t have the permissions to push LookML, we may want to create a dimension in the front end. If your Looker admin has enabled the custom fields feature and given you access, you can add new custom dimensions and measures yourself. What we can do if you have the permissions is create a yesno dimension–something along the lines of “Is Greater than 3 Days”–that you can then pivot on to see side by side comparisons for conversions, pivoted to see yes (is >= 3 days) or no (is < 3 days). Feel free to shoot an email to support@looker.com if you need help with the specifics of this issue too!

Cheers,
David