Getting null values instead of actual values when blending data

Hi, I am blending data and using left outer join, however I keep getting null values instead of actual values. As far as I'm concerned null should only appear when the value is missing or 0, however in this case I am supposed to be getting a conversion rate in %. What am I doing wrong? 

As can be seen in table 1, I am measuring exits (metric) and using the event_label (which is the name) as the dimension.
In table 2, I am measuring the pop-up views (metric) and using the event_label (which is the name as the dimension).

Then I am using left outer join and using the event_label field to join both tables.

Then in the set up section of the new chart I am using event_label as the dimension and added the new conversion rate metric using the formula: SUM(Exit From Pop-up) / SUM(Pop-up Viewed).

Despite, this the values for the conversion rate are still being shown as null, when in reality there should be percentages displayed.

Screenshot 2024-08-14 120747.pngScreenshot 2024-08-14 120758.pngScreenshot 2024-08-14 120819.pngScreenshot 2024-08-14 120836.pngScreenshot 2024-08-14 121510.png

0 8 2,260
8 REPLIES 8

Blends are calculated by Studio and each Table has a row limit applied.

My guess is that the size of the table is too high and second table does not return the same keys of the first one. 

What is the table limit? Can I check it somewhere? I have the same issue like above

Hi Francesco, The strange part is that sometimes when I refresh the sheet the value appears, whereas at other times I keep refreshing and they don't...

I think that's because when Looker Studio chooses what rows it will use in the blending limitation it picks randomly which one is used. So with every refresh the blending rechooses the rows and this impacts which ones have a result and which ones not.

I found this really helpful: https://www.youtube.com/watch?v=mjAPgAeKaFY

if you can share the demo report we can check 

I notice that this will happen often when I use more than one filter for a source in a blend, or when one of my filters are wrong. I actually just had this issue yesterday with a report and found that one of my filters was off by 1 character.

I would suggest first verifying that your filters are not the issue here by creating a new table with GA4 as the source (not the blend), and then add your filters one at a time to see if it breaks. If it breaks when you add any of the filters, you know that you need to adjust that filter. 

The basic issue is, on a join like left join, there can be missing elements from the right side.  So if you have a calculation that includes a right metric, this can cause null values to be included in your calculation.

To solve this, you need to use something like a Coalesce function which can force null values into a non-null value.  So instead of

sum(orders)

you should use

sum(coalesce(orders,0))

which will make null values 0.