Why the ADD/UPDATE/DELETE pattern doesn't work

[NigelG]

As I asked in this thread…

Airtable daily update

I have a daily file of listings that I need to update in Airtable from a CSV. What I could do is clear down the Airtable table every time and load it fresh, but not sure if that is the best way of doing it? The alternative is to do an “overlap” match and update the ones that match, add the ones that don’t or delete if they are not on the new file. That seems like I would need to set up three “Overlaps” on the same data and do A = B then A not in B and B not in A. API call to Airtable for eac…

but now about Bubble rather than Airtable, the “pattern” are presented with…

Will never work using the “Find Overlap” as presented.

This is because you can only choose data from one input or another. However…

You MUST have the CSV (the updated data) as it is needed to make the updates.
You MUST have the Source (Airtable/Bubble) as it is needed to match to the unique id on PATCH

So it is impossible to get this to work as stated.

What you have to do, and I am open to alternate solutions, is recombine the output of the Overlap with the Source database so that you have all the required fields.

This then means you have duplicate fields with (1) after them in the combined output.

There is a Parabola bug that makes this also fail unless you force the names to be different.

Bug with Bubble PATCH

There is a bug with the Bubble “PATCH” implementation. If you need to combine two tables (and you really do) to do a PATCH from an input that is updating rows on the Bubble table, then you will most likely have two fields named the same thing. So DATE from Bubble and DATE (1) from a Google Sheet (as an example). When you PATCH Bubble’s DATE field with {Date(1)} (chosen from the list provided) it will Ignore that you want the updated field and always take the field from Bubble (so never update…

Is there another way, as this is very clumsy?

[daniel]

Hey @NigelG,

Thanks for pointing that out. I just replied to your thread regarding the potential bug. Please keep me posted on that front.

Here is another pattern that could work. Feel free to adjust to your different export types. However, any columns from your updated data source that share the same column name as your source data will be appended with a (1).

Lastly, are you receiving any errors from Bubble, or is it sucessfully updating using the wrong date column?