Hi all, I’ve written a little middleware app to handle a one way sync between a project management suite and Coda. It’s been working wonderfully so far, however, I’ve noticed that when I ‘upsert’ rows in Coda via the API the column type will change from a text or date field to a drop down for some (but not all!) of the columns.
While this isn’t a severe problem, it’s a little annoying. I couldn’t find any information on updating column types in the API manual. Is there a way to specify a column type when making an upsert, or get the API request to respect the existing column types?
Hmm, that doesn’t seem like the expected behavior. Are you able to reproduce the issue in a fresh table? Any idea what’s different about the columns being flipped?
One of the app’s function is to delete all of the rows in the table and insert 400ish rows to refresh the table to make sure all of the data matches, and I’ve tried it with two different tables on different docs. It seems most likely to happen with columns that have only a few different possible options (i.e. columns where the content is either ‘ready’ or ‘not ready’ for all of the rows). I’m fairly sure that I’m only passing strings and numbers to Coda, but I’ll double check to make sure.
Also worth mentioning that I’m using Node and Express, along with coda-js.
Hi Ethan! Can you try using the disableParsing url parameter: https://coda.io/developers/apis/v1#operation/upsertRows
This should disable any type inference on previously-unfilled columns.