We have a table look and we need to have action implementation on that , need help here

[SachinS]

We have a table look and we have around 13 columns in that table look and on that last 3 columns client will come and put the data and we need to capture that data and update the biquery table back and also use this data for some ML work , but in the current approach I am  using data action and mentioned  the cloud function url  but since we have lot of rows in the table look , if we add this clound function url on all these columns it will not be best for scalability and production , can you let me know is there any optimized way for this

[andy4]

My opinion is that the Looker UI is not optimized to write back to your database. You can use the Looker action feature to accomplish this, but the end user will need to refresh the results once the write-back is completed. In addition, modern end users expect that write-back interaction to resolve more or less instantaneously (which an API call and a manual refresh is not), and expect features like bulk updating multiple fields at once (definitely not possible with a looker action). If you are hitting issues here, I recommend creating a custom front end which pulls data from Looker to do an inital load of a web-based data grid, then let the end user update that directly - probably with the ability to periodically send their updates to the backend source of truth.