ChristianRRL
Honored Contributor II

Quick follow-up on this @Retired_mod (or to anyone else in the Databricks multi-verse who is able to help clarify this case).

I understand that the proposed solution would work for a "one-to-one" case where many files are landing in a specific dbfs path to be ingested by auto loader into a specific bronze table. OR another similar case where we have "many-to-one" where we are expecting different data from different sources going into a single bronze table, but critically all the different sourced data would have the same format (i.e. all different source data have the same column length/type/order and it's cleaned up in a single bronze raw location).

In my case, what I'm asking about is if auto loader supports a "many-to-many" case (i.e. "many source database_table.csv-to-many target database_table_bronze"), or else what is the suggested Databricks approach to this? In this context, I mean that we have many source locations with many different kinds of tables being loaded into a single "dropzone" location. From this single "dropzone" location, I would expect that either auto loader should natively support my use-case, or if not that there should be a fairly standard approach to dealing with this.

I'm not the only person in the world dealing with this, and I found a similar example of someone else posting this on Stack Overflow: https://stackoverflow.com/questions/69572265/ingest-several-types-of-csvs-with-databricks-auto-loade.... The proposed solution for them was to use "pathGlobFilter", however I believe this only handles the first two cases and not the "many-to-many" case I'm posting here.

For additional context, here's the code I have so far below:

ChristianRRL_1-1702920099476.png