topic Autoloader update table when new changes are made in Get Started Discussions

Autoloader update table when new changes are made

anonymous_567 — Wed, 17 Apr 2024 14:15:03 GMT

Hello,

Everyday a new file of the same name gets sent to my storage account with old and new data appended at the end. Columns may also be added during one of these file updates. This file does a complete overwrite of the previous file. Is it possible to configure autoloader so that it does a full scan of the changes and updates the delta table only where changes exist?

Below is an example of the file updates I am referring to
Day 1 file: A
Day 2 file: AB

Day 3 file: A'BC

*each letter represents new data and
**a letter followed by an apostrophe is altered data

Re: Autoloader update table when new changes are made

Ajay-Pandey — Wed, 17 Apr 2024 14:22:29 GMT

Hi @anonymous_567

With the help of autoloader, you can not do it as autoloader work at the file level, not the data level, to achieve this you can go with SCD or CDF approach.

Re: Autoloader update table when new changes are made

anonymous_567 — Wed, 17 Apr 2024 14:52:07 GMT

What if the only changes made to the new file being ingested is new data appended at the end of the file?

Re: Autoloader update table when new changes are made

data-grassroots — Wed, 17 Apr 2024 19:31:01 GMT

This may be helpful - the bit on allow overwrite

https://docs.databricks.com/en/ingestion/auto-loader/faq.html