This is request from STrategic customer T-Mobile. They point out that there is metadata in the parquet files already, so why do they have to spend compute to count all the rows to get a simple count of rows. More details:
Regarding the rowcount functionality request for Parquet in ADF by T-Mobile, I followed up with the customer and they advised as follows:
“The count should be instant because it is in the metadata of the file. right now you have to run a select count from the serverless endpoint. so counting the startup time of synapse serverless could be 3- 5 min”
As a reminder/context, your guidance was for him to not ask for a specific technical implementation, but to ask for a functional requirement/sla.
There is already an open feature request, but we were hoping that this could help to prioritize it.
ToddChitt
on:
Copy Job needs more functionality
on:
Support for Signed GitHub Commits
on:
Bug: Web URL Conditional Formatting Ignores "Show ...
on:
Apply Conditional Formatting to All Columns/Series...
yicw
on:
In Fabric view, let us assign workspace colors bas...
on:
Introducing Excel-Style Table in Power BI