Fastapi: [QUESTION] Support for concurrent dataframe operations with async await

[scheung38]

Support for concurrent Panda Dataframe GET access

Does FastAPI support concurrent Dataframe read/GET access to various different columns for validation checks, where I am not expecting any dependencies between the columns.

i.e after df is created, instead of sequentially calling

check_for_NA
check_account_types
check_duplicates

how to async trigger these above and collect results are ready and perform a AND function

Environment

• OS: macOS: 11.1 Big Sur

• FastAPI Version fastapi 0.63.0:

  (.venv) pyth3 --version
  Python 3.9.1

[Mause]

Firstly, Fastapi does not have any integration with pandas.

To answer your question though, you could run those operations in a thread pool, and perform the AND once they all complete? I would be surprised if this actually improved performance though

[scheung38]

@Mause thanks I know there is no integration with Pandas, but based on the async await nature that FastApi handles I thought I'd ask to see if there are benefits, if so how to approach it.

[Mause]

Pandas doesn't have async support, so your only option here is to use threads, though as I said, it's not guaranteed that this will give a performance improvement

[hellocoldworld]

This is not related in any way to fastapi, but it's an interesting question. I would encourage you to post this question on a site like StackOverflow