[Feature Request] Providing a cache option for `pd.read_sql()`

Providing a cache option for pd.read_sql()

Is there any work being done to allow a cache mechanism for pandas sql queries? I typically connect pandas to a sql database in a jupyter notebook. However, when I re-run cells it would be nice if these queries were cached in a file unless cache busting is explicit.

query = "SELECT * FROM table WHERE %(test)s = 'test';"

df = pd.read_sql(query, db, params={
    'test': 'name'
}, cache=True)

I would imagine the behavior would that the query string, db engine connection string, and params are hashed and saved as a file in some directory.

[Ouwen]

@gfyoung sounds good to me!

[Ouwen]

For my own purposes I am just going to create a function wrapper outside of pandas. It would simply be provided a directory to save files given some query, params, and database connection string.

However, my thoughts are that this problem is common enough to developers and researchers that it would be convenient if the pandas library provided support.