fips encoded as an int

[TomGoBravo]

The DataFrame returned by the client has fips represented as an int64 but my understanding is they are better represented as 2 or 5 character strings. A quick fix that seems to work for me is df.fips = df.fips.apply(lambda v: f"{v:0>{2 if v < 100 else 5}}")

[cc7768]

Hi @TomGoBravo -- Thanks for pinging us.

I'm not particularly opinionated on this -- I believe we chose to use integers so

1. We could do comparisons like fips < 100 => states or (fips > 6000) and (fips < 7000) => all CA counties (There could be other ways to do this as well though)
2. It's a little less data to pass from database

Am I leaving anything out @sglyon?

Unless we have a more compelling reason than (1) or (2), I'm not opposed to just changing them at the database level which would make this change happen globally.

[sglyon]

My preference is still to store them as int.

@TomGoBravo would it be helpful if we had a keyword arg on the client that is something like fips_as_str that would apply that transformation to the fips column on each request before returning the data frame?

[TomGoBravo]

This is no big deal either way to me, just bringing up something I noticed.
In code I work on I'm trying to stick to an opaque string to identify a region and factor out logic that relates the regions to each other without depending on the structure of the region identifier.