search

jrs0 / hbr_uhbw

Development of high bleeding risk tools for UHBW
GNU General Public License v3.0
0 stars 0 forks source link

Clean up the primary care attributes column processing #7

Open jrs0 opened 5 months ago

jrs0 commented 5 months ago

Currently, the data is fetched in icb.py which lists the columns that go into the select statement, and then there is another long list of flag columns in from_icb.py that require turning into bool. These lists should be in the same place.

jrs0 commented 5 months ago

Discovered that bool is a bad idea for the flag columns, because down the line it is not possible to store NaN. Going to use an integer type that can store NaN, and then use floats for proper numeric columns. This will allow them to be distinguished later on.