I have a vague idea of:

1) define catchment boundary
2) feed in .csv containing identifier + postcode
3) for each line, look up postcode and return one of (in catchment, out of catchment, too close to boundary to be sure)
4) Feed this back to .csv
5) Manual check on postcodes very close to boundary, since one end of the road may be in, but the other not.

Anyone built / seen something like this? It'd be for a few hundred queries at a time.