In our application we want to generate pairs where all the following are true:
Classification matches exactly
Name is close
Address is close
However if we add all three of these at once, we get the union of them all (and the classification exact means 45M pairs). It would be lovely if there was an intersection flag we could pass in, rather than our current approach of having to generate the three sets of pairs separately and then using MultiIndex.intersect() on them all.
In our application we want to generate pairs where all the following are true:
However if we add all three of these at once, we get the union of them all (and the classification exact means 45M pairs). It would be lovely if there was an intersection flag we could pass in, rather than our current approach of having to generate the three sets of pairs separately and then using MultiIndex.intersect() on them all.