Hash matches with collisions

I'm trying to improve certain matches that happen often, on many many dimension columns. So my choices as I understand are: 1. Match on many columns 2. Hash the columns, store the hash, match on that. If I use something like 'checksum' is there a known algorithm to handle the collisions? So the match would work like: 1. Partition two tables where checksum hashes match. 2. If there's more than one row returned, filter the results by matching all the dim columns Anyone tried using checksum to make hashes faster on a large number of columns? I should say, I'm looking for a SQL/TSQL solution. I can easily acheive this match with python or c#