Read the smaller file line by line, and store the relevant entries in a hash table. That's what I meant with "store in memory".