3 ms·
Hi, I'm curious how you deal with the potential for hash collisions across a large data set - is that a post-join check?
by rkerno 3y ago
Hi, I'm curious how you deal with the potential for hash collisions across a large data set - is that a post-join check?
- puzpuzpuz-hn 3y agoHi, if you're asking about the hash table itself, then currently we use linear probing, i.e. k/v pairs with a collision are inserted sequentially starting with the hash%capacity index.