2 ms·
My first guess is that if you always store the full string you don't need to scan the database to see if you already have the same string. Essentially you choos
by Etheryte 3y ago
My first guess is that if you always store the full string you don't need to scan the database to see if you already have the same string. Essentially you choose to use more space but reduce load. Regardless of whether you do the string deduping on inserts or async later on, you have to do it at some point and the unpredictable performance overhead might be undesirable.
- cm2187 3y agoWell it should be a dictionary lookup, it should be pretty fast and predictable. And for freeing it up, it should be a good candidate for reference counting.