Isn't it the other way around? I believe I've read in a book that hash join goes like that:
...first, a single pass through the file with FEWER records... (partitioning phase) ...second, a single pass through the "other" file... (probing phase)Bold text
- Yeah, you are right. I'll make some changes to this when I get some free time (hopefully later tonight). bDerrly 13:03, 10 March 2006 (UTC)
- Fixed this problem (hjzla, March 18). Also added some more info.
I read this and I don't understand what a hash join is. It needs to be written more clear. ie. I was expecting to see a table structure, SQL, and an explanation why a statement is a hash join:
The "Classic hash join" section states "This algorithm is simple, but it requires that the smaller join relation fits into memory, ..." Shouldn't it be something like "This algorithm is simple, but it can loose its performance benefit if the hash table doesn't fit into memory..."? Randallbsmith (talk) 21:14, 11 August 2011 (UTC)