Talk:Hash join

I've downgraded to a draft task for now; we need more implementations (preferably independent ones!) so that we can be sure that a sufficient common understanding if the task exists. Once there are four language implementations, we can promote back. Also, what about language/runtime systems with built-in hash table support? –Donal Fellows (talk) 22:47, 30 November 2013 (UTC)

Agree on the draft status for now. And despite the, er, fireworks accompanying its creation, I suspect it'll be a good task. I don't see how the built-in-ness of hashes plays one way or the other. The task merely assumes that an appropriate hash implementation will be used, whether built-in or imported or implemented as part of the solution. --TimToady (talk) 04:05, 3 December 2013 (UTC)

And it is now a full task, with 5 implementations (right now). I'll do a bit more editing (adding in a print requirement and giving test data) but that won't break the existing implementations. –Donal Fellows (talk) 10:48, 12 December 2013 (UTC)

clarification needed

Is Popeye's entry to be not listed, either because it has no nemesis, or because it wasn't in the 2nd relation list? -- Gerard Schildberger (talk) 00:41, 17 December 2013 (UTC)

Can the nemeses (plural) be listed on one line (as the REXX example shows)?   To me, it looks cleaner, more succint, less screen (output) clutter. -- Gerard Schildberger (talk) 00:41, 17 December 2013 (UTC)