I've seen many times the Lookup stage designed with Sort on both links, the input and the reference link , that I'm wondering if the best practices changed on this one !
I'm talking about a Normal Lookup with the Input Hash Partitioned on Keys and the Reference Entire Partitioned.
What would be the need of sorting the data in this case ? I heard rumors the Lookup would be faster if the data is sorted !
Any opinion of that ?
Lookup design best practices
Moderators: chulett, rschirm, roy
Lookup design best practices
Thanks,
Emma
Emma