Need help in understanding the join mechanism

Post questions here relative to DataStage Enterprise/PX Edition for such areas as Parallel job design, Parallel datasets, BuildOps, Wrappers, etc.

Moderators: chulett, rschirm, roy

Post Reply
zulfi123786
Premium Member
Premium Member
Posts: 730
Joined: Tue Nov 04, 2008 10:14 am
Location: Bangalore

Need help in understanding the join mechanism

Post by zulfi123786 »

What happens when a join is performed on pair of keys which donot have the same metadata, i think for a join to be performed only the column name should be the same and the metadata of the keys can be different. For Example a join is performed on column "KEY" which is char 10 in left link and char 20 in right link, does DS perform implicit conversion from char 10 to char 20 so that the join can be performed??? if yes does it pad spaces??? what happens when char is joined with decimal???
ArndW
Participant
Posts: 16318
Joined: Tue Nov 16, 2004 9:08 am
Location: Germany
Contact:

Post by ArndW »

DataStage will do an implicit conversion, but it is much better to ensure that the metadata is identical and do an explicit conversion. Char() field conversion are always blank padded. I don't know the rule for conversion, perhaps DS uses the left link as the original and converts the other links to match that datatype.
Post Reply