No rows process through job

tonystark622 · Post by **tonystark622** » Fri Jun 20, 2003 11:22 am

Ah, so there is a way to build an index encompassing all the fields that I need (composite index)? I figured that there had to be. I'll dive back into the Universe documentation. Maybe in the meantime someone will post back with the syntax.

Thanks Craig.

Tony

chulett · Post by **chulett** » Fri Jun 20, 2003 12:40 pm

Actually, if you let your job create the hash and mark the fields you want to be in your 'primary key', it will create the hash with a composite primary key (assuming that is what you want). You can verify that it did what you wanted by doing a LIST DICT HASHNAME in the Administrator and looking at the definition of the @KEY PHrase record.

What I don't have handy is the syntax to create a composite index on non-pk fields, hopefully someone will post it.

-craig

It's been a long week... I'm hoping that I'm thinking correctly with PK = Unique Index for a hashfile like in a RDB. If not, I'm sure Ray or Kim or *someone* will happily point out the error of my ways. [:I]

tonystark622 · Post by **tonystark622** » Fri Jun 20, 2003 1:38 pm

I've lead such a sheltered life. I never knew about the .NOTFOUND thingee. I've always used IsNull() or Not(IsNull()).[:)]

Tony

chulett · Post by **chulett** » Fri Jun 20, 2003 3:29 pm

The .NOTFOUND status is *fairly* new and was somewhat problematical at first, meaning it didn't always work as advertised. [}:)]

Checking a key field for null or not null (as you noted) was the Original Recipe way of checking lookups and never gave us any problems that I recall. I believe that nowadays the Extra Crispy way works too. By 'works', I mean 100% on all platforms kind of 'works'. [:)]

-craig