F98741 TAM file sequence

tkivinen

Member
I read in a few other threads about "select * from pd7334.f98741" being used for the spec/TAM file build phase of package builds. Is that the full query, or is there an "order by" clause?

We have modified a few objects rather than making new "55" versions (P4210 namely), and I'm wondering if we are experiencing fragmentation of the TAM files. P4210 seems to take longer to start. Two thirds of our ERP8 central objects TAM tables were delivered (except F98710-98713 and F98306) with non-clustered indexes which will cause the data to be returned in physical order, rather than ordered.

First is this an issue? Is the flat-file XDB index built the same as the sql index, and is there much penalty for a non-sequential read of the TAM file?

And second, is there any harm in converting one of the indexes to clustered? The F98741 primary key is ESEVSK + ESEVSEQ, and the second index is ESEVSPEC + ESEVSEQ. Is this usually queried by ESEVSK?

Thanks for any help.

Todd
 
Back
Top