Re: Index "Fillfactors"

Dave Kosenko wrote: > > Greg Moye wrote: > > > Dave - I think I have (on)checked right after dbimport on 7.10.UC1 and 7.12.UC1. > > Ah, and now you provide the critical piece of info that had been missing! > > Dbimport will build the indexes FIRST, then load the table. Since FILLFACTOR only > applies at the time the index is built, you are not going to see any difference in > how sparse the leaf nodes are packed. > > Try rebuilding one of the indexes after the dbimport and running the oncheck. I'm > sure you will find a very different use of space. > > -- > Dave Kosenko, Informix Professional Services > snip.... EGAD...are you serious?? Why in the world would it make any sense to create the index first, then create and load the table data??? Since the "Create Index..." comes after the "Create Table" in the SQL file, I assumed that's the order it happened in. It also looks that way when Dbimport is running. So basically, dbimport is working in a basackwards fashion, and lying about it at the terminal. Yuck. So this means I have to edit out all of the create index statements and manually reconstruct them to run seperately, to keep dbimport from screwing up....could this be any more obnoxious?? Further, it would appear that I would have to edit out all primary key constraints in the "Create Table", and convert them to "Alter Table" to be executed later to get the primary key "fillfactored" right. Sorry Dave - I realize it isn't your fault - but I needed to vent a little....thanks for the info - even though it really wasn't what I wanted to hear. -- Greg Moye Electronic Data Systems Corp. - Plano TX