Re: C-ISAM Maximum File Size?

From: Jonathan Leffler <jleffler@earthlink.net> > >Robin Grayson wrote: > > > Does anybody know what the maximum file size is for a C-ISAM set of > > data files? > > > > We currently have a data file approaching 2GB, with an index file not > > far behind. > >My esteemed but Obnoxious colleague is behind the times. So, no change there, then.... >You can actually achieve >file sizes somewhere in the region of 2^45 bytes (give or take >a couple of factors of two). If you have exploited the variable 2^45 ~= 35GB? >There is not yet a 7.25 SE; we're thinking about it. The C-ISAM >datablade does not currently exploit this technology. And there >are cogent arguments for saying that you should aim to avoid 2 GB >data files, related to the depth of the B-tree and the number of >index nodes that have to be read to do random access by primary >key on the files. One item up for consideration is another new >version of C-ISAM which would not be index compatible with the >current version, but which would increase the index node sizes from >512/1024 bytes up to say 4/8 kB. This would dramatically decrease >the depth of the B-tree, at the cost of compatability. That raises >conversion and reversion issues, amongst other things, and dual-mode >working (with both smaller and bigger index node sizes). <WOODEN SPOON> And of course, Informix could consider a similar (but more challenging) break with the past to allow bigger chunk sizes in IDS? </WOODEN SPOON> :0) ______________________________________________________ Get Your Private, Free Email at http://www.hotmail.com

On Wed, 19 Apr 2000 09:13:38 BST, Obnoxio The Clown <obnoxio@hotmail.com> wrote: >And of course, Informix could consider a similar (but more challenging) >break with the past to allow bigger chunk sizes in IDS? *standing ovations* Bye, Olaf -- A long december, and there´s reason to believe, maybe this year will be better than the last... [Counting Crows]