CISAM .idx structures

Is there any published tech specs on the CISAM idx structures for CISAM v 7.2.X? We are having issues with some of our reads/writes to the .dat files and we wanted to do some low level investigations on the indecies. If we can't get that some info on what typically is held in an idx field? Thanks in advance Mike

Mike wrote: > Is there any published tech specs on the CISAM idx structures for > CISAM v 7.2.X? There is the manual for C-ISAM. And in the manual are the appendices. And in the appendices is the information that is published. It is pretty much the best information there is on the subject. > We are having issues with some of our reads/writes to the .dat files > and we wanted to do some low level investigations on the indecies. Errr...what do you mean? Reads occur using the index you select - or sequentially if you don't use any index. Writes to the .dat file happen when you add or update a record (or delete, I suppose). That's the beauty of using C-ISAM - and also the pain of using it. > If we can't get that some info on what typically is held in an idx > field? It depends on the options you used when creating the index, but the basic information is the key value and one (or more) rowid values which have the same key value. That applies strictly to the leaf nodes; at the root and twig nodes, the information is a bit different, but still similar (a key value and where to find more information). There are some issues related to massively duplicate indexes (eg an index on gender alone on a table with millions of records), and to leading and trailing compression. There are complications if you're using variable length records - the variable portion of the record is stored in the index file. -- Jonathan Leffler #include <disclaimer.h> Email: jleffler@earthlink.net, jleffler@us.ibm.com Guardian of DBD::Informix v2003.04 -- http://dbi.perl.org/