Hi, I have a text-based, variable record length database (a set of bibliographic references) of approx. 1 megabite and growing, for which I'm creating a command line dbms with perl. Could anybody here tell me if it's going to be quicker to address the data via an ancillary fixed length (text-based) "pointer" file, or with a simple scan of the bibliography file? The pointer file will store the starting position of a citation (each has a number, the citation number) in the citation file in a line corresponding to the citation number. When a request is issued for a particular citation (the citation number is supplied by the user), the perl application will look at the appropriate spot in the pointer file (i.e., seek to citation number * fixed-length of pointer file record), draw from there the starting position and length of the citation in the variable length citation file, then seek to that position in the citation file, read for length, and return the citation to the user. I suppose it all depends, but could anybody here address this helpfully? thanks, Eric Lawson -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
