Re: Need to know CISAM definition to create schema

Jonathan Leffler <jleffler@earthlink.net> wrote in message news:<w9v0d.12721$w%6.7673@newsread1.news.pas.earthlink.net>... > Sanjiv wrote: > > I have an old application written using CISAM.Our goal is to convert > > all data into Oracle tables. Unix admin is going to extract data into > > ASCII files. I'm planning to use SQL*Loader. When I log into unix, all > > I see is a bunch of idx files, .dat files, udat files and .h files. > > Being new to all this, I have no idea how can I get CISAM file > > definitions to create schema or Oracle tables? Other than looking > > into each data file manually. Also we don't have Informix SE, which I > > think would have helped. > > > > Any help will be appreciated? > > As you've probably guessed from the lack of response, there isn't any > easy answer. > > Given the absence of any schema information, you're likely to have do > the reverse engineering yourself. You can use bcheck to establish the > row size and which parts of the row have which types (by looking at > the index specifications). After that, it's down to eyeballing binary > data dumps. Make sure you get both the hex and the ASCII listing. > > Once you know what the schema is, you can consider whether it is worth > using ISEXTRACT from the IIUG Software Archive to unload the data. > > Beware variable length records - bcheck should warn you if they are in > use. ISEXTRACT has not got the first clue about how to deal with them. Thanks Jonathan. Finally some response. It is definitely helpful. I will look into ISEXTRACT.