Re: get data from mainframe files, ADA

Billy Verreynne wrote:

> johnbishop_at_atl.mediaone.net wrote in message
> <7korhd$j9e$1_at_nnrp1.deja.com>...
>
> >The best (and cheapest) we've come up with is
> >
> >1) cron calls FTP to move the data
> >2) PERL transposes data (if need be)
> >3) SQL*Load data in.
>
> Sounds good. We did it pretty much the same way, but without using PERL.
> Data formatting was done in COBOL on the mainframe (modifying existing data
> extraction programs).
>
> The only addition we had was using pipes. This pretty much speeded up the
> whole process a lot.
>
> Something like:
> 1. create pipe 1 and start SQL*Loader reading from the pipe
> 2. create pipe 2 and start dd to convert EBCDIC to ASCII, reading from
> EBCDIC from pipe 2 and writing ASCII to pipe 1
> 3. fire up FTP with a binary transfer of mainframe data and download the
> data into pipe 2
>
> >The next best thing we came up with was something like an EDA/SQL
> >gateway.
>
> We looked at a similar option, but these mainframe SQL gateways were -very-
> expensive.
>
> >I'm open to just about anything.
>
> The pub around the corner and you pay? ;-)
>
> regards,
> Billy

If you have a lot of data, you shuld not ftp. It is faster to use pipes, and even faster if you have access to CLIO/S (IBM product).

--
Ron Reidy
Oracle DBA
Reidy Consulting, L.L.C. Received on Fri Jun 25 1999 - 07:58:44 CDT