Shell Programming and Scripting

I want to develop an application. Simply,

That is it. I have got suggestions to do this with Perl CGI. (my system has up to Perl only - no Python)

Is Perl CGI a better way to do it? Or are there anyother alternatives that are more effective than CGI?

PHP looked nice for the first three, but then you got into batch. Maybe put that file into a DB table, process it and select out the new file? Can the unix server push the file so it gets processed and returned, perhaps with a web query?

I don't think PHP would be good for me.. Since the system that I am working do not have any PHP codecs

We are talking about a huge file here more than a million records. Loading that into a temporary table is worthy when I have lot more functions to carry on with the data. All I do here is a simple comparison against a column and prepare the new records based on it.

Any other suggestions are welcome too

Some RDBMS allow you to map the file as a table - the data has to get over the rdbms wall somehow. I did it in Interbase/firefox using UDF with a SP.

PHP is open source code for the web server, similar to ASP/JSP.

DG.. Thanks for the interest shown..

I am working with Informix with too many restrictions. I am sure that I am not allowed to install Interbase/firefox in my system (working system not my own.. company's )

That is why I had to think about Perl CGI.

Well, no point in bothering the web server for the batch, normally. PERL can talk to RDBMS as a client, running a query per record. Mapped in files are virtually inside the DB engine, allowing them to be worked in SQL and efficiently related to other tables inside a single, big query. Google "INFORMIX external file as table", and voila: Using data file abstraction with external tables in Informix Dynamic Server- You want to write a query that goes through the file one time relating records to a join of db tables, and returns a cursor of augmented data to your client. You need a robust client setup, not something dopey like some odbc/jdbc cursors that query all the result into VM first so you can walk backward, forward, etc. You need a non-reversible cursor that just flows the data out through your client code immediately as it arrives.