We have some IBM mainframe programs written in COBOL that build flat files. We schedule JCL that runs the COBOL program and then FTPs the flat file down to a server. We point the QlikView script at that file, and read it in like any other text file. If the numbers are packed or otherwise not in display format, some additional manipulation may be required. But if you're in charge of the files themselves or their specification, just specify that they be in display format for simplicity.
Our preferred approach is to point QlikView directly at the source data, most of which is stored in DB2 databases. But we have some older IMS databases with no ODBC connection, and in some cases, it makes sense to do some of the data manipulation on the mainframe rather than in QlikView.
