Multiple Record Types

[ghutchin]

I've got a file that has multiple record types, each having a header with a record type at the beginning of each row. Each record type has a different record length.

My problem is that the file is on the Mainframe and need to process this file on our Unix server and the FTP stage is having an issue pushing this data to the server. Beyond that, even when the data is FTP-ed via a script initiated on the DS Server, there is no CR/LF at the end of each row. I've tried reading the file with both the CFF and the Sequential file stages and each needs either the metadata to describe the record length or describe the line delimitter.

Anyone run into anything similar to this? If so, please let me know how you got around this.

Thanks,

Gordon

[rwierdsm]

We are doing something similar here at our site.

Our Universal Copy software takes care of translating from EBCIDIC to ASCII and puts the proper CRLF in the data.

We created COBOL copy books to describe the various embedded record types and imported these definitions into the CFF stage. The first three characters in each row (sounds like you have something similar in your data) is used to direct each record type into its own stream (use the selection criteria tab in the CFF stage)

HTH,

Rob W.