pilgram | A python library for instagram filters | Computer Vision library

If you're using a mainframe COBOL compiler, you go to the documentation and select your version. Then you do a search for "file status key" and review what file status 46 means.

A sequential READ statement was attempted on a file open in the input or I-O mode and no valid next record had been established because:

• The preceding READ statement was unsuccessful but did not cause an at-end condition.
• The preceding READ statement caused an at-end condition.

Note that in your OPEN-FILES paragraph you PERFORM READ-NEXT-RECORD and then fall through into the rest of your code after already having reached end of file.

You probably want a STOP RUN or GOBACK at the end of your first paragraph.

Edit regarding printing record count: There really isn't a good way to have the record count appear at the top of the report because you don't know the record count until you've read the entire input file but you're printing the report lines as you go. Most of the time control totals like record counts are DISPLAYed (which by default goes to the SYSOUT DD) and the report(s) go to a different DD defined in FILE-CONTROL (via a WRITE, just like what you're doing).

Second edit regarding printing record count: As @GilbertLeBlanc points out you can store your output lines in a table until you've read all record in the input file. You do have to have enough space in the table to handle all the output records, and there are a number of different ways to do that.

• Your table could be statically defined with a large enough OCCURS clause to handle what you've been told is a reasonable number of records. This used to be very common, and there would be code to check if the reasonable number had been exceeded and abend if it was.
• Your table could be variably occurring with the UNBOUNDED phrase, subject to its limitations, and storage managed with the ALLOCATE statement and FREE statement.
• You could roll your own allocation and reallocation with LE Callable Services CEEGTST, CEEFRST, and CEECZST.

Gilbert also points out you can read the file twice, once to get the record count, then close and reopen to do your normal processing. This will work so long as you're not doing something tricky with your JCL like...