This is what I found worked for me. As I mentioned previously, I found that you must already have created a .ARD file and a .IND file (containing no index data) with the same name.
Application Group INVOICEAG contains 3 fields, NAME, INVNUMBER, RPTDATE with format YYYYMMDD.
Application INVOICE is defined as Generic Indexer with data type PDF.
These files exist in my directory:
INVOICEAG.INVOICE.TESTNAME.987654.20190221.ARD
INVOICEAG.INVOICE.TESTNAME.987654.20190221.ARD.IND
INVOICEAG.INVOICE.TESTNAME.987654.20190221.ARD.PDF
The content of the .IND file are:
COMMENT:
CODEPAGE:37
COMMENT:
GROUP_OFFSET:0
GROUP_LENGTH:0
GROUP_FILENAME:/ascent/Scan02/INVOICEAG.INVOICE.TESTNAME.987654.20190221.ARD.PDF
This arsload command worked to load a specific file:
arsload -h ONDENU -v -G MVS -A JOBNAME -b 1,NAME -b 2,INVNUMBER -b 3,RPTDATE -B MVS.JOB.1FLD.2FLD.3FLD /ascent/Scan02/INVOICEAG.INVOICE.TESTNAME.987654.20190221.ARD
Joe