I have a problem. Let med describe it.
I have 300 of ½" older (20 - 25 years old) tapes with governmental
data, all with the same structure.
Ebcdic, IBM-labelled, IBM variable.
The variable text files (tables) consist of different records, all in
the same files.
There are nine different physical record types with different (fixed)
record organisation, five of which include packed fields in different
locations.
Field with record type indicator (01..72) is located in the beginning
of each record, so every record type must be converted according to
this indicator and corresponding record description.
Similar situations have been rather common in the older days, but this
seems to be extra complex.

The converted recods are to be written into a pc Windows disk file in
the original order for further treatment.

I use my own ebcdic to ascii translation table.

I can solve the problem with what I know now, but it involves reading
each tape six times with different conversions and merginging the
correct converted records in right order into a file, which is a lot
of work, and tear and wear on the old tapes.

NovaXchange support have given up on this one.

We use also an older tape reading program from Vogon that can do
things like that, but there is a limit to the number of involved
fields, which has been exceeded here.

We have also an EBCDIC package from Vedit, but...

To summarise :
Convert ebcdic data on tape in variable format with packed numerical
fields in different places for different records.

Can you help ?


Yours

Jurek / RA