Page 1 of 1
Improvement suggestion for copybook filter.
Posted: Tue Oct 13, 2009 7:33 pm
by mauros64
I have many situation in which I need to convert a very big and complex mainframe file using the copybook filter.
It work fine but when I need to convert only a limited subset of the available fields is far more faster to delete the unneeded byte range and then use the filter on a copybook modified and restricted to match only the remaining fields.
It will be very useful to let the user specify in a copybook only the needed subset of the fields (for example putting an X character in the first column of the unneeded fields in the copybook, or in some other way) and do the conversion work only for the specified subset.
Thanks
Re: Improvement suggestion for copybook filter.
Posted: Wed Oct 14, 2009 10:19 am
by DataMystic Support
That sounds like a lot of manual effort to edit the copybook! - and what if it changes?
Why not a filter that removes fields based on column number or by field name, with commas or tabs between each name?
Would that work for you?
Re: Improvement suggestion for copybook filter.
Posted: Fri Oct 16, 2009 1:59 am
by mauros64
It sound better to remove field by field name or number.
But ...
If it is after the Copybook that file has already converted 60+ fields ( EBCDIC, Packed Decimal , etc...) just to pass it to the discad filter that keep 8 fields.
If it is before the Copybook it must know the Copybook format.
so to have a good performance (whithout the same knoledge of the copybook filter) it must be inside the Copybook filter.
I usually manage 1-2 GB data files, and discard not needed field is a big performance improvement.
Now I'm using a discard first ( with manual calculation) approach, it works but is a tedious work and error prone.
It's just a suggestion for the future, maybe other people will like it to.
Regards.
Re: Improvement suggestion for copybook filter.
Posted: Fri Oct 16, 2009 12:25 pm
by DataMystic Support
Thanks, I understand what you mean.
Ok, how about a field within the copybook filter that has options to
1. Only keep named fields
2. Discard named fields
3. Only keep numbered fields
4. Discard numbered fields
- so you can specify fields by name or number.
Then this is followed by a field containing space, tab or newline separated field names or numbers. How does that sound?
Re: Improvement suggestion for copybook filter.
Posted: Fri Oct 16, 2009 4:36 pm
by mauros64
It sound very complete and usefull.
It will be easy to use and it can have grat performance too.