[Scott Barrett]

Charlie Bursell wrote:

You do not need to define Directory Parse/Delete routines to get file name if Fileset/Local

Unless changed the message metadata will always contain it

I posted this before:

set klst [msgmetaget $mh DRIVERCTL]

Or, if in Xlate

set klst [xpmmetaget $xlateID DRIVERCTL]

# Always set a default

set filepath “”; keylget klst FILENAME filepath

filepath will contain full path.

[Scott Barrett]

This is a very common flat-file format.  I have it in several varieties as well, from several different hospitals.

They are just groupings of records that make 1 patient record, and all lines in the file are LF for UNIX, and CRLF for dos.

Essentially the previous method of programming used 01 (in this case) to indicate that a new patient record is beginning (much like MSH indicates a new message)….and also qualified each individual record with an account number in each.

Based on your last suggestion, I tried to convert all LF to CR, but that didn’t improve things.  I think I’ll write a pre-pass script to place a special character before the 01, then remove all LF, then change the special character to LF.  I think this will put an end to it.

[Scott Barrett]

It’s on a UNIX system, so there is a line-feed for each segment/record.

01……

02……

04……

01……

etc

Do you think that converting the line-feed character to something else before processing will make a difference?

[Scott Barrett]

Thanks Jim.

I got a little closer….

I set up the HRL as you suggested and set testing tool to read EOF instead of EOL, and defined the carriage return character into each FRL. This got it to read properly through the first complete recordset, but it won’t pick up the next 01 and so on.

[Author]

Replies