RE: Replace 1 Char with 3Char's

["Dennis Lovelady" <iseries@xxxxxxxxxxxx>]

> PGM
> DCL        VAR(&X0D) TYPE(*CHAR) LEN(1) VALUE(X'0D')
> DCL        VAR(&X25) TYPE(*CHAR) LEN(1) VALUE(X'25')
> DCL        VAR(&SEDCMD) TYPE(*CHAR) LEN(100)
> CHGVAR     VAR(&SEDCMD) VALUE('sed +
>              "s%[\][.]br[\]%\'||&x0D||'\'||&X25||'%g" +
>              /joe/temp/in.hl7 > /joe/temp/out.hl7')
>
> QSH        CMD(&SEDCMD)
> ENDPGM
>
> This replaces \.br\ with a crlf.

This essentially would have been my solution, except that I would not have
defined &X0A and &X0D.  Instead, I'd define &CRLF as len(2) with value
X'0A0D'.

> However, when I use this against a live file I have a problem.  The
> live
> file contains stand alone cr (x'0d') as line delimiters.  For some
> reason,
> the existing x'0d' characters get removed.

Line separator of X'0D'... where have I seen that before?  MAC, maybe?  How
to you get this file?  If via FTP and if BIN mode has been selected, is it
possible to change to ASCII mode instead, to solve this?  If the file comes
another way, who controls its format?

So are you saying that sed sees this all as one line now?  That could be an
issue.  I'll be more able to play with this when I arrive home.  First thing
I will try is the GNU binaries for sed et al.  If you want to shortcut this
(I have about 90 minute commute) you can find those for installation into
PASE at 

http://www-1.ibm.com/servers/enable/site/porting/iseries/overview/gnu_utilit
ies.html

Dennis Lovelady
http://www.linkedin.com/in/dennislovelady
--
The best way to save face is to keep the lower part full