Date: Fri, 4 Nov 2005 18:38:04 +0000
Reply-To: iw1junk@COMCAST.NET
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Ian Whitlock <iw1junk@COMCAST.NET>
Subject: Re: Forcing SAS to print the resolved value of macro variables in
the log
Scott,
The simple answer.
1 options mprint;
2 %let foo=bar;
3 options symbolgen ;
SYMBOLGEN: Macro variable FOO resolves to bar
4 title "&foo";
SYMBOLGEN: Macro variable FOO resolves to bar
5 data &foo ;
6 &foo = "&foo" ;
SYMBOLGEN: Macro variable FOO resolves to bar
SYMBOLGEN: Macro variable FOO resolves to bar
7 run ;
NOTE: The data set WORK.BAR has 1 observations and 1 variables.
The following is based on "Make Your Own Macro Processor" in NESUG 1999
hints at a more sophisticated solution. (Search words "whitlock"
"processor" at <http://www.lexjansen.com/nesug/index.htm>)
filename code temp ;
data _null_ ;
array lkup (2) $ 32 _temporary_ ( "foo", "out" ) ;
length line $ 256 ;
infile cards truncover ;
input line $char100. ;
if index ( line , '&' ) then
do ;
do i = 1 to dim ( lkup ) ;
line = tranwrd ( line
, '&'||trim(lkup[i])||'.'
, trim(symget(lkup[i])) ) ;
end ;
end ;
file code ;
put line $char50. ;
cards4 ;
options mprint;
%let foo=bar;
%let out = junk ;
title "&foo.";
data &out. ;
&foo. = "&out." ;
run ;
/* more code */
;;;;
%inc code / source2 ;
The PRX functions together with RESOLVE might speed up and allow more
sophisticated processing, eg &&&var &&v&i etc. I think the paper has the
code in a macro and the array of macro variable names is constructed
automatically from the %LET statements. Note that in the above code the
"." ender is essential, but with more testing it could be eliminated.
Ian Whitlock
==================
Date: Sat, 5 Nov 2005 01:37:48 +1100
Reply-To: Scott Bass <usenet739_yahoo_com_au@ALFREDO.CC.UGA.EDU>
Sender: "SAS(r) Discussion"
From: Scott Bass <usenet739_yahoo_com_au@ALFREDO.CC.UGA.EDU>
Subject: Re: Forcing SAS to print the resolved value of macro
variables in
the log
Comments: To: sas-l
Thanks Ron, that was a typo. That will teach me to try to improve the
posted testcase without retesting. I've corrected the code below.
"Scott Bass" <usenet739_yahoo_com_au> wrote in message
news:436a97b1$0$1745$5a62ac22@per-qv1-newsreader-01.iinet.net.au...
> Hi,
>
> Summary:
>
> Submit this code:
>
> 1 options mprint;
> 2 %let foo=bar;
> 3 title "&foo"; <<<<<
> 4 %macro foo;
> 5 title "&foo";
> 6 %mend;
> 7 %foo;
> MPRINT(FOO): title "bar"; <<<<<
>
> I wish there was some way I could force line 3 in the log to be the
> resolved value of the macro variable, like line 7, rather than the macro
> variable reference.
>
> Details/Rationale:
>
> We have an SDLC requirement to document the test cases within our
> programs. The problem is, our test case documentation was always out of
> sync with the test cases in the programs themselves.
>
> Therefore, I wrote a Perl script (thanks comp.lang.perl but, man, what a
> TOUGH crowd on the newbies) plus a bit of SAS (thanks Richard) that
parses
> .sas or more often .log files for specially formatted comment blocks, and
> generates the test case documentation. This is analogous to JavaDoc or
> Perl POD, where the documentation for the program is embedded in the
> program itself.
>
> I've encountered a problem, for which I need to come up with a
> workaround/best practice. If the documentation block contains macro
> variable references, outside of a macro, the SAS log will print the macro
> variable reference, rather than the resolved value. Therefore, the
> reference, rather than the value, ends up in the doc.
>
> My Perl script does handle/parse out the "MPRINT (FOO):" text, so the
> immediate workaround is to force the end user to use a macro if the test
> case block contains macro variables. However, this does add
> complexity/additional code, just to get the log to read "properly". I
was
> hoping there was some magical, undocumented, whizbang SAS option that
> would give me what I want.
>
> Or some brilliant idea/alternate approach/workaround from a fellow
> SAS-L'er.
>
> Cheers,
> Scott
>
| 
