Re: Any way in the CL program to find the length of the character datastored in a variable? -- MIDRANGE-L

On Sun, 06 Mar 2005 13:52:21 -0600, Vernon Hamberg <vhamberg@xxxxxxxxxxx> wrote:
> You're right, I see. And QCLSCAN does not have the ability to scan from the
> right.
> 
> But I found something that works directly from CL in V5R3 and above - the
> MI function TRIML (Trim Length). Here is a sample that shows how it works.
> Member type must be CLLE--
> 
>   PGM        PARM(&TESTSTRING)
> 
>   DCL        VAR(&TESTSTRING) +
>                TYPE(*CHAR) +
>                LEN(50)
>   DCL        VAR(&STRING) +
>                TYPE(*CHAR) +
>                LEN(51)
>   DCL        VAR(&TRIMCHAR) +
>                TYPE(*CHAR) +
>                LEN(1) +
>                VALUE(' ')
>   DCL        VAR(&TRIMLEN) +
>                TYPE(*INT) +
>                LEN(4)
>   DCL        VAR(&NULL) +
>                TYPE(*CHAR) +
>                LEN(1) +
>                VALUE(x'00')
>   DCL        VAR(&TRIMLENC) +
>                TYPE(*CHAR) +
>                LEN(10)
> 
>   CHGVAR     VAR(&STRING) +
>                VALUE((&TESTSTRING *cat &NULL))
>   CALLPRC    PRC('triml') +
>                PARM(&STRING (&TRIMCHAR *BYVAL)) +
>                RTNVAL(&TRIMLEN)

I did not think of this and I didnt know v5r3 could pass arguments by
value.  You can also use *tcat to put the null immed after the last
non blank and then call the "strlen" c runtime.
  dcl &ch4 *char 4
  dcl &lx *dec 5
  chgvar &string (&string *tcat &null)
 CALLPRC    PRC('strlen') PARM(&String) RTNVAL(%BIN(&CH4))
 chgvar &lx %bin(&ch4)

I like the approach of writing your own proc in RPG  that way, you
dont have to do any pre or post work in the CL program:

 CALLPRC    PRC('CharLx') PARM(&ADDR) RTNVAL(&LX) 

the RPG proc uses the OPDESC info that is automatically passed when CL
calls a procedure to calc the size of the CL character variable that
holds the string:
** ------------------ CharLx -----------------------------
** calc the length of cl character variable
pCharLx           b                   export
dCharLx           pi             7p 0 opdesc
d InCharVar                   9999a   const options(*VarSize)
d #               s             10i 0
d Sx              s             10i 0
 /free
        ceedod( 1: #: #: #: #: Sx: *Omit ) ;
        return    %len(%trimr(%subst(InCharVar:1:Sx))) ;
 /end-free
p                 e

-Steve

>   CHGVAR     VAR(&TRIMLENC) +
>                VALUE(&TRIMLEN)
>   SNDPGMMSG  MSG('Trimmed length of "' *CAT &TESTSTRING *CAT '" is' *BCAT
> &TRIMLENC)
> 
>   ENDPGM
> 
> The 'triml' function needs a null-terminated string as input (append it to
> incoming parameter here), the second parameter has to be passed by value,
> so earlier versions cannot use this - create a command that uses RPGLE and
> returns the length.
> 
> You will need to bind to a system service program, so create the module
> first - assume the member name is YOURLIB/TESTTRIML. Then
> 
> CRTPGM PGM(YOURLIB/TESTTRIML) BNDSRVPGM(QC2UTIL1)
> 
> to create the program.
> 
> (NOTE: Use any suitable debug or activation group parameters you want.)
> 
> It MIGHT be possible to restore this to a V5R2 machine, if correctly saved.
> i did not have success, however, in my short test. It DID compile and save
> to V5R2, however.
> 
> In my test
> 
> call vern/testtriml 'This is a string'
> 
> resulted in the message
> 
> Trimmed length of "This is a string                                  " is
> 0000000016
> 
> Of course, you could always build a command around the BIFs of RPGLE.   ;-)
> 
> HTH
> Vern
> 
> At 09:05 AM 3/6/2005, you wrote:
> >Isn't the problem with QCLSCAN that you cannot test for a multiple word
> >variable?  When you see the first blank the scan stops and the result is
> >invalid.
> >
> >-----Original Message-----
> >From: midrange-l-bounces@xxxxxxxxxxxx
> >[mailto:midrange-l-bounces@xxxxxxxxxxxx]On Behalf Of Vernon Hamberg
> >Sent: Sunday, March 06, 2005 1:19 AM
> >To: Midrange Systems Technical Discussion
> >Subject: Re: Any way in the CL program to find the length of the
> >character datastored in a variable?
> >
> >
> >This is interesting - which really is the bigger hit on performance. I
> >think it depends. Using %SST in a loop  could very likely take more time
> >than a single call to a program that uses a more efficient loop - probably
> >in some way that uses pointers and doesn't need to resolve the %SST for
> >each test character. I can see that a 256-character variable that has only
> >the 1st 3 characters filled would take longer with %SST than with a call to
> >QCLSCAN. But it probably does not matter too much, unless this is repeated
> >over several hundreds of thousands of records or values, say.
> >
> >Still, it would be an interesting test. Maybe I'll try it when I have some
> >time.
> >
> >Vern
> >
> >At 10:43 AM 3/5/2005, you wrote:
> > >Here's the V5R2 link:
> > >
> > >http://publib.boulder.ibm.com/iseries/v5r2/ic2924/index.htm?info/apis/qclsc
> >a
> > >n.htm
> > >
> > >But I would use a substring method myself - no reason to call another
> > >program and take the performance hit when you can do it in a loop with
> >%SST.
> > >
> > >
> > > >            Siva
> > > >            <matchrefree@yaho
> > > >            o.com>                                                     To
> > > >            Sent by:                  Midrange Systems Technical
> > > >            midrange-l-bounce         Discussion
> > > >            s@xxxxxxxxxxxx            <midrange-l@xxxxxxxxxxxx>
> > > >                                                                       cc
> > > >
> > > >            03/04/2005 11:26                                      Subject
> > > >            PM                        RE: Any way in the CL program to
> > > >                                      find the length of the character
> > > >                                      datastored in a variable?
> > > >            Please respond to
> > > >            Midrange Systems
> > > >                Technical
> > > >               Discussion
> > > >            <midrange-l@midra
> > > >                nge.com>
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > > Thanks Doug....
> > > > Is it an API?... Wats the Parm list for this...  Coudnt get any in my
> > > > manuals...
> > > >
> > > > Thanks
> > > > Siva.
> > > >
> > > >
> > > > Doug Hart <DougHart@xxxxxxxxxxxx> wrote:
> > > >
> > > > QCLSCAN
> > > >
> > > >
> > > > ---
> > > > Doug Hart
> > > >
> > > >
> > > >
> > > > -----Original Message-----
> > > > From: midrange-l-bounces@xxxxxxxxxxxx
> > > > [mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of Siva
> > > > Sent: Friday, March 04, 2005 11:02 PM
> > > > To: MIDRANGE-L@xxxxxxxxxxxx
> > > > Subject: Any way in the CL program to find the length of the character
> > > > datastored in a variable?
> > > >
> > > >
> > > > Is there any way in the CL program to find the length of the character
> >data
> > > > stored in a variable?
> > > >
> > > > I find there is no direct CL command to acheive this... may be some
> > > > roundabout way there to do this... but nothing striking at this
> >moment....
> > > >
> > > > i think of some prgm calls with the required variable as the parm and
> >the
> > > > receiviong prgm process the variable etc... but couldnt get any logic
> >here
> > > > :-)......
> > > >
> > > >
> > > > Thanks
> > > > Siva.
> > >
> > >--
> > >This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
> > >To post a message email: MIDRANGE-L@xxxxxxxxxxxx
> > >To subscribe, unsubscribe, or change list options,
> > >visit: http://lists.midrange.com/mailman/listinfo/midrange-l
> > >or email: MIDRANGE-L-request@xxxxxxxxxxxx
> > >Before posting, please take a moment to review the archives
> > >at http://archive.midrange.com/midrange-l.
> >
> >--
> >This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
> >To post a message email: MIDRANGE-L@xxxxxxxxxxxx
> >To subscribe, unsubscribe, or change list options,
> >visit: http://lists.midrange.com/mailman/listinfo/midrange-l
> >or email: MIDRANGE-L-request@xxxxxxxxxxxx
> >Before posting, please take a moment to review the archives
> >at http://archive.midrange.com/midrange-l.
> >
> >--
> >This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
> >To post a message email: MIDRANGE-L@xxxxxxxxxxxx
> >To subscribe, unsubscribe, or change list options,
> >visit: http://lists.midrange.com/mailman/listinfo/midrange-l
> >or email: MIDRANGE-L-request@xxxxxxxxxxxx
> >Before posting, please take a moment to review the archives
> >at http://archive.midrange.com/midrange-l.
> 
> --
> This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
> To post a message email: MIDRANGE-L@xxxxxxxxxxxx
> To subscribe, unsubscribe, or change list options,
> visit: http://lists.midrange.com/mailman/listinfo/midrange-l
> or email: MIDRANGE-L-request@xxxxxxxxxxxx
> Before posting, please take a moment to review the archives
> at http://archive.midrange.com/midrange-l.
> 
>