Extracting number from line

Hi all,

I'm racking my brain the whole day long for a rather simple question.
Please help me with your infinite knowledge.

I want to count the number of pages in a postscript file by using the
psselect-command according to

psselect -p- file.ps >/dev/null

Example: me@home: psselect -p- input.ps >/dev/null
me@home: [1] [2] [3] [4] [5] [6] Wrote 6 pages, 603850 bytes

And now, I want to extract the page number (here: 6) to process it for other
things. How do I extract this number from this line? Because I want to use t
he
commands within a shell script and apply it generally the output line shows
the page number at different positions within the line (thus 'cut' doesnt se
em
to work here properly).
Any suggestions ?

thanks,
rncdnet

In article <ba35ffd8.0504110629.67f3027b@posting.google.com>,
R.K. <rncdnet@netscape.net> wrote:
>Hi all,
>
>I'm racking my brain the whole day long for a rather simple question.
>Please help me with your infinite knowledge.
>
>I want to count the number of pages in a postscript file by using the
>psselect-command according to
>
>psselect -p- file.ps >/dev/null
>
>Example: me@home: psselect -p- input.ps >/dev/null
>	 me@home: [1] [2] [3] [4] [5] [6] Wrote 6 pages, 603850 bytes
>
>And now, I want to extract the page number (here: 6) to process it for othe
r
>things. How do I extract this number from this line? Because I want to use 
the
>commands within a shell script and apply it generally the output line shows
>the page number at different positions within the line (thus 'cut' doesnt s
eem
>to work here properly).
>Any suggestions ?

Don't use cut.  cut is stupid.  And this is an AWK newsgroup.

I suppose the following shell would work:

x=$(psselect -p- input.ps 2>&1 >/dev/null | gawk 'sub(/^.*Wrote */,"") {prin
t $1}')

Kenny McCormack wrote:
> In article <ba35ffd8.0504110629.67f3027b@posting.google.com>,
> R.K. <rncdnet@netscape.net> wrote:
> 
>
>
> Don't use cut.  cut is stupid.  And this is an AWK newsgroup.
>
> I suppose the following shell would work:
>
> x=$(psselect -p- input.ps 2>&1 >/dev/null | gawk 'sub(/^.*Wrote */,"") {pr
int $1}')
>
No need for sub:

x=$(psselect -p- input.ps 2>&1 >/dev/null |gawk '/Wrote/{print $(NF-3)}'

Ed.

Le Mon, 11 Apr 2005 07:29:06 -0700, R.K. a écrit_:

> Hi all,
>
> I'm racking my brain the whole day long for a rather simple question.
> Please help me with your infinite knowledge.
>
> I want to count the number of pages in a postscript file by using the
> psselect-command according to
>
> psselect -p- file.ps >/dev/null
>
> Example: me@home: psselect -p- input.ps >/dev/null
> 	 me@home: [1] [2] [3] [4] [5] [6] Wrote 6 pages, 603850 bytes
>
> And now, I want to extract the page number (here: 6) to process it for oth
er
> things. How do I extract this number from this line? Because I want to use
 the
> commands within a shell script and apply it generally the output line show
s
> the page number at different positions within the line (thus 'cut' doesnt 
seem
> to work here properly).
> Any suggestions ?

An overkill in fonky style but it ensures to get what's between
"Wrote" and " pages,"
$ awk 'BEGIN{RS=" pages,";FS="Wrote "};/Wrote /,/ pages,/{print $2}'

In article <KpWdnTbqr8nNDcffRVn-tg@comcast.com>,
Ed Morton  <morton@lsupcaemnt.com> wrote:
... 
>No need for sub:
>
>x=$(psselect -p- input.ps 2>&1 >/dev/null |gawk '/Wrote/{print $(NF-3)}'
>
>	Ed.

6 of one, half dozen of the other.

Mine seems a little more generic.  YMMV.

In article <ba35ffd8.0504110629.67f3027b@posting.google.com>,
R.K. <rncdnet@netscape.net> wrote:

% Please help me with your infinite knowledge.

I assume this means you want a lot of answers.

% Example: me@home: psselect -p- input.ps >/dev/null
% 	 me@home: [1] [2] [3] [4] [5] [6] Wrote 6 pages, 603850 bytes

You could use
(psselect -p- input.ps >/dev/null) 2>&1 |
awk -F'Wrote | page' '/Wrote/ { print $2 }'

On the other hand, since psselect depends on the postscript file
following the document structuring conventions, you don't really
need to use it, just parse out the relevant DSC yourself:

awk '/^%%Pages: [0-9]+/ { print $NF; exit }' input.ps

or count all the '/^%%Page:/' entries.
--

Patrick TJ McPhee
North York  Canada
ptjm@interlog.com