Tech Tip: Extract Pages From a PDF


There are a number of ways to extract a range of pages from a PDF file: there are PDF related toolkits for doing it, or you can use Ghostscript directly.

For example, to extract pages 22-36 from a 100-page PDF file using pdftk:

  $ pdftk A=100p-inputfile.pdf cat A22-36 output outfile_p22-p36.pdf

Or use a combination of xpdf-utils (or poppler-tools) with psutils and the ps2pdf command (which ships as part of Ghostscript):

  $ pdftops 100p-inputfile.pdf - | psselect -p22-36 | \
         ps2pdf14 - outfile_p22-p36.pdf

Or, just use Ghostscript (which, unlike pdftk, is installed nearly everywhere; and you've been using it in the last command anyway):

  $ gs -sDEVICE=pdfwrite -dNOPAUSE -dBATCH -dSAFER \
       -dFirstPage=22 -dLastPage=36 \
       -sOutputFile=outfile_p22-p36.pdf 100p-inputfile.pdf

Regarding speed and efficiency of the processing and more important the quality of the output file, the 2nd method above is for sure the worst of the 3. The conversion of the original PDF to PostScript and back to PDF (also known as "refrying" the PDF) is very unlikely to completely preserve advanced PDF features (such as transparency information, font hinting, overprinting information, color profiles, trapping instructions, etc.).

The 3rd method uses Ghostscript only (which the 2nd one uses anyway, because ps2pdf14 is nothing more than a wrapper script around a more or less complicated Ghostscript commandline. The 3rd method also preserves all the important PDF objects on your pages as they are, without any "roundtrip" conversions....

The only drawback of the 3rd method is that it's a longer, more complicated command line to type. But you can overcome that drawback if you save it as a bash function. Just put these lines in your ~/.bashrc file:

function pdfpextr()
    # this function uses 3 arguments:
    #     $1 is the first page of the range to extract
    #     $2 is the last page of the range to extract
    #     $3 is the input file
    #     output file will be named "inputfile_pXX-pYY.pdf"
    gs -sDEVICE=pdfwrite -dNOPAUSE -dBATCH -dSAFER \
       -dFirstPage=${1} \
       -dLastPage=${2} \
       -sOutputFile=${3%.pdf}_p${1}-p${2}.pdf \

Now you only need to type (after starting a new copy bash or sourcing .bashrc) the following:

  $ pdfpextr 22 36 inputfile.pdf

which will result in the file inputfile_p22-p36.pdf in the same directory as the input file.



Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

You should do:

mol's picture

You should do:
-sOutputFile="${3%.pdf}_p${1}-p${2}.pdf" \
or it won't work with filenames with spaces

Hi, i tried the first one,

Seeley's picture

Hi, i tried the first one, and didn't work, only work if you write it like this:

$ pdftk book.pdf cat 22-26 output book2.pdf

i'm using Ubuntu karmic by

Seeley's picture

i'm using Ubuntu karmic by the way
and sorry my bad english, i'm spanish

pdftk is fantastic

Wojtek's picture


Thank you for the great article, it exactly answered to the question I was having for some time now.

I tried solution n°1 and it works perfectly (on Windows 7). My problem was to split into single pages a large PDF document and a

pdftk a.pdf burst output "hello world %d.pdf

made my day.


Or Just "Print"...

waparmley's picture

The only "printer" I use with my laptop and my netbook is cups-pdf. If I really need a paper copy of something I "print" a pdf and send it over to the one computer that is connected to a hardware printer.

So ... if I want to extract pages from a pdf I just open it in Evince then select Print from the menu, then set the page range that I want. My pages are then "printed" to a new pdf.

Seems to work fine for me -- anyone see any downside to this?

Works great, but..

pete12's picture

Hi, I used this function and it seems to work really well. There is only one problem that doesn't seem to effect actual performance, but when I test ran it on the linux mint user manual I got this error:

GPL Ghostscript 8.64: ERROR: A pdfmark destination page 54 points beyond the last page 3.

At first I thought it was a link leading to page 54, but is it just an index value or something of the sort?

Not that it matters much, thanks for the function!

Probably a bogus error message

Anonymous's picture

This is probably a bogus error message (should rather be a warning). If you check your output, most likely everything is there.

for a gui application -> PdfShuffler

gourgi's picture