How to do mail merge on top of a PDF?

Question

I often get a PDF from our designer (built in Adobe InDesign) which is supposed to be sent out to thousands of people.

I've got the list with all the people, and it's easy doing a mail merge in OpenOffice.org. However, OpenOffice.org doesn't support the advanced PDF. I just want to output some text onto each page and print it out.

Here's how I do it now: print out 6.000 copies of the PDF, then put all of them into the printer again and just print out name, address and other information on top of it. But that's expensive.

Sadly, I can't make the PDF to an image and use that in OpenOffice.org because it grinds the computer to a halt. It also takes extremely long time to send this job to the printer.

So, is there an easy way to do this mail merge (preferably in Python) without paying for third party closed solutions?

Answer 1

Probably the best way would be to generate another PDF with the missing text, and overlay one PDF over the other. A quick Google found this link showing how to do it in Acrobat, and I'm sure there are other methods as well.

http://forums.macrumors.com/showthread.php?t=508226

Answer 2

Now I've made an account. I fixed it by using the ingenious pdftk.

In my quest I totally overlook the feature "background" and "overlay". My solution was this:

pdftk names.pdf background boat_background.pdf output out.pdf

Creating the names.pdf you can easily do with Python reportlab or similar PDF-creation scripts. It's best using code to do that, creating 6k pages took several hours in LibreOffice/OpenOffice, while it took just a few seconds using Python.

Answer 3

Someone asked for specifics. I didn't want to sully my top answer with it, because you can do it how you like (and just knowing pdftk is up to it should give people the idea).

But here's some scripts I used ages ago:

csv_to_pdf.py

#!/usr/bin/python
# This makes one PDF page per name in the CSV file
# csv_to_pdf.py <CSV_FILE>

import csv
import sys
from reportlab.pdfgen.canvas import Canvas
from reportlab.lib.units import cm, mm

in_db = csv.reader(open(sys.argv[1], "rb"));
outname = sys.argv[1].replace("csv", "pdf")
pdf = Canvas(outname)
in_db.next()

i = 0
for rad in in_db:
        pdf.setFontSize(11)
        adr = rad[1]

        tekst = pdf.beginText(2*cm, 26*cm)

        for a in adr.split('\n'):
            if not a.strip():
                continue
            if a[-1] == ',':
                a = a[:-1]
            tekst.textLine(a)
        pdf.drawText(tekst)
        pdf.showPage()

        i += 1
        if i % 1000 == 0:
                print i
pdf.save()

When you've ran this, you have a file with thousands of pages, only with a name on it. This is when you can background the fancy PDF under all of them:

pdftk <YOUR_NEW_PDF_FILE.pdf> background <DESIGNED_FILE.pdf> <MERGED.pdf>

Answer 4

For a no-mess, no-fuss solution, use iText to simply add the text to the pdf. For example, you can do the following to add text to a pdf document once loaded:

PdfContentByte cb= ...;
cb.BeginText();
cb.SetFontAndSize(font, fontSize);
float x = ...;
float y = ...;
cb.SetTextMatrix(x, y);
cb.ShowText(fieldValue);
cb.EndText();    

From there on, save it as a different file, and print it.

However, I've found that form fields are the way to go with pdf document generation from templates.

If you have a template with form fields (added with Adobe Acrobat), you have one of two choices :

• Create a FDF file, which is essentially a list of values for the fields on the form. A FDF is a simple text document which references the original document so that when you open up the PDF, the document loads with the field values supplied by the FDF.
• Alternatively, load the template with with a library like iText / iTextSharp, fill the form fields manually, and save it as a seperate pdf.

A sample FDF file looks like this (stolen from Planet PDF) :

%FDF-1.2
%âãÏÓ
1 0 obj
<<<
 /F(Example PDF Form.pdf)
 /Fields[
  <<
  /T(myTextField)
  /V(myTextField default value)
  >>
  ]
 >>
>> endobj trailer
<>
%%EOF

Because of the simple format and the small size of the FDF, this is the preferred approach, and the approach should work well in any language.

As for filling the fields programmatically, you can use iText in the following way :

PdfAcroForm acroForm = writer.AcroForm;
acroForm.Put(new PdfName(fieldInfo.Name), new PdfString(fieldInfo.Value));

Answer 5

You could probably look at a PDF library like iText. If you have some programming knowledge and a bit of time you could write some code that adds the contact information to the PDFs

Answer 6

One easy way would be to create a fillable pdf form from the original document in Acrobat and do a mail merge with the form and a csv.

PDF mail merges are relatively easy to do in python and pdftk. Fdfgen (pip install fdfgen) is a python library that will create an fdf from a python array, so you can save the excel grid to a csv, make sure that the csv headers match the name of the pdf form field you want to fill with that column, and do something like

import csv
import subprocess

from fdfgen import forge_fdf

PDF_FORM = 'path/to/form.pdf'
CSV_DATA = 'path/to/data.csv'

infile = open(CSV_DATA, 'rb')
reader = csv.DictReader(infile)
rows = [row for row in reader]
infile.close()

for row in rows:
    # Create fdf
    filename = row['filename'] # Construct filename
    fdf_data = [(k,v) for k, v in row.items()]
    fdf = forge_fdf(fdf_data_strings=fdf_data)
    fdf_file = open(filename+'.fdf', 'wb')
    fdf_file.write(fdf)
    fdf_file.close()

    # Use PDFTK to create filled, flattened, pdf file
    cmds = ['pdftk', PDF_FORM, 'fill_form', filename+'.fdf',
            'output', filename+'.pdf', 'flatten', 'dont_ask']
    process = subprocess.Popen(cmds, stdout=subprocess.PIPE)
    stdout, stderr = process.communicate()
    returncode = process.poll()
    os.remove(filename+'.fdf')

I've encountered this problem enough to write my own free solution, PdfZero. PdfZero has a mail merge feature to merge spreadsheets with PDF forms. You will still need to create a PDF form, but you can upload the form and csv to pdfzero, select which form fields you want filled with which columns, create a naming convention for each filled pdf using the csv data if needed, and batch generate the filled PDfs.

DISCLAIMER: I wrote PdfZero