Merging PDF files

Basic Example

from pypdf import PdfWriter

merger = PdfWriter()

for pdf in ["file1.pdf", "file2.pdf", "file3.pdf"]:
    merger.append(pdf)

merger.write("merged-pdf.pdf")
merger.close()

For more details, see an excellent answer on StackOverflow by Paul Rooney.

Showing more merging options

from pypdf import PdfWriter

merger = PdfWriter()

input1 = open("document1.pdf", "rb")
input2 = open("document2.pdf", "rb")
input3 = open("document3.pdf", "rb")

# Add the first 3 pages of input1 document to output
merger.append(fileobj=input1, pages=(0, 3))

# Insert the first page of input2 into the output beginning after the second page
merger.merge(position=2, fileobj=input2, pages=(0, 1))

# Append entire input3 document to the end of the output document
merger.append(input3)

# Write to an output PDF document
output = open("document-output.pdf", "wb")
merger.write(output)

# Close file descriptors
merger.close()
output.close()

append

append has been slightly extended in PdfWriter. See PdfWriter.append for more details.

Examples

# Append the first 10 pages of source.pdf
writer.append("source.pdf", (0, 10))

# Append the first and 10th page from reader and create an outline
writer.append(reader, "page 1 and 10", [0, 9])

During merging, the relevant named destination will also imported.

If you want to insert pages in the middle of the destination, use merge (which provides an insertion position). You can insert the same page multiple times, if necessary even using a list-based syntax:

writer.append(reader, [0, 1, 0, 2, 0])

will insert the pages 1 and 2 with page 0 before, in the middle and after.

add_page / insert_page

It is recommended to use append or merge instead.

Merging forms

When merging forms, some form fields may have the same names, preventing access to some data.

A grouping field should be added before adding the source PDF to prevent that. The original fields will be identified by adding the group name.

For example, after calling reader.add_form_topname("form1"), the field previously named field1 will now identified as form1.field1 when calling reader.get_form_text_fields(True) or reader.get_fields().

After that, you can append the input PDF completely or partially using writer.append or writer.merge. If you insert a set of pages, only those fields will be listed.

reset_translation

During cloning, if an object has been already cloned, it will not be cloned again, and a pointer to this previously cloned object is returned instead. Because of that, if you add/merge a page that has already been added, the same object will be added the second time. If you modify any of these two pages later, both pages can be modified independently.

To reset, call writer.reset_translation(reader).

Advanced cloning

In order to prevent side effects between pages/objects and all objects linked cloning is done during the merge.

This process will be automatically applied if you use PdfWriter.append/merge/add_page/insert_page. If you want to clone an object before attaching it “manually”, use the clone method of any PdfObject:

cloned_object = object.clone(writer)

If you try to clone an object already belonging to the writer, it will return the same object:

assert cloned_object == object.clone(writer)

The same holds true if you try to clone an object twice. It will return the previously cloned object:

assert object.clone(writer) == object.clone(writer)

Please note that if you clone an object, you will clone all the objects below as well, including the objects pointed by IndirectObject. Due to this, if you clone a page that includes some articles ("/B"), not only the first article, but also all the chained articles and the pages where those articles can be read will be copied. This means that you may copy lots of objects which will be saved in the output PDF as well.

In order to prevent this, you can provide the list of fields in the dictionaries to be ignored:

new_page = writer.add_page(reader.pages[0], excluded_fields=["/B"])

Merging rotated pages

If you are working with rotated pages, you might want to call transfer_rotation_to_content() on the page before merging to avoid wrongly rotated results:

for page in writer.pages:
    if page.rotation != 0:
        page.transfer_rotation_to_content()
    page.merge_page(background, over=False)