Chapter 2
Merging and Splitting

cpdf -merge in1.pdf [<range>] in2.pdf [<range>] [<more names/ranges>]
     [-collate] [-retain-numbering] [-remove-duplicate-fonts]
     [-merge-add-bookmarks [-merge-add-bookmarks-use-titles]]
     [-no-process-struct-trees]
     -o out.pdf

cpdf -split in.pdf -o <format> [-chunk <chunksize>]

cpdf -split-bookmarks <level> in.pdf [-utf8] -o <format>

cpdf -split-max <file size> in.pdf -o <format>

cpdf -spray in.pdf -o a.pdf [-o b.pdf [-o ...]]

2.1 Merging

The -merge operation allow the merging of several files into one. Ranges can be used to select only a subset of pages from each input file in the output. The output file consists of the concatenation of all the input pages in the order specified on the command line. Actually, the -merge can be omitted, since this is the default operation of cpdf.

cpdf -merge a.pdf 1 b.pdf 2-end -o out.pdf

Take page one of a.pdf and all but the first page of b.pdf, merge them and produce out.pdf.

cpdf -merge -idir files -o out.pdf

Merge all files from directory files, producing out.pdf.

Merge maintains and merges bookmarks, named destinations, annotations, tagged PDF information, and so on. PDF features which cannot be merged are retained if they are from the document which first exhibits that feature.

The -collate option collates pages: that is to say, it takes the first page from the first document and its range, then the first page from the second document and its range and so on. When all first pages have been taken, it begins on the second from each range, and so on.

The -retain-numbering option keeps the PDF page numbering labels of each document intact, rather than renumbering the output pages from 1.

The -remove-duplicate-fonts option ensures that fonts used in more than one of the inputs only appear once in the output.

The -merge-add-bookmarks option adds a top-level bookmark for each file, using the filename. Any existing bookmarks are retained. The -merge-add-bookmarks-use-titles, when used in conjunction with -merge-add-bookmarks, will use the title from each PDF’s metadata instead of the filename.

The -no-process-struct-tree option prevents processing of structure trees. Instead of merging them, they are simply copied across wholesale. This can be useful with -collate which can presently cause an increased size. Or, to simply save time when a merged structure tree is not required.

2.2 Splitting

The -split operation splits a PDF file into a number of parts which are written to file, their names being generated from a format. The optional -chunk option allows the number of pages written to each output file to be set.

cpdf -split a.pdf -o out%%%.pdf

Split a.pdf to the files out001.pdf, out002.pdf etc.

cpdf a.pdf even AND -split -chunk 10 -o dir/out%%%.pdf

Split the even pages of a.pdf to the files out001.pdf, out002.pdf etc. with at most ten pages in each file. The directory (folder) dir must exist.

If the output format does not provide enough numbers for the files generated, the result is unspecified. The following format operators may be used:

%, %%,  %%% etc.   Sequence num berpadded to thenum berofpercentsign s
               @F  O riginalfilen am e withou textension
               @N  Sequence num berwithoutpadding zeroes
               @S  Startpageofthischunk
               @E  En dpage ofthischu nk
               @B  Bo okmark name atthispage

Numbers padded to a fixed width field by zeroes may be obtained for @S and @E by following them with more @ signs e.g @E@@@ for a fixed width of three.

2.3 Splitting on Bookmarks

The -split-bookmarks <level> operation splits a PDF file into a number of parts, according to the page ranges implied by the document’s bookmarks. These parts are then written to file with names generated from the given format.

Level 0 denotes the top-level bookmarks, level 1 the next level (sub-bookmarks) and so on. So -split-bookmarks 1 creates breaks on level 0 and level 1 boundaries.

cpdf -split-bookmarks 0 a.pdf -o out%%%.pdf

Split a.pdf to the files out001.pdf, out002.pdf on bookmark boundaries.

There may be many bookmarks on a single page (for instance, if paragraphs are bookmarked or there are two subsections on one page). The splits calculated by -split-bookmarks ensure that each page appears in only one of the output files. It is possible to use the @ operators above, including operator @B which expands to the text of the bookmark:

cpdf -split-bookmarks 0 a.pdf -o @B.pdf

Split a.pdf on bookmark boundaries, using the bookmark text as the filename.

The bookmark text used for a name is converted from unicode to 7 bit ASCII, and the following characters are removed, in addition to any character with ASCII code less than 32:

 / ? < > \ : * | " ˆ + =

To prevent this process, and convert bookmark names to UTF8 instead, add -utf8 to the command.

2.4 Splitting to Maximum Size

The -split-max operation splits a file into chunks of no more than the given size, starting at the beginning. The suffixes kB, KiB, MB, MiB, GB, and GiB may be used to give the size. For example:

cpdf -split-max 100kB in.pdf -o out%%%.pdf

Split in.pdf into parts of no more than 100kB, if possible.

Should the operation not be possible for the given size, an error message is printed and no output (not even partial output) is produced.

2.5 Spraying

Spraying is a sort of de-collation. It takes one input file, and writes pages in turn to one or more outputs:

cpdf -spray in.pdf -o a.pdf -o b.pdf

Place odd pages of the input file in one file, and the even in another.

This is the only time more than one -o is allowed.

2.6 Encrypting with Split operations

The encryption parameters described in Chapter 4 may be added to the command line to encrypt each split PDF. Similarly, the -recrypt switch described in Chapter 1 may by given to re-encrypt each file with the existing encryption of the source PDF.

2.7 Splitting and structure trees

Presently, cpdf will include the input structure tree in each output PDF. To remove a structure tree entirely, preprocess the file with -remove-dict-entry /StructTreeRoot. If you do not require structure trees, this will lead to faster splitting, and smaller outputs.