PDF merging

View Source

NativeElixirPdfUtilities.Merge combines a non-empty list of PDF binaries and returns a new PDF.

alias NativeElixirPdfUtilities.Merge

with {:ok, first} <- File.read("first.pdf"),
     {:ok, second} <- File.read("second.pdf"),
     {:ok, merged} <- Merge.merge([first, second]),
     :ok <- File.write("merged.pdf", merged) do
  :ok
end

Input order determines page order. merge/1 accepts PDF binaries rather than paths so callers control file access, storage, and error handling.

Supported inputs

Merging uses the shared PDF reader and therefore supports:

  • classic cross-reference tables
  • cross-reference streams and object streams
  • incremental revisions and hybrid-reference files
  • active object generations and free entries
  • inherited page Resources, MediaBox, CropBox, and Rotate values, including indirectly referenced page rectangles
  • the stream filters and resource limits described in the PDF reader guide

Malformed, encrypted, unsupported, or resource-intensive inputs fail through the shared diagnostic contract. The merger does not decrypt PDFs.

Output behavior

The merger emits a PDF 1.7 document with a new cross-reference table, trailer, catalog, and flat page tree. Active input objects receive new object numbers, their indirect references are rewritten, and stream bytes and filter declarations are preserved. Effective inherited page resources, media and crop boxes, and rotation values are written onto the merged pages where required. Existing page-level entries such as BleedBox, TrimBox, ArtBox, and UserUnit remain on copied page dictionaries. If required page geometry is missing or malformed, the merger returns a diagnostic that identifies the problem. It does not invent a page size.

The merger combines page content and appends each input's top-level outline items in input order. Outline destinations are remapped to the corresponding output pages. It rebuilds the top-level catalog, so it does not carry metadata, viewer preferences, portfolios, or AcroForm configuration into the output.

Internal links are remapped to the corresponding pages from their own input, including names resolved through legacy destination dictionaries or name trees. Resolvable named links become explicit destinations in the merged PDF; unresolved named links are omitted. External URI links remain.

Outline item and aggregate title-byte limits apply to the combined output, even when each input outline individually fits within its limits. Assembly also reserves object capacity for the generated outline tree. See PDF outlines and bookmarks for preservation and limit rules.

Errors

An empty input list and invalid PDF input return diagnostics:

case Merge.merge(pdf_binaries) do
  {:ok, merged_pdf} ->
    merged_pdf

  {:error, {_reason, diagnostic}} ->
    Logger.warning(diagnostic.message)
end

See Diagnostics for the common error shape and the fields available for logging and support.