So there is more to this than simply combining the data from each file. The actual usable data for each file is "packaged" with markups and other code (similar to HTML and other document formats). You actually have to decode each PDF file, combine the necessary parts, then re-encode with new "packaging." This requires a working knowledge of the PDF specifications and structure, available freely from Adobe here.
I used this information to write a script sufficient for my needs. It does not account for every possibility, however, so particularly merging any documents requiring PDF-1.4 and higher, this will need quite a bit of work.
/**
* Merges all given PDF files into one.
*
* @param {Folder} directory the folder to store the output file
* @param {string} name the desired name of the output file
* @param {File} pdf1 the first PDF file
* @param {File} pdf2 the second PDF file
* @param {File} opt_pdf3 [optional] the third PDF file; add as many more as you like
*
* @return {File} the merged file
*/
function mergePdfs(directory, name, pdf1, pdf2, opt_pdf3) {
if (name.slice(-4) != '.pdf') {
name = name + '.pdf';
}
var newObjects = ['1 0 obj
<</Type/Catalog/Pages 2 0 R >>
endobj'];
var pageAddresses = [];
for (var argumentIndex = 2; argumentIndex < arguments.length; argumentIndex++) {
var bytes = arguments[argumentIndex].getBlob().getBytes();
var xrefByteOffset = '';
var byteIndex = bytes.length - 1;
while (!/sstartxrefs/.test(xrefByteOffset)) {
xrefByteOffset = String.fromCharCode(bytes[byteIndex]) + xrefByteOffset;
byteIndex--;
}
xrefByteOffset = +(/sd+s/.exec(xrefByteOffset)[0]);
var objectByteOffsets = [];
var trailerDictionary = '';
var rootAddress = '';
do {
var xrefTable = '';
var trailerEndByteOffset = byteIndex;
byteIndex = xrefByteOffset;
for (byteIndex; byteIndex <= trailerEndByteOffset; byteIndex++) {
xrefTable = xrefTable + String.fromCharCode(bytes[byteIndex]);
}
xrefTable = xrefTable.split(/s*trailers*/);
trailerDictionary = xrefTable[1];
if (objectByteOffsets.length < 1) {
rootAddress = /d+s+d+s+R/.exec(//Roots*d+s+d+s+R/.exec(trailerDictionary)[0])[0].replace('R', 'obj');
}
xrefTable = xrefTable[0].split('
');
xrefTable.shift();
while (xrefTable.length > 0) {
var xrefSectionHeader = xrefTable.shift().split(/s+/);
var objectNumber = +xrefSectionHeader[0];
var numberObjects = +xrefSectionHeader[1];
for (var entryIndex = 0; entryIndex < numberObjects; entryIndex++) {
var entry = xrefTable.shift().split(/s+/);
objectByteOffsets.push([[objectNumber, +entry[1], 'obj'], +entry[0]]);
objectNumber++;
}
}
if (/s*/Prev/.test(trailerDictionary)) {
xrefByteOffset = +(/s*d+s/.exec(/s*/Prevs*d+s/.exec(trailerDictionary)[0])[0]);
}
} while (/s*/Prev/.test(trailerDictionary));
var rootObject = getObject(rootAddress, objectByteOffsets, bytes);
var pagesAddress = /d+s+d+s+R/.exec(//Pagess*d+s+d+s+R/.exec(rootObject)[0])[0].replace('R', 'obj');
var pagesObject = getObject(pagesAddress, objectByteOffsets, bytes);
var objects = getDependencies(pagesObject, objectByteOffsets, bytes);
var newObjectsInsertionIndex = newObjects.length;
for (var objectIndex = 0; objectIndex < objects.length; objectIndex++) {
var newObjectAddress = [(newObjects.length + 3) + '', 0 + '', 'obj'];
if (!Array.isArray(objects[objectIndex])) {
objects[objectIndex] = [objects[objectIndex]];
}
objects[objectIndex].unshift(newObjectAddress);
var objectAddress = objects[objectIndex][1].match(/d+s+d+s+obj/)[0].split(/s+/);
objects[objectIndex].splice(1, 0, objectAddress);
if (//Types*/Page[^s]/.test(objects[objectIndex][2])) {
objects[objectIndex][2] = objects[objectIndex][2].replace(//Parents*d+s+d+s+R/.exec(objects[objectIndex][2])[0], '/Parent 2 0 R');
pageAddresses.push(newObjectAddress.join(' ').replace('obj', 'R'));
}
var addressRegExp = new RegExp(objectAddress[0] + '\s+' + objectAddress[1] + '\s+' + 'obj');
objects[objectIndex][2] = objects[objectIndex][2].replace(addressRegExp.exec(objects[objectIndex][2])[0], newObjectAddress.join(' '));
newObjects.push(objects[objectIndex]);
}
for (var referencingObjectIndex = newObjectsInsertionIndex; referencingObjectIndex < newObjects.length; referencingObjectIndex++) {
var references = newObjects[referencingObjectIndex][2].match(/d+s+d+s+R/g);
if (references != null) {
var string = newObjects[referencingObjectIndex][2];
var referenceIndices = [];
var currentIndex = 0;
for (var referenceIndex = 0; referenceIndex < references.length; referenceIndex++) {
referenceIndices.push([]);
referenceIndices[referenceIndex].push(string.slice(currentIndex).indexOf(references[referenceIndex]) + currentIndex);
referenceIndices[referenceIndex].push(references[referenceIndex].length);
currentIndex += string.slice(currentIndex).indexOf(references[referenceIndex]);
}
for (var referenceIndex = 0; referenceIndex < references.length; referenceIndex++) {
var objectAddress = references[referenceIndex].replace('R', 'obj').split(/s+/);
for (var objectIndex = newObjectsInsertionIndex; objectIndex < newObjects.length; objectIndex++) {
if (arrayEquals(objectAddress, newObjects[objectIndex][1])) {
var length = string.length;
newObjects[referencingObjectIndex][2] = string.slice(0, referenceIndices[referenceIndex][0]) + newObjects[objectIndex][0].join(' ').replace('obj', 'R') +
string.slice(referenceIndices[referenceIndex][0] + referenceIndices[referenceIndex][1]);
string = newObjects[referencingObjectIndex][2];
var newLength = string.length;
if (!(length == newLength)) {
for (var subsequentReferenceIndex = referenceIndex + 1; subsequentReferenceIndex < references.length; subsequentReferenceIndex++) {
referenceIndices[subsequentReferenceIndex][0] += (newLength - length);
}
}
break;
}
}
}
}
}
for (var objectIndex = newObjectsInsertionIndex; objectIndex < newObjects.length; objectIndex++) {
if (Array.isArray(newObjects[objectIndex])) {
if (newObjects[objectIndex][3] != undefined) {
newObjects[objectIndex] = newObjects[objectIndex].slice(2);
} else {
newObjects[objectIndex] = newObjects[objectIndex][2];
}
}
}
}
newObjects.splice(1, 0, '2 0 obj
<</Type/Pages/Count ' + pageAddresses.length + ' /Kids [' + pageAddresses.join(' ') + ' ]>>
endobj');
newObjects.splice(2, 0, '3 0 obj
<</Title (' + name + ') /CreationDate (D' +
Utilities.formatDate(new Date(), CalendarApp.getDefaultCalendar().getTimeZone(), 'yyyyMMddHHmmssZ').slice(0, -2) + "'00) /ModDate (D" + Utilities.formatDate(new Date(),
CalendarApp.getDefaultCalendar().getTimeZone(), 'yyyyMMddHHmmssZ').slice(0, -2) + "'00)>>
endobj");
var byteOffsets = [0];
var bytes = [];
var header = '%PDF-1.3
';
for (var headerIndex = 0; headerIndex < header.length; headerIndex++) {
bytes.push(header.charCodeAt(headerIndex));
}
bytes.push('%'.charCodeAt(0));
for (var characterCode = -127; characterCode < -123; characterCode++) {
bytes.push(characterCode);
}
bytes.push('
'.charCodeAt(0));
bytes.push('
'.charCodeAt(0));
while (newObjects.length > 0) {
byteOffsets.push(bytes.length);
var object = newObjects.shift();
if (Array.isArray(object)) {
var streamKeyword = /streams*
/.exec(object[0])[0];
if (streamKeyword.indexOf('
') > streamKeyword.length - 3) {
streamKeyword = streamKeyword.slice(0, -1);
} else if (streamKeyword.indexOf('
') > streamKeyword.length - 5) {
streamKeyword = streamKeyword.slice(0, -2);
}
var streamIndex = object[0].indexOf(streamKeyword) + streamKeyword.length;
for (var objectIndex = 0; objectIndex < streamIndex; objectIndex++) {
bytes.push(object[0].charCodeAt(objectIndex))
}
bytes = bytes.concat(object[1]);
for (var objectIndex = streamIndex; objectIndex < object[0].length; objectIndex++) {
bytes.push(object[0].charCodeAt(objectIndex));
}
} else {
for (var objectIndex = 0; objectIndex < object.length; objectIndex++) {
bytes.push(object.charCodeAt(objectIndex));
}
}
bytes.push('
'.charCodeAt(0));
bytes.push('
'.charCodeAt(0));
}
var xrefByteOffset = bytes.length;
var xrefHeader = 'xref
';
for (var xrefHeaderIndex = 0; xrefHeaderIndex < xrefHeader.length; xrefHeaderIndex++) {
bytes.push(xrefHeader.charCodeAt(xrefHeaderIndex));
}
var xrefSectionHeader = '0 ' + byteOffsets.length + '
';
for (var xrefSectionHeaderIndex = 0; xrefSectionHeaderIndex < xrefSectionHeader.length; xrefSectionHeaderIndex++) {
bytes.push(xrefSectionHeader.charCodeAt(xrefSectionHeaderIndex));
}
for (var byteOffsetIndex = 0; byteOffsetIndex < byteOffsets.length; byteOffsetIndex++) {
for (var byteOffsetStringIndex = 0; byteOffsetStringIndex < 10; byteOffsetStringIndex++) {
bytes.push(Utilities.formatString('%010d', byteOffsets[byteOffsetIndex]).charCodeAt(byteOffsetStringIndex));
}
bytes.push(' '.charCodeAt(0));
if (byteOffsetIndex == 0) {
for (var generationStringIndex = 0; generationStringIndex < 5; generationStringIndex++) {
bytes.push('65535'.charCodeAt(generationStringIndex));
}
for (var keywordIndex = 0; keywordIndex < 2; keywordIndex++) {
bytes.push(' f'.charCodeAt(keywordIndex));
}
} else {
for (var generationStringIndex = 0; generationStringIndex < 5; generationStringIndex++) {
bytes.push('0'.charCodeAt(0));
}
for (var keywordIndex = 0; keywordIndex < 2; keywordIndex++) {
bytes.push(' n'.charCodeAt(keywordIndex));
}
}
bytes.push('
'.charCodeAt(0));
bytes.push('
'.charCodeAt(0));
}
for (var trailerHeaderIndex = 0; trailerHeaderIndex < 9; trailerHeaderIndex++) {
bytes.push('trailer
'.charCodeAt(trailerHeaderIndex));
}
var idBytes = Utilities.computeDigest(Utilities.DigestAlgorithm.MD5, (new Date).toString());
var id = '';
for (var idByteIndex = 0; idByteIndex < idBytes.length; idByteIndex++) {
id = id + ('0' + (idBytes[idByteIndex] & 0xFF).toString(16)).slice(-2);
}
var trailer = '<</Size ' + (byteOffsets.length) + ' /Root 1 0 R /Info 2 0 R /ID [<' + id + '> <' + id + '>]>>
startxref
' + xrefByteOffset + '
%%EOF';
for (var trailerIndex = 0; trailerIndex < trailer.length; trailerIndex++) {
bytes.push(trailer.charCodeAt(trailerIndex));
}
return directory.createFile(Utilities.newBlob(bytes, 'application/pdf', name));
function getObject(objectAddress, objectByteOffsets, bytes) {
objectAddress = objectAddress.split(/s+/);
for (var addressIndex = 0; addressIndex < 2; addressIndex++) {
objectAddress[addressIndex] = +o
