Merge PDF files on iOS
I came out with this solution:
// Documents dir
NSArray *paths = NSSearchPathForDirectoriesInDomains(NSDocumentDirectory, NSUserDomainMask, YES);
NSString *documentsDirectory = [paths objectAtIndex:0];
// File paths
NSString *pdfPath1 = [documentsDirectory stringByAppendingPathComponent:@"1.pdf"];
NSString *pdfPath2 = [documentsDirectory stringByAppendingPathComponent:@"2.pdf"];
NSString *pdfPathOutput = [documentsDirectory stringByAppendingPathComponent:@"out.pdf"];
// File URLs
CFURLRef pdfURL1 = (CFURLRef)[[NSURL alloc] initFileURLWithPath:pdfPath1];
CFURLRef pdfURL2 = (CFURLRef)[[NSURL alloc] initFileURLWithPath:pdfPath2];
CFURLRef pdfURLOutput = (CFURLRef)[[NSURL alloc] initFileURLWithPath:pdfPathOutput];
// File references
CGPDFDocumentRef pdfRef1 = CGPDFDocumentCreateWithURL((CFURLRef) pdfURL1);
CGPDFDocumentRef pdfRef2 = CGPDFDocumentCreateWithURL((CFURLRef) pdfURL2);
// Number of pages
NSInteger numberOfPages1 = CGPDFDocumentGetNumberOfPages(pdfRef1);
NSInteger numberOfPages2 = CGPDFDocumentGetNumberOfPages(pdfRef2);
// Create the output context
CGContextRef writeContext = CGPDFContextCreateWithURL(pdfURLOutput, NULL, NULL);
// Loop variables
CGPDFPageRef page;
CGRect mediaBox;
// Read the first PDF and generate the output pages
NSLog(@"GENERATING PAGES FROM PDF 1 (%i)...", numberOfPages1);
for (int i=1; i<=numberOfPages1; i++) {
page = CGPDFDocumentGetPage(pdfRef1, i);
mediaBox = CGPDFPageGetBoxRect(page, kCGPDFMediaBox);
CGContextBeginPage(writeContext, &mediaBox);
CGContextDrawPDFPage(writeContext, page);
CGContextEndPage(writeContext);
}
// Read the second PDF and generate the output pages
NSLog(@"GENERATING PAGES FROM PDF 2 (%i)...", numberOfPages2);
for (int i=1; i<=numberOfPages2; i++) {
page = CGPDFDocumentGetPage(pdfRef2, i);
mediaBox = CGPDFPageGetBoxRect(page, kCGPDFMediaBox);
CGContextBeginPage(writeContext, &mediaBox);
CGContextDrawPDFPage(writeContext, page);
CGContextEndPage(writeContext);
}
NSLog(@"DONE!");
// Finalize the output file
CGPDFContextClose(writeContext);
// Release from memory
CFRelease(pdfURL1);
CFRelease(pdfURL2);
CFRelease(pdfURLOutput);
CGPDFDocumentRelease(pdfRef1);
CGPDFDocumentRelease(pdfRef2);
CGContextRelease(writeContext);
The biggest issue here is memory allocation. As you can see, in this approach you have to read both PDF files you want to merge with and, at the same time, generate the output. The releases only occur at the end. I tried combining a PDF file with 500 pages (~15MB) with another containing 100 pages (~3MB) and it produced a new one with 600 pages (of course!) having only ~5MB size (magic?). The execution took around 30 seconds (not so bad, considering an iPad 1) and allocated 17MB (ouch!). The app luckily didn't crash, but I think iOS would love to kill an app consuming 17MB like this one. ;P
I've made a little refactor on Jonathan's code to join any PDF file of any size:
+ (NSString *)joinPDF:(NSArray *)listOfPaths {
// File paths
NSString *fileName = @"ALL.pdf";
NSString *pdfPathOutput = [[NSSearchPathForDirectoriesInDomains(NSDocumentDirectory, NSUserDomainMask, YES) objectAtIndex:0] stringByAppendingPathComponent:fileName];
CFURLRef pdfURLOutput = ( CFURLRef)CFBridgingRetain([NSURL fileURLWithPath:pdfPathOutput]);
NSInteger numberOfPages = 0;
// Create the output context
CGContextRef writeContext = CGPDFContextCreateWithURL(pdfURLOutput, NULL, NULL);
for (NSString *source in listOfPaths) {
CFURLRef pdfURL = ( CFURLRef)CFBridgingRetain([[NSURL alloc] initFileURLWithPath:source]);
//file ref
CGPDFDocumentRef pdfRef = CGPDFDocumentCreateWithURL((CFURLRef) pdfURL);
numberOfPages = CGPDFDocumentGetNumberOfPages(pdfRef);
// Loop variables
CGPDFPageRef page;
CGRect mediaBox;
// Read the first PDF and generate the output pages
DLog(@"GENERATING PAGES FROM PDF 1 (%@)...", source);
for (int i=1; i<=numberOfPages; i++) {
page = CGPDFDocumentGetPage(pdfRef, i);
mediaBox = CGPDFPageGetBoxRect(page, kCGPDFMediaBox);
CGContextBeginPage(writeContext, &mediaBox);
CGContextDrawPDFPage(writeContext, page);
CGContextEndPage(writeContext);
}
CGPDFDocumentRelease(pdfRef);
CFRelease(pdfURL);
}
CFRelease(pdfURLOutput);
// Finalize the output file
CGPDFContextClose(writeContext);
CGContextRelease(writeContext);
return pdfPathOutput;
}
Hope that helps