How to use NSXMLParser to parse parent-child elements that have the same name

This is a common problem with parsers like this one, of “type SAX”, where you have to manually keep track of the current depth of the XML tree you’re in. The problem, as always, is that loading the entire tree in a DOM structure in memory can be impossible, depending on the size of the data you want to manipulate.

The following code shows a class that does this job:

#import <Foundation/Foundation.h>

@interface Test : NSObject <NSXMLParserDelegate> 
{
@private
    NSXMLParser *xmlParser;
    NSInteger depth;
    NSMutableString *currentName;
    NSString *currentElement;
}

- (void)start;

@end

This is the implementation:

#import "Test.h"

@interface Test ()
- (void)showCurrentDepth;
@end


@implementation Test

- (void)dealloc
{
    [currentElement release];
    [currentName release];
    [xmlParser release];
    [super dealloc];
}

- (void)start
{
    NSString *xml = @"<?xml version=\"1.0\" encoding=\"UTF-8\" ?><Node><name>Main</name><Node><name>Child 1</name></Node><Node><name>Child 2</name></Node></Node>";
    xmlParser = [[NSXMLParser alloc] initWithData:[xml dataUsingEncoding:NSUTF8StringEncoding]];
    [xmlParser setDelegate:self];
    [xmlParser setShouldProcessNamespaces:NO];
    [xmlParser setShouldReportNamespacePrefixes:NO];
    [xmlParser setShouldResolveExternalEntities:NO];
    [xmlParser parse];

}

#pragma mark -
#pragma mark NSXMLParserDelegate methods

- (void)parserDidStartDocument:(NSXMLParser *)parser 
{
    NSLog(@"Document started", nil);
    depth = 0;
    currentElement = nil;
}

- (void)parser:(NSXMLParser *)parser parseErrorOccurred:(NSError *)parseError 
{
    NSLog(@"Error: %@", [parseError localizedDescription]);
}

- (void)parser:(NSXMLParser *)parser didStartElement:(NSString *)elementName 
  namespaceURI:(NSString *)namespaceURI 
 qualifiedName:(NSString *)qName 
    attributes:(NSDictionary *)attributeDict
{
    [currentElement release];
    currentElement = [elementName copy];

    if ([currentElement isEqualToString:@"Node"])
    {
        ++depth;
        [self showCurrentDepth];
    }
    else if ([currentElement isEqualToString:@"name"])
    {
        [currentName release];
        currentName = [[NSMutableString alloc] init];
    }
}

- (void)parser:(NSXMLParser *)parser didEndElement:(NSString *)elementName 
  namespaceURI:(NSString *)namespaceURI 
 qualifiedName:(NSString *)qName
{

    if ([elementName isEqualToString:@"Node"]) 
    {
        --depth;
        [self showCurrentDepth];
    }
    else if ([elementName isEqualToString:@"name"])
    {
        if (depth == 1)
        {
            NSLog(@"Outer name tag: %@", currentName);
        }
        else 
        {
            NSLog(@"Inner name tag: %@", currentName);
        }
    }
}        

- (void)parser:(NSXMLParser *)parser foundCharacters:(NSString *)string
{
    if ([currentElement isEqualToString:@"name"]) 
    {
        [currentName appendString:string];
    } 
}

- (void)parserDidEndDocument:(NSXMLParser *)parser 
{
    NSLog(@"Document finished", nil);
}

#pragma mark -
#pragma mark Private methods

- (void)showCurrentDepth
{
    NSLog(@"Current depth: %d", depth);
}

@end

This is the result of running a command line tool that triggers the “start” method above:

Document started
Current depth: 1
Outer name tag: Main
Current depth: 2
Inner name tag: Child 1
Current depth: 1
Current depth: 2
Inner name tag: Child 2
Current depth: 1
Current depth: 0
Document finished

Leave a Comment