C vs Ruby for SAX Parsing

results

# Data file is protein sequence database psd7003.xml (683MB) from http://www.cs.washington.edu/research/xmldatasets/www/repository.html#pir
# Structure is list of ~260,000 <ProteinEntry> elements
# Goal is to read data file, and process each <ProteinEntry> onto its own line of output
# This simulates filtering/splitting a large list of similar XML elements 

$ time ruby xmlsplit.rb test.xml > output_rb

real	2m35.965s
user	2m28.158s
sys	0m3.340s

$ time ./xmlsplit test.xml > output_c

real	0m18.755s
user	0m17.029s
sys	0m1.667s

xmlsplit.c

#include <stdio.h>
#include <string.h>
#include <libxml/parser.h>

int append = 0;

char data[100000];
int idx = 0;

void OnStartElementNs(
    void *ctx,
    const xmlChar *localname,
    const xmlChar *prefix,
    const xmlChar *URI,
    int nb_namespaces,
    const xmlChar **namespaces,
    int nb_attributes,
    int nb_defaulted,
    const xmlChar **attributes
)
{
    if (strcmp(localname, "ProteinEntry") == 0) {
        append = 1;
    }
    if (!append) return;
    idx += sprintf(data + idx, "<%s>", localname);
}

void OnEndElementNs(void* ctx, const xmlChar* localname, 
                    const xmlChar* prefix, const xmlChar* URI)
{
    if (!append) return;
    idx += sprintf(data + idx, "</%s>", localname);
    if (strcmp(localname, "ProteinEntry") == 0) {
        append = 0;
        idx = 0;
        printf("%s\n", data);
    }
}

void OnCharacters(void *ctx, const xmlChar *ch, int len) 
{
    if (!append || len == 1) return;
    char *s = malloc(len + 1);
    strncpy(s, (const char *)ch, len);
    s[len] = (char) NULL;
    idx += sprintf(data + idx, "%s", s);
}

xmlSAXHandler make_sax_handler()
{
    xmlSAXHandler handler;
    memset(&handler, 0, sizeof(xmlSAXHandler));
    handler.initialized = XML_SAX2_MAGIC;
    handler.startElementNs = OnStartElementNs;
    handler.endElementNs = OnEndElementNs;
    handler.characters = OnCharacters;
    return handler;
}

int read_xmlfile(FILE *f)
{
    char chars[1024];
    int res = fread(chars, 1, 4, f);
    if (res <= 0) {
        return 1;
    }
    xmlSAXHandler handler = make_sax_handler();
    xmlParserCtxtPtr ctxt = xmlCreatePushParserCtxt(
        &handler, NULL, chars, res, NULL
    );
    while ((res = fread(chars, 1, sizeof(chars), f)) > 0) {
        if (xmlParseChunk(ctxt, chars, res, 0)) {
            xmlParserError(ctxt, "xmlParseChunk");
            return 1;
        }
    }
    xmlParseChunk(ctxt, chars, 0, 1);
    xmlFreeParserCtxt(ctxt);
    xmlCleanupParser();
    return 0;
}

int main(int argc, char **argv)
{
    if (argc != 2) {
        printf("Expected one argument 'filepath'\n");
        exit(-1);
    }
    FILE *f = fopen(argv[1], "r");
    read_xmlfile(f);
    fclose(f);
    return 0;
}

xmlsplit.rb

require 'nokogiri'

class Parser < Nokogiri::XML::SAX::Document
    def initialize
        @buffer = []
        @append = false
    end

    def start_element(name, attrs=[])
        @append = true if name == "ProteinEntry"
        return unless @append
        @buffer << "<#{name}>"
    end

    def characters(string)
        return unless @append
        @buffer << string
    end

    def end_element(name)
        return unless @append
        @buffer << "</#{name}>"
        if name == "ProteinEntry"
            @append = false
            puts @buffer.join ''
            @buffer = []
        end
    end
end

Nokogiri::XML::SAX::Parser.new(Parser.new).parse(File.open(ARGV[0]))