XML SAX Parser

SAX stands for Simple API for XML. A SAX parser reads the XML document as a stream and fires event callbacks each time it encounters a start tag, end tag, or text content. It never builds a complete tree in memory. This makes SAX ideal for processing gigabyte-scale XML files where a DOM tree would exhaust available memory.

SAX Event Model

XML Input:
<order id="O1">
  <item>USB Cable</item>
  <price>99</price>
</order>

SAX fires these events in sequence:
startDocument()
startElement("order", {id: "O1"})
  startElement("item", {})
    characters("USB Cable")
  endElement("item")
  startElement("price", {})
    characters("99")
  endElement("price")
endElement("order")
endDocument()

SAX Parser in Python

import xml.sax

class ProductHandler(xml.sax.ContentHandler):

    def __init__(self):
        self.current_element = ""
        self.current_product = {}

    def startElement(self, name, attrs):
        self.current_element = name
        if name == "product":
            self.current_product = {"id": attrs["id"]}

    def characters(self, content):
        content = content.strip()
        if content:
            if self.current_element == "name":
                self.current_product["name"] = content
            elif self.current_element == "price":
                self.current_product["price"] = content

    def endElement(self, name):
        if name == "product":
            print(f"{self.current_product['id']}: "
                  f"{self.current_product['name']} — "
                  f"INR {self.current_product['price']}")
        self.current_element = ""

parser = xml.sax.make_parser()
handler = ProductHandler()
parser.setContentHandler(handler)
parser.parse("catalog.xml")

Key Points to Remember

• SAX parses XML as a stream, firing events — never builds a tree in memory.
• Write handler methods: startElement, endElement, characters.
• SAX is faster and uses far less memory than DOM for large files.
• SAX is forward-only — you cannot go back or access nodes out of order.
• Use SAX when processing very large XML files or when you only need a subset of the data.