fast-xml-parser

Validate XML or Parse XML to JS/JSON very fast without C/C++ based libraries and no callback

**Note**: If you are using v3, your code may start failing in parsing and validation both. I apologize for the breaking changes. But code was supposed to be changed to support large files and many other options. Please refer the code example below for more detail.

You can use this library online (press try me button above), or as command from CLI, or in your website, or in npm repo.

• This library let you validate the XML data syntactically.
• Or you can transform/covert/parse XML data to JS/JSON object.
• Or you can transform the XML in traversable JS object which can later be converted to JS/JSON object.

How to use

Installation

$npm install fast-xml-parser

or using yarn

$yarn add fast-xml-parser

Usage

var fastXmlParser = require('fast-xml-parser');
var jsonObj = fastXmlParser.parse(xmlData);

/* upto 2.9.x
var options = {
    attrPrefix : "@_",
    attrNodeName: false,
    textNodeName : "#text",
    ignoreNonTextNodeAttr : true,
    ignoreTextNodeAttr : true,
    ignoreNameSpace : true,
    ignoreRootElement : false,
    textNodeConversion : true,
    textAttrConversion : false,
    arrayMode : false
};
*/
//from 3.0.0
var options = {
    attributeNamePrefix : "@_",
    attrNodeName: false,
    textNodeName : "#text",
    ignoreAttributes : true,
    ignoreNameSpace : false,
    allowBooleanAttributes : false,
    parseNodeValue : true,
    parseAttributeValue : false,
    trimValues: true,
    decodeHTMLchar: false,
};
if(fastXmlParser.validate(xmlData)=== true){//optional
	var jsonObj = fastXmlParser.parse(xmlData,options);
}

//Intermediate obj
var tObj = fastXmlParser.getTraversalObj(xmlData,options);
var jsonObj = fastXmlParser.convertToJson(tObj,options);

OPTIONS :

• attributeNamePrefix : prepend given string to attribute name for identification
• attrNodeName: (Valid name) Group all the attributes as properties of given name.
• ignoreAttributes : Ignore attributes to be parsed.
• ignoreNameSpace : Remove namespace string from tag and attribute names.
• allowBooleanAttributes : a tag can have attributes without any value
• parseNodeValue : Parse the value of text node to float, integer, or boolean.
• parseAttributeValue : Parse the value of an attribute to float, integer, or boolean.
• trimValues : trim string values of an attribute or node
• decodeHTMLchar : decodes any named and numerical character HTML references excluding CDATA part.

To use from command line

$xml2js [-ns|-a|-c] <filename> [-o outputfile.json]
$cat xmlfile.xml | xml2js [-ns|-a|-c] [-o outputfile.json]

-ns : To include namespaces (bedefault ignored) -a : To ignore attributes -c : To ignore value conversion (i.e. "-3" will not be converted to number -3)

To use it on webpage

1. Download and include parser.js

var result = parser.validate(xmlData);
if(result !== true) cnosole.log(result.err);
var jsonObj = parser.parse(xmlData);

Or use directly from CDN

Comparision

I decided to created this library when I couldn't find any library which can convert XML data to json without any callback and which is not based on any C/C++ library.

Libraries that I compared

• xml-mapping : fast, result is not satisfactory
• xml2js : fast, result is not satisfactory
• xml2js-expat : couldn't test performance as it gives error on high load. Installation failed on travis and on my local machine using 'yarn'.
• xml2json : based on node-expat which is based on C/C++. Installation failed on travis.
• fast-xml-parser : very very fast.

Why not C/C++ based libraries? Installation of such libraries fails on some OS. You may require to install missing dependency manually.

Benchmark report

file size	fxp 3.0 validator (rps)	fxp 3.0 parser (rps)	xml2js 0.4.19 (rps)
1.5k	16581.06758	14032.09323	4615.930805
1.5m	14918.47793	13.23366098	5.90682005
13m	1.834479235	1.135582008	-1
1.3k with CDATA	30583.35319	43160.52342	8398.556349
1.3m with CDATA	27.29266471	52.68877009	7.966000795
1.6k with cdata,prolog,doctype	27690.26082	41433.98547	7872.399268
98m	0.08473858148	0.2600104004	-1

• -1 indicates error or incorrect output.

Changes from v3

• It can handle big file now (I have tested up to 98mb). Performance report is given above.
• Meaningful error messages from validator

"err": {
    "code": "InvalidAttr",
    "msg": "Attributes for rootNode have open quote"
}

• Updated options : check snippet aboove
• Parse boolean values as well. E.g. "true" to true
• You can set pasrer not to trim whitespaces from attribute or tag /node value.
• You can set pasrer to HTML decode Tag / node and attribute values. However CDATA value will not be HTML decoded.
• Tag / node value will not be parsed if CDATA presents.
• You can set validator and parser to allow boolean values.
• Few validation and parsing bugs are also fixed

Some of my other NPM pojects

• stubmatic : A stub server to mock behaviour of HTTP(s) / REST / SOAP services. Stubbing redis is on the way.
• fast-lorem-ipsum : Generate lorem ipsum words, sentences, paragraph very quickly.

TODO

• P2: validating XML stream data
• P2: validator cli
• P2: fast XML prettyfier