fast-xml-parser

Validate XML, Parse XML to JS/JSON and vice versa, or parse XML to Nimn rapidly without C/C++ based libraries and no callback

This project welcomes contributors. If you have a feature you'd like to see implemented or a bug you'd liked fixed, the best and fastest way to make that happen is to implement it and submit a PR. Basic knowledge of JS is sufficient. Feel free to ask for any guidance.

Users

List of some applications/projects using Fast XML Parser. (Raise an issue to submit yours)

Join this project as collaborator / maintainer.

Main Features

• Validate XML data syntactically
• Transform XML to JSON or Nimn
• Transform JSON back to XML
• Works with node packages, in browser, and in CLI (press try me button above for demo)
• Faster than any pure JS implementation.
• It can handle big files (tested up to 100mb).
• Various options are available to customize the transformation
  • You can parse CDATA as separate property.
  • You can prefix attributes or group them to separate property. Or can ignore them from result completely.
  • You can parse tag's or attribute's value to primitive type: string, integer, float, hexadecimal, or boolean. And can optionally decode for HTML char.
  • You can remove namespace from tag or attribute name while parsing
  • It supports boolean attributes, if configured.

How to use

To use it in NPM package install it first

$npm install fast-xml-parser or using yarn$yarn add fast-xml-parser

To use it from CLI Install it globally with -g option.

$npm install fast-xml-parser -g

To use it on a webpage include it from a CDN

XML to JSON

varjsonObj=parser.parse(xmlData[,options]);

varparser=require('fast-xml-parser');varhe=require('he');varoptions={attributeNamePrefix : "@_",attrNodeName: "attr",//default is 'false'textNodeName : "#text",ignoreAttributes : true,ignoreNameSpace : false,allowBooleanAttributes : false,parseNodeValue : true,parseAttributeValue : false,trimValues: true,cdataTagName: "__cdata",//default is 'false'cdataPositionChar: "\\c",localeRange: "",//To support non english character in tag/attribute values.parseTrueNumberOnly: false,attrValueProcessor: a=>he.decode(a,{isAttributeValue: true}),//default is a=>atagValueProcessor : a=>he.decode(a)//default is a=>a};if(parser.validate(xmlData)===true){//optional (it'll return an object in case it's not valid)varjsonObj=parser.parse(xmlData,options);}// Intermediate objvartObj=parser.getTraversalObj(xmlData,options);varjsonObj=parser.convertToJson(tObj,options);

Note: he library is used in this example

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 : This options has been removed from 3.3.4. Instead, use tagValueProcessor, and attrValueProcessor. See above example.
• cdataTagName : If specified, parser parse CDATA as nested tag instead of adding it's value to parent tag.
• cdataPositionChar : It'll help to covert JSON back to XML without losing CDATA position.
• localeRange: Parser will accept non-English character in tag or attribute name. Check #87 for more detail. Eg localeRange: "a-zA-Zа-яёА-ЯЁ"
• parseTrueNumberOnly: if true then values like "+123", or "0123" will not be parsed as number.
• tagValueProcessor : Process tag value during transformation. Like HTML decoding, word capitalization, etc. Applicable in case of string only.
• attrValueProcessor : Process attribute value during transformation. Like HTML decoding, word capitalization, etc. Applicable in case of string only.
• stopNodes : an array of tag names which are not required to be parsed. Instead their values are parsed as string.

To use from command line

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

• -ns : To include namespaces (by default ignored)
• -a : To ignore attributes
• -c : To ignore value conversion (i.e. "-3" will not be converted to number -3)
• -v : validate before parsing
• -V : only validate

To use it on webpage

varresult=parser.validate(xmlData);if(result!==true)console.log(result.err);varjsonObj=parser.parse(xmlData);

JSON / JS Object to XML

varParser=require("fast-xml-parser").j2xParser;//default options need not to setvardefaultOptions={attributeNamePrefix : "@_",attrNodeName: "@",//default is falsetextNodeName : "#text",ignoreAttributes : true,cdataTagName: "__cdata",//default is falsecdataPositionChar: "\\c",format: false,indentBy: " ",supressEmptyNode: false,tagValueProcessor: a=>he.encode(a,{useNamedReferences: true}),// default is a=>aattrValueProcessor: a=>he.encode(a,{isAttributeValue: isAttribute,useNamedReferences: true})// default is a=>a};varparser=newParser(defaultOptions);varxml=parser.parse(json_or_js_obj);

OPTIONS :

With the correct options, you can get the almost original XML without losing any information.

• attributeNamePrefix : Identify attributes with this prefix otherwise treat them as a tag.
• attrNodeName: Identify attributes when they are grouped under single property.
• ignoreAttributes : Don't check for attributes. Treats everything as tag.
• encodeHTMLchar : This option has been removed from 3.3.4. Use tagValueProcessor, and attrValueProcessor instead. See above example.
• cdataTagName : If specified, parse matching tag as CDATA
• cdataPositionChar : Identify the position where CDATA tag should be placed. If it is blank then CDATA will be added in the last of tag's value.
• format : If set to true, then format the XML output.
• indentBy : indent by this char when format is set to true
• supressEmptyNode : If set to true, tags with no value (text or nested tags) are written as self closing tags.
• tagValueProcessor : Process tag value during transformation. Like HTML encoding, word capitalization, etc. Applicable in case of string only.
• attrValueProcessor : Process attribute value during transformation. Like HTML encoding, word capitalization, etc. Applicable in case of string only.

Benchmark

XML to JSON

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.

JSON to XML

report

file size	fxp 3.2 js to xml	xml2js 0.4.19 builder
1.3k	160148.9801	10384.99401
1.1m	173.6374831	8.611884025

Limitations

Currently FXP fails to parse XML with attributes has ">" in the value. This problem is left open as change in regex for its fix is degrading the performance. And the parser become very slow in case of long attrbute names.

Worth to mention

• BigBit standard) : A standard to reprent any number in the universe in comparitively less space and without precision loss. A standard to save space to represent any text string in comparision of UTF encoding.
• imglab : Speedup and simplify image labeling / annotation. Supports multiple formats, one click annotation, easy interface and much more. There are more than 20k images are annotated every month.
• अनुमार्गक (anumargak) : The fastest and simple router for node js web frameworks with many unique features.
• stubmatic : A stub server to mock behaviour of HTTP(s) / REST / SOAP services, incuding DynamoDB calls. You can also mock binary formats.
• मुनीम (Muneem) : A webframework made for all team members. Faster tha fastify, express, koa, hapi and others.
• शब्दावली (shabdawali) : Amazing human like typing effects beyond your imagination.

Contributors

This project exists thanks to all the people who contribute. [Contribute].

Backers

Thank you to all our backers! 🙏 [Become a backer]

Sponsors

Support this project by becoming a sponsor. Your logo will show up here with a link to your website. [Become a sponsor]