Convert XML text to Javascript object / JSON text (and vice versa).
There are many XML to JavaScript object / JSON converters out there, but could not satisfy the following requirements:
-
Maintain Order of Elements: Instead of converting
<a/><b/><a/>
to{a:[{},{}],b:{}}
, I wanted to preserve order of elements by doing this:{"elements":[{"type":"element","name":"a"},{"type":"element","name":"b"},{"type":"element","name":"a"}]}
. -
Fully XML Compliant: Can parse: elements, attributes, texts, comments, CData, DOCTYPE, XML declarations, and Processing Instructions.
-
Reversible: Whether converting xml→json or json→xml, the result can be converted back to its original form.
-
Minimal Dependencies: This library depends only on one external npm module.
-
Change Property Key Name: Usually output of XML attributes are stored in
@attr
,_atrr
,$attr
,$
, orwhatever
in order to avoid conflicting with name of sub-elements. This library store them inattributes
, but most importantly, you can change this to whatever you like. -
Support Upwards Traversal: By setting
{addParent: true}
option, an extra property namedparent
will be generated along each element so that its parent can be referenced. Therefore, anywhere during the traversal of an element, its children and its parent can be easily accessed. -
Portable Code: Written purely in JavaScript which means it can be used in Node environment and browser environment (via bundlers like browserify/JSPM/Webpack).
-
Support Command Line: To quickly convert xml or json files, this module can be installed globally or locally (i.e. use it as script in package.json).
-
Typings Info Included: Support type checking and code suggestion via intellisense. Thanks to the wonderful efforts by DenisCarriere
Most XML to JSON converters (including online converters) convert <a/>
to some compact output like {"a":{}}
instead of non-compact output like {"elements":[{"type":"element","name":"a"}]}
.
While compact output might work in most situations, there are cases when elements of different names are mixed inside a parent element. Lets use <a x="1"/><b x="2"/><a x="3"/>
as an example.
Most converters will produce compact like {a:[{_:{x:"1"}},{_:{x:"3"}}], b:{_:{x:"2"}}}
,
which has merged both <a>
elements into an array! If you try to convert this back to xml, you will get <a x="1"/><a x="3"/><b x="2"/>
which has not preserved the order of elements! This is an inherent limitation in the compact representation
because output like {a:{_:{x:"1"}}, b:{_:{x:"2"}}, a:{_:{x:"3"}}}
is illegal (same property name a
should not appear twice in an object).
The non-compact output, which is supported by this library, will produce more information and always guarantees the order of the elements as they appeared in the XML file.
Another drawback of compact output is the resultant element can be an object or an array and therefore makes the client code a little awkwards in terms of the extra check needed on object type before processing.
NOTE: Although non-compact output is more accurate representation of original XML than compact version, the non-compact version is verbose and consumes more space.
This library provides both options. Use {compact: false}
if you are not sure because it preserves everything;
otherwise use {compact: true}
if you want to save space and you don't care about mixing elements of same type and losing their order.
npm install --save xml-js
You can also install it globally to use it as a command line convertor (see Command Line).
npm install --global xml-js
var convert = require('xml-js');
var xml =
'<?xml version="1.0" encoding="utf-8"?>' +
'<note importance="high" logged="true">' +
' <title>Happy</title>' +
' <todo>Work</todo>' +
' <todo>Play</todo>' +
'</note>';
var result1 = convert.xml2json(xml, {compact: true, spaces: 4});
var result2 = convert.xml2json(xml, {compact: false, spaces: 4});
console.log(result1, '\n', result2);
To see the result of this code, see the output above in Synopsis section.
Or run and edit this code live in the browser.
XML | JS/JSON compact | JS/JSON non-compact |
---|---|---|
<a/> |
{"a":{}} |
{"elements":[{"type":"element","name":"a"}]} |
<a/><b/> |
{"a":{},"b":{}} |
{"elements":[{"type":"element","name":"a"},{"type":"element","name":"b"}]} |
<a><b/></a> |
{"a":{"b":{}}} |
{"elements":[{"type":"element","name":"a","elements":[{"type":"element","name":"b"}]}]} |
<a> Hi </a> |
{"a":{"_text":" Hi "}} |
{"elements":[{"type":"element","name":"a","elements":[{"type":"text","text":" Hi "}]}]} |
<a x="1.234" y="It's"/> |
{"a":{"_attributes":{"x":"1.234","y":"It's"}}} |
{"elements":[{"type":"element","name":"a","attributes":{"x":"1.234","y":"It's"}}]} |
<?xml?> |
{"_declaration":{}} |
{"declaration":{}} |
<?go there?> |
{"_instruction":{"go":"there"}} |
{"elements":[{"type":"instruction","name":"go","instruction":"there"}]} |
<?xml version="1.0" encoding="utf-8"?> |
{"_declaration":{"_attributes":{"version":"1.0","encoding":"utf-8"}}} |
{"declaration":{"attributes":{"version":"1.0","encoding":"utf-8"}}} |
<!--Hello, World!--> |
{"_comment":"Hello, World!"} |
{"elements":[{"type":"comment","comment":"Hello, World!"}]} |
<![CDATA[<foo></bar>]]> |
{"_cdata":"<foo></bar>"} |
{"elements":[{"type":"cdata","cdata":"<foo></bar>"}]} |
This library provides 4 functions: js2xml()
, json2xml()
, xml2js()
, and xml2json()
. Here are the usages for each one (see more details in the following sections):
var convert = require('xml-js');
result = convert.js2xml(js, options); // to convert javascript object to xml text
result = convert.json2xml(json, options); // to convert json text to xml text
result = convert.xml2js(xml, options); // to convert xml text to javascript object
result = convert.xml2json(xml, options); // to convert xml text to json text
To convert JavaScript object to XML text, use js2xml()
. To convert JSON text to XML text, use json2xml()
.
var convert = require('xml-js');
var json = require('fs').readFileSync('test.json', 'utf8');
var options = {ignoreComment: true, spaces: 4};
var result = convert.json2xml(json, options);
console.log(result);
The below options are applicable for both js2xml()
and json2xml()
functions.
Option | Default | Description |
---|---|---|
spaces |
0 |
Number of spaces to be used for indenting XML output. Passing characters like ' ' or '\t' are also accpeted. |
compact |
false |
Whether the input object is in compact form or not. |
fullTagEmptyElement |
false |
Whether to produce element without sub-elements as full tag pairs <a></a> rather than self closing tag <a/> . |
indentCdata |
false |
Whether to write CData in a new line and indent it. Will generate <a>\n <![CDATA[foo]]></a> instead of <a><![CDATA[foo]]></a> . See discussion |
ignoreDeclaration |
false |
Whether to ignore writing declaration directives of xml. For example, <?xml?> will be ignored. |
ignoreInstruction |
false |
Whether to ignore writing processing instruction of xml. For example, <?go there?> will be ignored. |
ignoreAttributes |
false |
Whether to ignore writing attributes of the elements. For example, x="1" in <a x="1"></a> will be ignored |
ignoreComment |
false |
Whether to ignore writing comments of the elements. That is, no <!-- --> will be generated. |
ignoreCdata |
false |
Whether to ignore writing CData of the elements. That is, no <![CDATA[ ]]> will be generated. |
ignoreDoctype |
false |
Whether to ignore writing Doctype of the elements. That is, no <!DOCTYPE > will be generated. |
ignoreText |
false |
Whether to ignore writing texts of the elements. For example, hi text in <a>hi</a> will be ignored. |
To convert XML text to JavaScript object, use xml2js()
. To convert XML text to JSON text, use xml2json()
.
var convert = require('xml-js');
var xml = require('fs').readFileSync('test.xml', 'utf8');
var options = {ignoreComment: true, alwaysChildren: true};
var result = convert.xml2js(xml, options); // or convert.xml2json(xml, options)
console.log(result);
The below options are applicable for both xml2js()
and xml2json()
functions.
Option | Default | Description |
---|---|---|
compact |
false |
Whether to produce detailed object or compact object. |
trim |
false |
Whether to trim whitespace characters that may exist before and after the text. |
sanitize (Deprecated) |
false |
Whether to replace & < > with & < > respectively, in the resultant text. |
nativeType |
false |
Whether to attempt converting text of numerals or of boolean values to native type. For example, "123" will be 123 and "true" will be true |
addParent |
false |
Whether to add parent property in each element object that points to parent object. |
alwaysArray |
false |
Whether to always put sub element, even if it is one only, as an item inside an array. <a><b/></a> will be a:[{b:[{}]}] rather than a:{b:{}} (applicable for compact output only). |
alwaysChildren |
false |
Whether to always generate elements property even when there are no actual sub elements. <a></a> will be {"elements":[{"type":"element","name":"a","elements":[]}]} rather than {"elements":[{"type":"element","name":"a"}]} (applicable for non-compact output). |
instructionHasAttributes |
false |
Whether to parse contents of Processing Instruction as attributes or not. <?go to="there"?> will be {"_instruction":{"go":{"_attributes":{"to":"there"}}}} rather than {"_instruction":{"go":"to=\"there\""}} . See discussion. |
ignoreDeclaration |
false |
Whether to ignore parsing declaration property. That is, no declaration property will be generated. |
ignoreInstruction |
false |
Whether to ignore parsing processing instruction property. That is, no instruction property will be generated. |
ignoreAttributes |
false |
Whether to ignore parsing attributes of elements.That is, no attributes property will be generated. |
ignoreComment |
false |
Whether to ignore parsing comments of the elements. That is, no comment will be generated. |
ignoreCdata |
false |
Whether to ignore parsing CData of the elements. That is, no cdata will be generated. |
ignoreDoctype |
false |
Whether to ignore parsing Doctype of the elements. That is, no doctype will be generated. |
ignoreText |
false |
Whether to ignore parsing texts of the elements. That is, no text will be generated. |
The below option is applicable only for xml2json()
function.
Option | Default | Description |
---|---|---|
spaces |
0 |
Number of spaces to be used for indenting JSON output. Passing characters like ' ' or '\t' are also accepted. |
To change default key names in the output object or the default key names assumed in the input JavaScript object / JSON, use the following options:
Option | Default | Description |
---|---|---|
declarationKey |
"declaration" or "_declaration" |
Name of the property key which will be used for the declaration. For example, if declarationKey: '$declaration' then output of <?xml?> will be {"$declaration":{}} (in compact form) |
instructionKey |
"instruction" or "_instruction" |
Name of the property key which will be used for the processing instruction. For example, if instructionKey: '$instruction' then output of <?go there?> will be {"$instruction":{"go":"there"}} (in compact form) |
attributesKey |
"attributes" or "_attributes" |
Name of the property key which will be used for the attributes. For example, if attributesKey: '$attributes' then output of <a x="hello"/> will be {"a":{$attributes:{"x":"hello"}}} (in compact form) |
textKey |
"text" or "_text" |
Name of the property key which will be used for the text. For example, if textKey: '$text' then output of <a>hi</a> will be {"a":{"$text":"Hi"}} (in compact form) |
cdataKey |
"cdata" or "_cdata" |
Name of the property key which will be used for the cdata. For example, if cdataKey: '$cdata' then output of <![CDATA[1 is < 2]]> will be {"$cdata":"1 is < 2"} (in compact form) |
doctypeKey |
"doctype" or "_doctype" |
Name of the property key which will be used for the doctype. For example, if doctypeKey: '$doctype' then output of <!DOCTYPE foo> will be {"$doctype":" foo} (in compact form) |
commentKey |
"comment" or "_comment" |
Name of the property key which will be used for the comment. For example, if commentKey: '$comment' then output of <!--note--> will be {"$comment":"note"} (in compact form) |
parentKey |
"parent" or "_parent" |
Name of the property key which will be used for the parent. For example, if parentKey: '$parent' then output of <a></b></a> will be {"a":{"b":{$parent:_points_to_a}}} (in compact form) |
typeKey |
"type" |
Name of the property key which will be used for the type. For example, if typeKey: '$type' then output of <a></a> will be {"elements":[{"$type":"element","name":"a"}]} (in non-compact form) |
nameKey |
"name" |
Name of the property key which will be used for the name. For example, if nameKey: '$name' then output of <a></a> will be {"elements":[{"type":"element","$name":"a"}]} (in non-compact form) |
elementsKey |
"elements" |
Name of the property key which will be used for the elements. For example, if elementsKey: '$elements' then output of <a></a> will be {"$elements":[{"type":"element","name":"a"}]} (in non-compact form) |
Two default values mean the first is used for non-compact output and the second is for compact output.
TIP: You probably want to set
{textKey: 'value', cdataKey: 'value', commentKey: 'value'}
for non-compact output to make it more consistent and easier for your client code to go through the contents of text, cdata, and comment.
Because any good library should support command line usage, this library is no difference.
npm install -g xml-js // install this library globally
xml-js test.json --spaces 4 // xml result will be printed on screen
xml-js test.json --spaces 4 --out test.xml // xml result will be saved to test.xml
xml-js test.xml --spaces 4 // json result will be printed on screen
xml-js test.xml --spaces 4 --out test.json // json result will be saved to test.json
If you want to use it as script in package.json (can also be helpful in task automation via npm scripts)
npm install --save xml-js // no need to install this library globally
In package.json, write a script:
...
"dependencies": {
"xml-js": "latest"
},
"scripts": {
"convert": "xml-js test.json --spaces 4"
}
npm run convert // task 'scripts.convert' will be executed
Usage: xml-js src [options]
src Input file that need to be converted.
Conversion type xml->json or json->xml will be inferred from file extension.
Options:
--help, -h Display this help content.
--version, -v Display version number of this module.
--out Output file where result should be written.
--spaces Specifies amount of space indentation in the output.
--full-tag XML elements will always be in <a></a> form.
--no-decl Declaration directive <?xml?> will be ignored.
--no-inst Processing instruction <?...?> will be ignored.
--no-attr Attributes of elements will be ignored.
--no-text Texts of elements will be ignored.
--no-cdata CData of elements will be ignored.
--no-doctype DOCTYPE of elements will be ignored.
--no-comment Comments of elements will be ignored.
--trim Any whitespaces surrounding texts will be trimmed.
--compact JSON is in compact form.
--native-type Numbers and boolean will be converted (coerced) to native type instead of text.
--always-array Every element will always be an array type (applicable if --compact is set).
--always-children Every element will always contain sub-elements (applicable if --compact is not set).
--text-key To change the default 'text' key.
--cdata-key To change the default 'cdata' key.
--doctype-key To change the default 'doctype' key.
--comment-key To change the default 'comment' key.
--attributes-key To change the default 'attributes' key.
--declaration-key To change the default 'declaration' key.
--instruction-key To change the default 'processing instruction' key.
--type-key To change the default 'type' key (applicable if --compact is not set).
--name-key To change the default 'name' key (applicable if --compact is not set).
--elements-key To change the default 'elements' key (applicable if --compact is not set).
To perform tests on this project, download the full repository from GitHub (not from npm) and then do the following:
cd xml-js
npm install
npm test
For live testing, use npm start
instead of npm test
.
Use this link to report an issue or bug. Please include a sample code where the code is failing.