Click here to Skip to main content
Click here to Skip to main content

JSON Spirit: A C++ JSON Parser/Generator Implemented with Boost Spirit

, 10 May 2014
Rate this:
Please Sign up or sign in to vote.
A C++ JSON parser/generator written using Boost::spirit

Table of Contents

Introduction

JSON is a text file format similar to XML, but less verbose. It has been called "XML lite". This article describes JSON Spirit, a C++ library that reads and writes JSON files or streams. It is written using the Boost Spirit parser generator. If you are already using Boost, you can use JSON Spirit without any additional dependencies.

Key features:

  • supports ASCII or Unicode
  • std::vector or std::map implementations for JSON Objects
  • object library or header file only use

Using the Code

The JSON Spirit source code is available as a Microsoft Visual Studio 2005 C++ "solution". However, it should compile and work on any platform compatible with Boost. JSON Spirit has been built and tested with Visual C++ 2005, 2008, 2010, and G++ version 4.2.3 on Linux. It has been tested with Visual C++ using Boost versions 1.34.0, 1.37.0, 1.39.0, and 1.42.0. It has also been tested with STLPort.

Platform independent CMake files are included, kindly supplied by Uwe Arzt, with CPack declarations for packaging and a "make install" target supplied by Amzlkej.

The Visual C++ solution consists of five projects:

  • The JSON Spirit library and header files
  • An application running the library's unit tests
  • Three small programs demonstrating how to use JSON Spirit

JSON Spirit Value

A JSON value can hold either a JSON array, JSON object, string, integer, double, bool, or null. The interface of the JSON Spirit Value class is shown below. The Value class for Unicode is analogous; for details, see the section on Unicode support.

enum Value_type{ obj_type, array_type, str_type,
     bool_type, int_type, real_type, null_type };

class Value
{
    public:

        Value();  // creates null value

        Value( const char*          value );
        Value( const std::string&   value );
        Value( const Object&        value );
        Value( const Array&         value );
        Value( bool                 value );
        Value( int                  value );
        Value_impl( boost::int64_t  value );
        Value_impl( boost::uint64_t value );
        Value( double               value );

        bool operator==( const Value& lhs ) const;

        Value_type type() const;

        const std::string& get_str()   const;
        const Object&      get_obj()   const;
        const Array&       get_array() const;
        bool               get_bool()  const;
        int                get_int()   const;
        boost::int64_t     get_int64()  const;
        boost::uint64_t    get_uint64() const;
        double             get_real()  const;

        Object& get_obj();
        Array&  get_array();

        template< typename > T get_value() const;

        bool is_uint64() const;
        bool is_null() const;

        static const Value null;

    private:

        ...
};

You obtain the Value's type by calling Value::type(). You can then call the appropriate getter function. Generally, you will know a file's format, so you will know what type the JSON values should have. A std::runtime_error exception is thrown if you try to get a value of the wrong type, for example, if you try to extract a string from a value containing an integer.

The template getter function get_value() is an alternative to get_int(), get_real(), etc. Example usage would be:

int    i = value_1.get_value< int >();
double d = value_2.get_value< double >();

A top level Value read from a file or stream normally contains an Array or an Object. An Array is a std::vector of values. An Object is, by default, a std::vector of JSON pairs.

typedef std::vector< Pair > Object;
typedef std::vector< Value > Array;

A Pair is a structure that holds a std::string and a Value.

struct Pair
{
    Pair( const std::string& name, const Value& value );

    bool operator==( const Pair& lhs ) const;

    std::string name_;
    Value value_;
};

JSON Arrays and Objects can themselves contain other Arrays or Objects, forming a tree.

Note, JSON Spirit provides an alternative std::map based Object; see the Map Implementation section below. In this case, a Pair type is not needed as an mObject is a std::map of names to values.

typedef std::map< std::string, mValue > mObject;

Reading JSON

You can read JSON data from a stream or a string:

bool read( const std::string&  s, Value& value );
bool read( std::istream&  is,     Value& value );

For example:

ifstream is( "json.txt" );
Value value;
read( is, value );

You can also read JSON data by supplying a pair of string iterators.

bool read( std::string::const_iterator& begin,
           std::string::const_iterator  end, Value& value );

After a successful read, the iterator "begin" will point one past the last character of the text for the value just read. This allows the decoding of a string containing multiple top level values. A subsequent call to read will read the next value in the string.

Similarly the stream reading functions now allow a sequence of top-level values to be read one at a time. Before version 3.0, a stream was converted to a string before being parsed. This was fine for files, but not if, for example, you wanted to read multiple JSON values from a socket.

Writing JSON

To output JSON, you first create a Value object containing your data, then write the created Value to a stream or string. For example, the following code will create a small JSON file containing an object with three members:

Object addr_obj;

addr_obj.push_back( Pair( "house_number", 42 ) );
addr_obj.push_back( Pair( "road",         "East Street" ) );
addr_obj.push_back( Pair( "town",         "Newtown" ) );

ofstream os( "address.txt" );

write( addr_obj, os, pretty_print );

os.close();

The object addr_obj is automatically converted into a Value as it is passed to the write function. The file address.txt will contain:

{
    "house_number" : 42,
    "road" : "East Street",
    "town" : "Newtown"
}

There are two versions of each function, one to write to strings, the other to write to files:

std::string write( const Value&  value, Output_options options = none, unsigned int precision_of_doubles = 0 );
void write( const Value& value, std::ostream&  os, Output_options options = none, unsigned int precision_of_doubles = 0 );

The options parameter controls how the data is written:

enum Output_options
{  
    none = 0,
    pretty_print = 0x01, 
    raw_utf8 = 0x02, 
    remove_trailing_zeros = 0x04,
    single_line_arrays = 0x08,
    always_escape_nonascii = 0x10   
}; 

The pretty_print option causes white-space and line breaks to be added to the JSON text.

The raw_utf8 option is an extension to the JSON standard. It disables the escaping of non-printable characters, allowing UTF-8 sequences held in 8 bit char strings to pass through unaltered.

The remove_trailing_zeros option is really only present for backwards compatibility purposes. It is no longer needed as trailing spaces are now removed by default. The only effect it now has is to set the default number of decimal places used for outputting double to 16 instead of the normal default of 17.

The single_line_arrays option has the same effect as the pretty printing option except that arrays printed on single lines unless they contain composite elements, i.e. objects or arrays. This produced output like:

{
    "1" : [ 0, 0, 0, 0, 0 ],
    "2" : [
        [ 0, 0, 0, 0, 0 ],
        [ 0, 1, 2, 3, 4 ],
        [ 0, 2, 4, 6, 8 ],
        [ 0, 3, 6, 9, 12 ],
        [ 0, 4, 8, 12, 16 ]
    ]
}

The always_escape_nonascii option means that all unicode wide characters are escaped, i.e. outputted as "\uXXXX", even if they are printable under the current locale, ASCII printable chars are not escaped.

The precision_of_doubles parameter determines the precision used when outputting doubles. The default is 17.

You can apply more than one option by using the OR operator, i.e.:

write( addr_obj, os, pretty_print | raw_utf8 ); 

The default is to write the JSON text without any white-space and to escape non-printable characters. There is also a function called write_formatted. This is equivalent to:

write( addr_obj, os, pretty_print );

Error Detection

From version 3.00, JSON Spirit provides functions that report the position of format errors in the text being parsed. These functions are identical to the normal read functions except that instead of returning false if an error is found, they throw a json_spirit::Error_position exception. Note, these functions run about three times slower than the normal read functions.

The Error_position structure holds the line and column number where the first error was found.

struct Error_position
{
    ...
    unsigned int line_;
    unsigned int column_;
    std::string reason_;
};

Unicode Support

Unicode support is provided by std::wstring versions of the JSON Spirit Value, Array, Object, and Pair types. These are called wValue, wArray, wObject, and wPair. There are also std::wstring versions of each reader and writer function.

Note that there is no support for reading Unicode files and converting them to wstrings as this is not a task specific to JSON.

The Value and wValue classes are actually instantiations of the template class Value_impl.

Std::map Implementation

Before version 4.00, the JSON Spirit Object type was a std::vector of name/value Pairs. You now have the option of using mObject which is a name/value std::map. For the std::map version, use mValue instead of Value, mObject instead of Object, and mArray instead of Array. For the Unicode map version, use wmValue, wmObject, and wmArray.

The following table shows the time in seconds it takes on my PC to extract the data from a single object of varying sizes. The methods used are as per the demo programs. The vector version is faster until the number of object members reach around 10, but then gets exponentially slower.

size vector map
2 2.03253e-007 3.24672e-007
5 7.63788e-007 9.59516e-007
10 2.26948e-006 2.00929e-006
15 4.68965e-006 3.28509e-006
20 8.19667e-006 4.75871e-006
30 1.72695e-005 7.64189e-006
50 4.63171e-005 1.39316e-005
75 0.000102418 2.27205e-005
100 0.000179923 3.17732e-005
200 0.000709335 7.3061e-005
500 0.00440368 0.00022076
1000 0.0175385 0.000495554
10000 1.76533 0.00658268
100000 178.718 0.127325
1000000 17800.718 1.99467

Note: with a vector, object members will be written out in the same order they were read in. A map will sort members alphabetically. A vector object also allows members to have duplicate names. This might be useful in some circumstances but would be non-standard.

Stream_reader Class

There is a Stream_reader class to work around a bug that prevents multiple top level values that are not separated by white space from being read from a stream. The standard stream reading functions would fail in this case.

istringstream is( "[1][1,2][1,2,3]" );  // no white space separating arrays

Stream_reader< istringstream, Value > reader( is );

Value value;

const bool ok = reader.read_next( value ); // read first array
reader.read_next( value );                 // read second array
reader.read_next( value );                 // read third array

I originally thought that this was caused by a bug in the boost.multi_pass iterator. However, I now realise that it basically results from the intrinsic behaviour of the std::istream_iterator constructor in immediately reading a character from its stream.

Compiling and linking

There are two ways to use the library:

  • Build and link in the JSON Spirit an object library.
  • Just include the relevant header files, i.e., "header only" use.

Linking to the object library has the advantage that the heavily templated code only has to be compiled once. Compiling the code can take some time. If you just include the header file implementation, the code will be compiled each time it is included. The use of precompiled headers should help, but I have not investigated this.

The include files needed to read JSON text using the object library are json_spirit_reader.h and json_spirit_value.h. To generate JSON text, you need json_spirit_writer.h and json_spirit_value.h. Alternately, you can include json_spirit.h, which includes all three of the above.

The include files needed when using the header only implementation are json_spirit_reader_template.h and json_spirit_writer_template.h.

All JSON Spirit declarations are in the namespace json_spirit.

Reducing Build Times

As already described, you can choose between four different types of JSON value classes: Value, mValue, wValue and wmValue. You would normally chose a single type and stick with it, however the compiler will process the templates to generate sets of classes for all four types. This results in longer than necessary build times and larger intermediate files. You can disable value types you don't want by commenting out the relevant lines in json_spirit_value.h below:

#define JSON_SPIRIT_VALUE_ENABLED
#define JSON_SPIRIT_WVALUE_ENABLED
#define JSON_SPIRIT_MVALUE_ENABLED
#define JSON_SPIRIT_WMVALUE_ENABLED

Using JSON Spirit with Multiple Threads

If you intend to use JSON Spirit in more than one thread, you will need to uncomment the following line near the top of json_spirit_reader.cpp.

//#define BOOST_SPIRIT_THREADSAFE

In this case, Boost Spirit will require you to link against Boost Threads.

Advanced

Container Constructor

An advanced feature introduced in version 4.05 is a constructor in the Value classes for containers. A container is passed in as an iterator range. This allows you to create an Array from a container such as an std::vector. For example, the following code creates and outputs the array "[1,2,3]":

std::vector< int > int_vect;

int_vect.push_back( 1 );
int_vect.push_back( 2 );
int_vect.push_back( 3 );

const Value val( int_vect.begin(), int_vect.end() );

cout << write( val );

Any compatible container type can be used. The following code creates and outputs the array "[1.0,2.0,3.0]":

std::list< float > float_list;

double_vect.push_back( 1.0 );
double_vect.push_back( 2.0 );
double_vect.push_back( 3.0 );

const Value val( double_vect.begin(), double_vect.end() );

cout << write( val, remove_trailing_zeros );

Variant Constructor

Another advanced feature introduced in version 4.05 is a constructor in the Value classes for boost::variant objects. Any compatible variant type can be used. For example:

boost::variant< int, string, std::vector< bool > > var;

var = 1;
Value val( var );
cout << write( val ) << endl;

var = string( "hello" );
val = var;
cout << write( val ) << endl;

std::vector< bool > bool_vect;
bool_vect.push_back( true );
bool_vect.push_back( false );
bool_vect.push_back( true );
var = bool_vect;
val = var;
cout << write( val ) << endl;

Output:

1
"hello"
[true,false,true]

History

  • Version 1.00, 10 August 2007
  • 12 August 2007
    • Part of the Boost Spirit Applications Repository
  • Version 1.01, 20 August 2007
    • Fixed bug outputting escape characters
  • Version 1.02, 9 October 2007
    • Fixed bug inputting "\/"
    • No longer attaches a semantic action c_ecscape_ch_p, simplifying the code
    • Speed optimizations
  • Version 2.00, 9 November 2007
    • Unicode support
    • Writes out "/" without a "\" escape character
  • Version 2.01, 14 December 2007
    • Increased precision of floating point number output
  • Version 2.02, 12 February 2008, bug fixes
    • Value construction from explicit const char*
    • Writes out extended ASCII
  • Version 2.03, 1 March 2008
    • Added support for 64-bit integers
  • Version 2.04, 22 April 2008.
    • Allows the reading of top level values other than Arrays and Objects, see Daniel Friederich's message thread below
    • Fixed bug where, e.g., "[ 1 2 ]" is read as "[ 1, 2 ]"
    • Added template getter function get_value()
  • Version 2.05, 2 June 2008
    • Linux version added
    • Added commented out #define BOOST_SPIRIT_THREADSAFE
  • Version 2.06, 12 September 2008
    • Now works when MSVC 2005 is using STLPort
  • Version 3.00, 31 January 2009
    • Multiple top-level objects can now be read from a string or stream
    • Addition of functions that take a pair of string iterators
    • Addition of functions that report the position of format errors
    • Speed optimisations
    • Added utility functions obj_to_map, map_to_obj, find_value
  • Version 3.01, 11 May 2009
    • Added CMake files, many thanks to Uwe Arzt for supplying these
    • json_spirit_writer_test.cpp now includes limits.h, needed on some platforms
  • Version 4.00, 9 June 2009
    • Added alternative std::map implementation for objects
    • Added support for unsigned 64 bit integers
    • Allows numbers such as 3 to be read as integers or floating point numbers
  • Version 4.01, 13 June 2009
    • Changed source code license to the MIT License
    • Stopped GCC warning about missing switch statement
  • Version 4.02, 11 September 2009
    • Fixed compilation errors in the unit tests that occurred with 64 bit Linux GCC
    • Changed the internals of the Value_impl class to use boost variant, giving a speed improvement
    • Small changes to prevent some pedantic warning messages
  • Version 4.03, 12 January 2010
    • Implementation moved to header files to allow header only use
    • Now compiles under MSVC 2010 beta
    • Added Stream_reader class to work around a bug in boost spirit::multi_pass
    • Added CPack declarations to the CMake file
    • A std:runtime_error is now thrown instead of an assert firing on an attempt to get a value of the wrong type
    • CMake minimum boost version is 1.34, not 1.37
  • Version 4.04, 8 January 2011
    • Added the raw_utf8 feature to the write functions
    • Added a << std::dec before writing to a stream to ensure that integer variables appear in base 10 notation
    • Added additional .h files to CMakeLists.txt
    • Output doubles using precision of 17 instead of 16
    • Added option to remove trailing zeros when outputting doubles
    • Ensures that write functions return the state of a given IO stream to its original state on completion
  • Version 4.05, 12 September 2011
    • Added a new output option for single line arrays
    • Fixed non-standard zero digits following decimal point with "remove trailing zeros" output option
    • Reduced build times with new JSON_SPIRIT_VALUE_ENABLED type #defines
    • Added a constructor for variant types
    • Added a constructor for container iterators
  • Version 4.06, 14 May 2013
    • Added support for Javascript type comments, i.e. // and /* */
    • Added json_spirit_writer_options.h to CMake install
    • Improved error message on attempt to extract the wrong type of data from a value
    • Correct bug that reduced write performance
    • Added always_escape_nonascii writer flag
  • Version 4.07, 14 April 2014
    • Trailing zeros are now removed by default when writing out doubles
    • Added the precision_of_doubles parameter to write functions
  • Version 4.08, 10 May 2014
    • Undone change made in v4.07 where the parameters to the writer functions were changed from being ints to enums. This prevented the options from being or'ed.

License

This article, along with any associated source code and files, is licensed under The MIT License

Share

About the Author

John W. Wilkinson
Software Developer (Senior) Spirent Communications Plc
United Kingdom United Kingdom
No Biography provided

Comments and Discussions

 
GeneralRe: Optimising write performance - drastically! PinmemberJonas Bjering8-May-13 21:39 
QuestionCan I use this class in Black Berry 10 development? PinmemberMatra Divat5-Feb-13 4:16 
AnswerRe: Can I use this class in Black Berry 10 development? PinmemberJohn W. Wilkinson10-Feb-13 9:23 
Generalgratz [modified] Pinmemberr0d_4-Jan-13 3:06 
QuestionWhich are the project that work as follows... PinmemberTarun Batra10-Oct-12 20:34 
AnswerRe: Which are the project that work as follows... PinmemberJohn W. Wilkinson12-Oct-12 8:44 
QuestionRead and Write - real numbers (with "double" type) Pinmemberz8z831-Jul-12 23:52 
AnswerRe: Read and Write - real numbers (with "double" type) PinmemberJohn W. Wilkinson1-Aug-12 10:02 
GeneralRe: Read and Write - real numbers (with "double" type) Pinmemberz8z82-Aug-12 3:37 
GeneralRe: Read and Write - real numbers (with "double" type) PinmemberJohn W. Wilkinson2-Aug-12 10:11 
QuestionHave you thought about how to handle nan's and infinity's? PinmemberStephen B Chisholm11-Jul-12 4:27 
AnswerRe: Have you thought about how to handle nan's and infinity's? PinmemberJohn W. Wilkinson12-Jul-12 10:39 
Questionhow to compile JSON Spirit with VS 2010 c++ Pinmemberminhvc30-May-12 17:17 
AnswerRe: how to compile JSON Spirit with VS 2010 c++ PinmemberJohn W. Wilkinson5-Jun-12 9:41 
You should be able to open the provided VS2005 "solution" file. VS2010 will convert it and the associated project files to 2010 format.
 
John
GeneralRe: how to compile JSON Spirit with VS 2010 c++ [modified] Pinmemberminhvc18-Jun-12 16:44 
BugGCC 4.1.2 patch Pinmemberalex404719-May-12 21:23 
GeneralRe: GCC 4.1.2 patch PinmemberJohn W. Wilkinson20-May-12 9:05 
Questionjson_spirit_writer_options.h not installed on Linux Pinmemberadricanedo16-May-12 5:38 
AnswerRe: json_spirit_writer_options.h not installed on Linux PinmemberJohn W. Wilkinson16-May-12 8:59 
QuestionBuilding source under the linux environment [modified] Pinmemberpei hwa yin12-Apr-12 7:23 
AnswerRe: Building source under the linux environment PinmemberJohn W. Wilkinson13-Apr-12 9:01 
GeneralRe: Building source under the linux environment Pinmemberpei hwa yin18-Apr-12 6:41 
GeneralMy vote of 5 PinmemberMember 876780529-Mar-12 5:36 
GeneralRe: My vote of 5 PinmemberJohn W. Wilkinson29-Mar-12 9:13 
GeneralRe: My vote of 5 PinmemberMember 876780530-Mar-12 3:32 
GeneralRe: My vote of 5 PinmemberJohn W. Wilkinson1-Apr-12 9:25 
QuestionRe: My vote of 5 Pinmemberpei hwa yin12-Apr-12 10:01 
AnswerRe: My vote of 5 PinmemberJohn W. Wilkinson13-Apr-12 8:43 
GeneralRe: My vote of 5 Pinmemberpei hwa yin18-Apr-12 6:42 
AnswerRe: My vote of 5 Pinmember_scottymac_22-Apr-12 9:22 
QuestionBetter error messages in exceptions PinmemberxivVerge14-Mar-12 21:29 
AnswerRe: Better error messages in exceptions PinmemberJohn W. Wilkinson15-Mar-12 10:13 
GeneralMy vote of 5 Pinmemberyork52812-Mar-12 17:03 
GeneralRe: My vote of 5 PinmemberJohn W. Wilkinson15-Mar-12 9:56 
GeneralRe: My vote of 5 Pinmemberyork52816-Mar-12 9:26 
Questionwrite_format question? Pinmemberyork5286-Mar-12 20:56 
AnswerRe: write_format question? PinmemberJohn W. Wilkinson7-Mar-12 6:26 
GeneralRe: write_format question? Pinmemberyork5287-Mar-12 16:03 
Questionread_or_throw or read, difference between read_or_throw and read PinmemberYuancai (Charlie) Ye29-Feb-12 6:19 
AnswerRe: read_or_throw or read, difference between read_or_throw and read PinmemberJohn W. Wilkinson29-Feb-12 9:39 
QuestionImprove JSON spirit by eliminating parser engine repeatedly PinmemberYuancai (Charlie) Ye29-Feb-12 6:16 
AnswerRe: Improve JSON spirit by eliminating parser engine repeatedly PinmemberJohn W. Wilkinson29-Feb-12 10:03 
QuestionAdd support to comments PinmemberYuancai (Charlie) Ye29-Feb-12 6:10 
AnswerRe: Add support to comments PinmemberJohn W. Wilkinson29-Feb-12 9:30 
QuestionIs there a source code repository? Pinmembereturkay1-Jan-12 23:30 
AnswerRe: Is there a source code repository? PinmemberJohn W. Wilkinson3-Jan-12 8:57 
QuestionUnordered map [modified] Pinmembertekert212-Dec-11 8:03 
AnswerRe: Unordered map PinmemberJohn W. Wilkinson3-Jan-12 9:25 
QuestionProblem with (de)serializing \uHHHH literal [modified] PinmemberTomasz Roda28-Nov-11 2:23 
QuestionStreaming read PinmemberMember 827004726-Sep-11 17:18 

General General    News News    Suggestion Suggestion    Question Question    Bug Bug    Answer Answer    Joke Joke    Rant Rant    Admin Admin   

Use Ctrl+Left/Right to switch messages, Ctrl+Up/Down to switch threads, Ctrl+Shift+Left/Right to switch pages.

| Advertise | Privacy | Mobile
Web03 | 2.8.140902.1 | Last Updated 10 May 2014
Article Copyright 2007 by John W. Wilkinson
Everything else Copyright © CodeProject, 1999-2014
Terms of Service
Layout: fixed | fluid