Value Conversions

JiBX defines a default set of formats used for converting property values to and from text strings. These format definitions are effectively defined in a context outside the root binding element of the binding definition. Most of the conversions are based on the W3C XML Schema datatype definitions, but do differ from the standard schema definitions in that leading and trailing whitespace characters are not stripped from values (meaning the default conversions will report errors for valid schema values in cases where leading or trailing whitespace is used). If this causes problems for your data, the best solution for now is to define custom conversions which strip the leading and trailing whitespace from values. The default formats for simple values are shown in the following table (also see the Date Time Conversions page, for the standard conversions relating to date/time values).

Default Formats

Type Format Label Conversion
byte byte.default

Converts primitive byte values to and from the schema byte representation (integer values in the range of -128 to +127).

char char.default

Converts primitive char values to and from the schema unsigned short representation (integer values in the range of 0 to 65535). Although this is the default for char primitives, JiBX also supports an alternate conversion as single-character text values using the named format char.string (allowing the conversion to be applied on a specific value) and a pair of conversion methods. To make the single-character text conversion the default in your binding, just specify the format as a child of the root binding element as follows:

<format type="char"
double double.default

Converts primitive double values to and from the schema double representation (IEEE double-precision floating point).

float float.default

Converts primitive float values to and from the schema float representation (IEEE single-precision floating point).

int int.default

Converts primitive int values to and from the schema int representation (integer values in normal 32-bit signed value range).

long long.default

Converts primitive long values to and from the schema long representation (integer values in normal 64-bit signed value range).

short short.default

Converts primitive short values to and from the schema short representation (integer values in normal 16-bit signed value range).

boolean boolean.default

Converts primitive boolean values to and from the schema boolean representation ("true" or "false", or equivalently "1" or "0" - the former values are always used when converting to text).

byte[] byte-array.default

Converts byte arrays to and from the schema base64 representation to allow arbitrary binary data.

java.lang.String String.default

Identity converter for java.lang.String instances.

org.jibx.runtime.QName QName.default

Converter for namespace-qualified names using org.jibx.runtime.QName.

java.lang.Boolean Boolean.default

Converter for java.lang.Boolean instances to and from the schema boolean representation. As with boolean.default, this always uses the values "true" and "false" when converting to text.

java.lang.Object Object.default

Converts any object to a string representation by using the toString() method, and converts a string to any object by using a constructor that takes a single argument of type java.lang.String. If an optional value is not present when unmarshalling a null value is stored to the object reference. This is the default format conversion used for any object type without a more specific conversion defined.

These are the conversions you'll get unless you specify otherwise. If you use a format element to override one of these defaults you'll still be able to access the default by using the defined label for that format. Defining your own custom serialization formats is easy, basically requiring only a pair of static methods to convert to and from String representations. See the format element description for details of defining your own conversions. The org.jibx.runtime.Utility class defines a variety of conversion methods, including those used to implement the standard conversions listed in the above table, so it's a good starting point for developing your own variations.

java.math.BigDecimal is a special case. For 1.3 and 1.4.1 JDKs the default java.lang.Object conversion from the above table worked fine. In the 1.5 JDK Sun made a non-backward-compatible change to the format of the BigDecimal.toString() output, which meant that the resulting text does not always match the decimal schema representation (due to the use of exponents). Because of this JDK compatibility failure JiBX has an added optional conversion for BigDecimal which uses a method added in 1.5 JDK to obtain decimal-compatible output:

Non-default Formats

Type Format Label Conversion
java.math.BigDecimal BigDecimal.java5

Converts BigDecimal values to and from the schema decimal representation (string of digits with optional leading sign and optional embedded decimal point). This conversion can only be used for 1.5 and later JDKs.

In addition to the conversions listed in the tables, JiBX 1.1 and later support automatic conversion of Java 5 enum values to and from strings.