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, as described in the following table.
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"
serializer="org.jibx.runtime.Utility.serializeCharString"
deserializer="org.jibx.runtime.Utility.deserializeCharString"/>
|
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.util.Date |
Date.default |
Converts instances of java.util.Date to and from the schema
dateTime representation (a text representation like "2000-03-21T01:33:00", with
optional trailing fractional seconds, and difference from UTC). Since schema
doesn't have any concept equivalent to Java time zones, this conversion always
serializes times as UTC values (identified by a trailing "Z"). When
deserializing times which do not include a "Z" or offset from UTC it treats the
values as UTC. |
java.sql.Date |
SqlDate.default |
Converts instances of java.sql.Date to and from the schema
date representation (a text representation like
"2000-03-21"). |
java.sql.Time |
SqlTime.default |
Converts instances of java.sql.Time to and from the schema
time representation (a text representation like
"01:33:00" with optional trailing fractional seconds). |
java.sql.Timestamp |
Timestamp.default |
Converts instances of java.sql.Timestamp to and from the
schema dateTime representation, just as the Date:default conversion does
for java.util.Date instances. The only difference is that using the
timestamp value permits greater precision in the time value represented, down to
the nanosecond level. |
java.lang.String |
String.default |
Identity converter for java.lang.String instances. |
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 as well as some alternate date-related conversion
methods. The methods serializeDate
and deserializeDate
may be of special interest for developers working with xs:date values, since
they support conversion of xs:date format to and from
java.util.Date
instances (the equivalent to the default conversion
used for java.sql.Date
instances).
The default conversions listed above don't always match conversions used by
other frameworks for working with XML in Java. In particular, many other
frameworks convert the various schema date-related types to and from
java.util.Calendar
instances. This is an incorrect interpretation
of the schema types, which do not support the associated time zone which is a
necessary part of Calendar
instances. JiBX by default takes the
approach of converting to and from Java objects which are the nearest possible
equivalents to the schema types. You can always change these defaults for your
own purposes using the techniques described above.
In addition to the conversions listed in the table, JiBX 1.1 and later
support automatic conversion of Java 5 enum values to and from strings.