The Bean Introspection Utilities component of the Jakarta Commons subproject offers low-level utility classes that assist in getting and setting property values on Java classes that follow the naming design patterns outlined in the JavaBeans Specification, as well as mechanisms for dynamically defining and accessing bean properties.
The JavaBeans name comes from a Java API specification for a component architecture for the Java language. Writing Java classes that conform to the JavaBeans design patterns makes it easier for Java developers to understand the functionality provided by your class, as well as allowing JavaBeans-aware tools to use Java's introspection capabilities to learn about the properties and operations provided by your class, and present them in a visually appealing manner in development tools.
The JavaBeans Specification describes the complete set of characteristics that makes an arbitrary Java class a JavaBean or not -- and you should consider reading this document to be an important part of developing your Java programming skills. However, the required characteristics of JavaBeans that are important for most development scenarios are listed here:
String className = ...;
Class beanClass = Class.forName(className);
Object beanInstance = beanClass.newInstance();
get or set as the
prefix for the property name with it's first character capitalized. Thus,
you a JavaBean representing an employee might have
(among others) properties named firstName,
lastName, and hireDate, with method signatures
like this:
public class Employee {
public Employee(); // Zero-arguments constructor
public String getFirstName();
public void setFirstName(String firstName);
public String getLastName();
public void setLastName(String lastName);
public Date getHireDate();
public void setHireDate(Date hireDate);
public boolean isManager();
public void setManager(boolean manager);
public String getFullName();
}
is prefix instead of a get prefix if that makes
for a more understandable method name.fullName property is read-only, because there is no
setter method. It is also possible, but less common, to provide
write-only properties.BeanInfo class associated with
your bean class. See the JavaBeans Specification for full details.Using standard Java coding techniques, it is very easy to deal with JavaBeans if you know ahead of time which bean classes you will be using, and which properties you are interested in:
Employee employee = ...;
System.out.println("Hello " + employee.getFirstName() + "!");
The commons-beanutils package requires that the following additional packages be available in the application's class path at runtime:
As described above, the standard facilities of the Java programming language
make it easy and natural to access the property values of your beans using
calls to the appropriate getter methods.
But what happens in more sophisticated environments where you do not
necessarily know ahead of time which bean class you are going to be using,
or which property you want to retrieve or modify? The Java language provides
classes like java.beans.Introspector, which can examine a Java
class at runtime and identify for you the names of the property getter and
setter methods, plus the Reflection capabilities to dynamically call
such a method. However, these APIs can be difficult to use, and expose the
application developer to many unnecessary details of the underlying structure
of Java classes. The APIs in the BeanUtils package are intended to simplify
getting and setting bean properties dynamically, where the objects you are
accessing -- and the names of the properties you care about -- are determined
at runtime in your application, rather than as you are writing and compiling
your application's classes.
This is the set of needs that are satisfied by the static methods of the PropertyUtils class, which are described further in this section. First, however, some further definitions will prove to be useful:
The general set of possible property types supported by a JavaBean can be broken into three categories -- some of which are supported by the standard JavaBeans specification, and some of which are uniquely supported by the BeanUtils package:
int, a simple
object (such as a java.lang.String), or a more complex
object whose class is defined either by the Java language, by the
application, or by a class library included with the application.java.util.List (or an implementation of List) to be
indexed as well.java.util.Map to be "mapped". You can set and
retrieve individual values via a String-valued key.A variety of API methods are provided in the PropertyUtils class to get and set property values of all of these types. In the code fragments below, assume that there are two bean classes defined with the following method signatures:
public class Employee {
public Address getAddress(String type);
public void setAddress(String type, Address address);
public Employee getSubordinate(int index);
public void setSubordinate(int index, Employee subordinate);
public String getFirstName();
public void setFirstName(String firstName);
public String getLastName();
public void setLastName(String lastName);
}
Getting and setting simple property values is, well, simple :-). Check out the following API signatures in the Javadocs:
Using these methods, you might dynamically manipulate the employee's name in an application:
Employee employee = ...;
String firstName = (String)
PropertyUtils.getSimpleProperty(employee, "firstName");
String lastName = (String)
PropertyUtils.getSimpleProperty(employee, "lastName");
... manipulate the values ...
PropertyUtils.setSimpleProperty(employee, "firstName", firstName);
PropertyUtils.setSimpleProperty(employee, "lastName", lastName);
For indexed properties, you have two choices - you can either build a subscript into the "property name" string, using square brackets, or you can specify the subscript in a separate argument to the method call:
Only integer constants are allowed when you add a subscript to the property name. If you need to calculate the index of the entry you wish to retrieve, you can use String concatenation to assemble the property name expression. For example, you might do either of the following:
Employee employee = ...;
int index = ...;
String name = "subordinate[" + index + "]";
Employee subordinate = (Employee)
PropertyUtils.getIndexedProperty(employee, name);
Employee employee = ...;
int index = ...;
Employee subordinate = (Employee)
PropertyUtils.getIndexedProperty(employee, "subordinate", index);
In a similar manner, there are two possible method signatures for getting and setting mapped properties. The difference is that the extra argument is surrounded by parentheses ("(" and ")") instead of square brackets, and it is considered to be a String-value key used to get or set the appropriate value from an underlying map.
You can, for example, set the employee's home address in either of these two manners:
Employee employee = ...;
Address address = ...;
PropertyUtils.setMappedProperty(employee, "address(home)", address);
Employee employee = ...;
Address address = ...;
PropertyUtils.setMappedProperty(employee, "address", "home", address);
In all of the examples above, we have assumed that you wished to retrieve the value of a property of the bean being passed as the first argument to a PropertyUtils method. However, what if the property value you retrieve is really a Java object, and you wish to retrieve a property of that object instead?
For example, assume we really wanted the city property of the
employee's home address. Using standard Java programming techniques for direct
access to the bean properties, we might write:
String city = employee.getAddress("home").getCity();
The equivalent mechanism using the PropertyUtils class is called nested property access. To use this approach, you concatenate together the property names of the access path, using "." separators -- very similar to the way you can perform nested property access in JavaScript.
The PropertyUtils equivalent to the above Java expression would be:
String city = (String)
PropertyUtils.getNestedProperty(employee, "address(home).city");
Finally, for convenience, PropertyUtils provides method signatures that accept any arbitrary combination of simple, indexed, and mapped property access, using any arbitrary level of nesting:
which you might use like this:
Employee employee = ...;
String city = (String) PropertyUtils.getProperty(employee,
"subordinate[3].address(home).city");
The PropertyUtils class described in the preceding section is designed to provide dynamic property access on existing JavaBean classes, without modifying them in any way. A different use case for dynamic property access is when you wish to represent a dynamically calculated set of property values as a JavaBean, but without having to actually write a Java class to represent these properties. Besides the effort savings in not having to create and maintain a separate Java class, this ability also means you can deal with situations where the set of properties you care about is determined dynamically (think of representing the result set of an SQL select as a set of JavaBeans ...).
To support this use case, the BeanUtils package provides the
DynaBean interface, which must be implemented by a
bean class actually implementing the interface's methods, and the associated
DynaClass interface that defines the set of
properties supported by a particular group of DynaBeans, in much the same way
that java.lang.Class defines the set of properties supported by
all instances of a particular JavaBean class.
For example, the Employee class used in the examples above
might be implemented as a DynaBean, rather than as a standard JavaBean. You
can access its properties like this:
DynaBean employee = ...; // Details depend on which
// DynaBean implementation you use
String firstName = (String) employee.get("firstName");
Address homeAddress = (Address) employee.get("address", "home");
Object subordinate = employee.get("subordinate", 2);
One very important convenience feature should be noted: the
PropertyUtils property getter and setter methods understand how to access
properties in DynaBeans. Therefore, if the bean you pass as the first
argument to, say, PropertyUtils.getSimpleProperty() is really a
DynaBean implementation, the call will get converted to the appropriate
DynaBean getter method transparently. Thus, you can base your application's
dynamic property access totally on the PropertyUtils APIs, if you wish, and
use them to access either standard JavaBeans or DynaBeans without having to
care ahead of time how a particular bean is implemented.
Because DynaBean and DynaClass are interfaces, they may be implemented multiple times, in different ways, to address different usage scenarios. The following subsections describe the implementations that are provided as a part of the standard BeanUtils package, although you are encouraged to provide your own custom implementations for cases where the standard implementations are not sufficient.
BasicDynaBean and BasicDynaClassThe BasicDynaBean and BasicDynaClass implementation provides a basic set of dynamic property capabilities where you want to dynamically define the set of properties (described by instances of DynaProperty). You start by defining the DynaClass that establishes the set of properties you care about:
DynaProperty[] props = new DynaProperty[]{
new DynaProperty("address", java.util.Map.class),
new DynaProperty("subordinate", mypackage.Employee[].class),
new DynaProperty("firstName", String.class),
new DynaProperty("lastName", String.class)
};
BasicDynaClass dynaClass = new BasicDynaClass("employee", null, props);
Note that the 'dynaBeanClass' argument (in the constructor of
BasicDynaClass) can have the value of null. In this
case, the value of dynaClass.getDynaBeanClass will just be the
Class for BasicDynaBean.
Next, you use the newInstance() method of this DynaClass to
create new DynaBean instances that conform to this DynaClass, and populate
its initial property values (much as you would instantiate a new standard
JavaBean and then call its property setters):
DynaBean employee = dynaClass.newInstance();
employee.set("address", new HashMap());
employee.set("subordinate", new mypackage.Employee[0]);
employee.set("firstName", "Fred");
employee.set("lastName", "Flintstone");
Note that the DynaBean class was declared to be
DynaBean instead of BasicDynaBean. In
general, if you are using DynaBeans, you will not want to care about the
actual implementation class that is being used -- you only care about
declaring that it is a DynaBean so that you can use the
DynaBean APIs.
As stated above, you can pass a DynaBean instance as the first argument
to a PropertyUtils method that gets and sets properties, and it
will be interpreted as you expect -- the dynamic properties of the DynaBean
will be retrieved or modified, instead of underlying properties on the
actual BasicDynaBean implementation class.
ResultSetDynaClass (Wraps ResultSet in DynaBeans)A very common use case for DynaBean APIs is to wrap other collections of
"stuff" that do not normally present themselves as JavaBeans. One of the most
common collections that would be nice to wrap is the
java.sql.ResultSet that is returned when you ask a JDBC driver
to perform a SQL SELECT statement. Commons BeanUtils offers a standard
mechanism for making each row of the result set visible as a DynaBean,
which you can utilize as shown in this example:
Connection conn = ...;
Statement stmt = conn.createStatement();
ResultSet rs = stmt.executeQuery
("select account_id, name from customers");
Iterator rows = (new ResultSetDynaClass(rs)).iterator();
while (rows.hasNext()) {
DynaBean row = (DynaBean) rows.next();
System.out.println("Account number is " +
row.get("account_id") +
" and name is " + row.get("name"));
}
rs.close();
stmt.close();
RowSetDynaClass (Disconnected ResultSet as DynaBeans)Although ResultSetDynaClass is
a very useful technique for representing the results of an SQL query as a
series of DynaBeans, an important problem is that the underlying
ResultSet must remain open throughout the period of time that the
rows are being processed by your application. This hinders the ability to use
ResultSetDynaClass as a means of communicating information from
the model layer to the view layer in a model-view-controller architecture
such as that provided by the Struts
Framework, because there is no easy mechanism to assure that the result set
is finally closed (and the underlying Connection returned to its
connection pool, if you are using one).
The RowSetDynaClass class represents a different approach to
this problem. When you construct such an instance, the underlying data is
copied into a set of in-memory DynaBeans that represent the result.
The advantage of this technique, of course, is that you can immediately close
the ResultSet (and the corresponding Statement), normally before you even
process the actual data that was returned. The disadvantage, of course, is
that you must pay the performance and memory costs of copying the result data,
and the result data must fit entirely into available heap memory. For many
environments (particularly in web applications), this tradeoff is usually
quite beneficial.
As an additional benefit, the RowSetDynaClass class is defined
to implement java.io.Serializable, so that it (and the
DynaBeans that correspond to each row of the result) can be conveniently
serialized and deserialized (as long as the underlying column values are
also Serializable). Thus, RowSetDynaClass represents a very
convenient way to transmit the results of an SQL query to a remote Java-based
client application (such as an applet).
The normal usage pattern for a RowSetDynaClass will look
something like this:
Connection conn = ...; // Acquire connection from pool
Statement stmt = conn.createStatement();
ResultSet rs = stmt.executeQuery("SELECT ...");
RowSetDynaClass rsdc = new RowSetDynaClass(rs);
rs.close();
stmt.close();
...; // Return connection to pool
List rows = rsdc.getRows();
...; // Process the rows as desired
WrapDynaBean and WrapDynaClassOK, you've tried the DynaBeans APIs and they are cool -- very simple
get() and set() methods provide easy access to all
of the dynamically defined simple, indexed, and mapped properties of your
DynaBeans. You'd like to use the DynaBean APIs to access all
of your beans, but you've got a bunch of existing standard JavaBeans classes
to deal with as well. This is where the
WrapDynaBean (and its associated
WrapDynaClass) come into play. As the name
implies, a WrapDynaBean is used to "wrap" the DynaBean APIs around an
existing standard JavaBean class. To use it, simply create the wrapper
like this:
MyBean bean = ...;
DynaBean wrapper = new WrapDynaBean(bean);
String firstName = wrapper.get("firstName");
Note that, although appropriate WrapDynaClass instances are
created internally, you never need to deal with them.
You bought into the DynaBeans because it saves coding all those POJO JavaBeans but you're here because lazy caught your eye and wondered whats that about? What makes these flavors of DynaBean lazy are the following features:
DynaClass which implements
the MutableDynaClass
interface. This provides the ability to add and remove a DynaClass's
properties. Lazy beans use this feature to automatically add
a property which doesn't exist to the DynaClass when
the set(name, value) method is called.index being set then the List or
Array is automatically grown so that it is.get(name, index) or
set(name, index, value)) results in either a new List
or Array being instantiated. If the indexed property has not been
defined in the DynaClass then it is automatically added and a default List
implementation instantiated.get(name, key) or
set(name, key, value)) results in a new Map
being instantiated. If the mapped property has not been defined in the DynaClass
then it is automatically added and a default Map implementation
instantiated.DynaClass as a DynaBean or regular bean and
doesn't exist in the DynaBean then LazyDynaBean wiill
try to instantiate the bean using a default empty constructor.LazyDynaBean is the standard lazy bean
implementation. By default it is associated with a LazyDynaClass
which implements the MutableDynaClass interface - however
it can be used with any MutableDynaClass implementation. The question is how do
I use it? - well it can be as simple as creating a new bean and then calling the getters/setters...
DynaBean dynaBean = new LazyDynaBean();
dynaBean.set("foo", "bar"); // simple
dynaBean.set("customer", "title", "Mr"); // mapped
dynaBean.set("customer", "surname", "Smith"); // mapped
dynaBean.set("address", 0, addressLine1); // indexed
dynaBean.set("address", 1, addressLine2); // indexed
dynaBean.set("address", 2, addressLine3); // indexed
LazyDynaMap is a light wieght DynaBean
facade to a Map with all the usual lazy features. Its light weight because it doesn't
have an associated DynaClass containing all the properties. In fact it actually implements
the DynaClass interface itself (and MutableDynaClass) and derives all the DynaClass
information from the actual contents of the Map. A LazyDynaMap can be created around an
existing Map or can instantiate its own Map. After any DynaBean
processing has finished the Map can be retrieved and the DynaBean facade discarded.
If you need a new Map then to use....
DynaBean dynaBean = new LazyDynaMap(); // create DynaBean
dynaBean.set("foo", "bar"); // simple
dynaBean.set("customer", "title", "Mr"); // mapped
dynaBean.set("address", 0, addressLine1); // indexed
Map myMap = dynaBean.getMap() // retrieve the Map
or to use with an existing Map ....
Map myMap = .... // exisitng Map
DynaBean dynaBean = new LazyDynaMap(myMap); // wrap Map in DynaBean
dynaBean.set("foo", "bar"); // set properties
LazyDynaClass extends BasicDynaClass
and implements the MutableDynaClass interface. It can be used with other
DynaBean implementations, but it is the default DynaClass used by LazyDynaBean.
When using the LazyDynaBean there may be no need to have anything to do with the DynaClass
However sometimes there is a requirement to set up the DynaClass first - perhaps to
define the type of array for an indexed property, or if using the DynaBean in restricted mode (see note below)
is required. Doing so is straight forward...
Either create a LazyDynaClass first...
MutableDynaClass dynaClass = new LazyDynaClass(); // create DynaClass
dynaClass.add("amount", java.lang.Integer.class); // add property
dynaClass.add("orders", OrderBean[].class); // add indexed property
dynaClass.add("orders", java.util.TreeMapp.class); // add mapped property
DynaBean dynaBean = new LazyDynaBean(dynaClass); // Create DynaBean with associated DynaClass
or create a LazyDynaBean and get the DynaClass...
DynaBean dynaBean = new LazyDynaBean(); // Create LazyDynaBean
MutableDynaClass dynaClass =
(MutableDynaClass)dynaBean.getDynaClass(); // get DynaClass
dynaClass.add("amount", java.lang.Integer.class); // add property
dynaClass.add("myBeans", myPackage.MyBean[].class); // add 'array' indexed property
dynaClass.add("myMap", java.util.TreeMapp.class); // add mapped property
NOTE: One feature of MutableDynaClass is that it
has a Restricted property. When the DynaClass is restricted no properties can be added
or removed from the DynaClass. Neither the LazyDynaBean or LazyDynaMap
will add properties automatically if the DynaClass is restricted.
So far, we've only considered the cases where the data types of the dynamically accessed properties are known, and where we can use Java casts to perform type conversions. What happens if you want to automatically perform type conversions when casting is not possible? The BeanUtils package provides a variety of APIs and design patterns for performing this task as well.
BeanUtils and ConvertUtils ConversionsA very common use case (and the situation that caused the initial creation
of the BeanUtils package) was the desire to convert the set of request
parameters that were included in a
javax.servlet.HttpServletRequest received by a web application
into a set of corresponding property setter calls on an arbitrary JavaBean.
(This is one of the fundamental services provided by the
Struts Framework, which uses
BeanUtils internally to implement this functionality.)
In an HTTP request, the set of included parameters is made available as a
series of String (or String array, if there is more than one value for the
same parameter name) instances, which need to be converted to the underlying
data type. The BeanUtils class provides
property setter methods that accept String values, and automatically convert
them to appropriate property types for Java primitives (such as
int or boolean), and property getter methods that
perform the reverse conversion. Finally, a populate() method
is provided that accepts a java.util.Map containing a set of
property values (keyed by property name), and calls all of the appropriate
setters whenever the underlying bean has a property with the same name as
one of the request parameters. So, you can perform the all-in-one property
setting operation like this:
HttpServletRequest request = ...;
MyBean bean = ...;
HashMap map = new HashMap();
Enumeration names = request.getParameterNames();
while (names.hasMoreElements()) {
String name = (String) names.nextElement();
map.put(name, request.getParameterValues(name));
}
BeanUtils.populate(bean, map);
The BeanUtils class relies on conversion methods defined in
the ConvertUtils class to perform the actual
conversions, and these methods are availablve for direct use as well.
WARNING - It is likely that the hard coded use of
ConvertUtils methods will be deprecated in the future, and
replaced with a mechanism that allows you to plug in your own implementations
of the Converter interface instead. Therefore,
new code should not be written with reliance on ConvertUtils.
The ConvertUtils class supports the ability to define and
register your own String --> Object conversions for any given Java class.
Once registered, such converters will be used transparently by all of the
BeanUtils methods (including populate()). To
create and register your own converter, follow these steps:
convert() method should accept the
java.lang.Class object of your application class (i.e.
the class that you want to convert to, and a String representing the
incoming value to be converted.ConvertUtils.register() method.The standard classes in org.apache.commons.beanutils are not
locale aware. This gives them a cleaner interface and makes then easier to use
in situations where the locale is not important.
Extended, locale-aware analogues can be found in
org.apache.commons.beanutils.locale
. These are built along the same
lines as the basic classes but support localization.
So far, the examples have covered the static utility classes (BeanUtils,
ConvertUtils and PropertyUtils). These are easy to use but are
somewhat inflexible. These all share the same registered converters and the same caches.
This functionality can also be accessed through utility objects (in fact, the static utility class use worker instances of these classes). For each static utility class, there is a corresponding class with the same functionality that can be instantiated:
| Static Utility Class | Utility Object |
|---|---|
| BeanUtils | BeanUtilsBean |
| ConvertUtils | ConvertUtilsBean |
| PropertyUtils | PropertyUtilsBean |
Creating an instances allow gives guarenteed control of the caching and registration to the code that creates it.
commons-beanutils-bean-collections.jar is available
(in addition to commons-beanutils-core.jar, of course :-).
org.apache.commons.beanutils.BeanComparator is a Comparator implementation
that compares beans based on a shared property value.
The Closure interface in commons-collections encapsulates a block of code that
executes on an arbitrary input Object. Commons-collections contains code that allows
Closures to be applied to the contents of a Collection. For more details, see the
commons-collections
documentation.
BeanPropertyValueChangeClosure is a Closure that sets a specified property
to a particular value. A typical usage is to combine this with commons-collections
so that all the beans in a collection can have a particular property set to a particular value.
For example, set the activeEmployee property to TRUE for an entire collection:
// create the closure
BeanPropertyValueChangeClosure closure =
new BeanPropertyValueChangeClosure( "activeEmployee", Boolean.TRUE );
// update the Collection
CollectionUtils.forAllDo( peopleCollection, closure );
The Predicate interface in commons-collections encapsulates an evaluation
of an input Object that returns either true or false. Commons-collections contains code
that allows
Predicates to be applied to be used to filter collections. For more details, see the
commons-collections
documentation.
BeanPropertyValueEqualsPredicate is a Predicate that evaluates a
set property value against a given value. A typical usage is
(in combination with commons-collections)
to filter collections on the basis of a property value.
For example, to filter a collection to find all beans where active employee is false use:
BeanPropertyValueEqualsPredicate predicate =
new BeanPropertyValueEqualsPredicate( "activeEmployee", Boolean.FALSE );
// filter the Collection
CollectionUtils.filter( peopleCollection, predicate );
The Transformer interface in commons-collections encapsulates the transformation
of an input Object into an output object. Commons-collections contains code
that allows
Transformers to be applied produce a collection of outputs from a collection of inputs.
For more details, see the
commons-collections
documentation.
BeanToPropertyTransformer is a Transformer implementation
that transforms a bean into it's property value.
For example, to find all cities that are contained in the address of each person property of each bean in
a collection:
// create the transformer
BeanToPropertyValueTransformer transformer = new BeanToPropertyValueTransformer( "person.address.city" );
// transform the Collection
Collection peoplesCities = CollectionUtils.collect( peopleCollection, transformer );
The BeanUtils package relies on introspection rather than reflection. This means that it will find only JavaBean compliant properties.
There are some subtleties of this specification that can catch out the unwary:
java.beans.Introspector searches widely for a custom BeanInfo
class. If your class has the same name as another with a custom BeanUtils
(typically a java API class) then the Introspector may use that instead of
creating via reflection based on your class. If this happens, the only solution is to
create your own BeanInfo.
BeanComparator relies on an internal Comparator to perform the actual
comparisions. By default,
org.apache.commons.collections.comparators.ComparableComparator
is used which imposes a natural order. If you want to change the order,
then a custom Comparator should be created and passed into the
appropriate constructor.
For example:
import org.apache.commons.collections.comparators.ComparableComparator;
import org.apache.commons.collections.comparators.ReverseComparator;
import org.apache.commons.beanutils.BeanComparator;
...
BeanComparator reversedNaturalOrderBeanComparator
= new BeanComparator("propertyName", new ReverseComparator(new ComparableComparator()));
Collections.sort(myList, reversedNaturalOrderBeanComparator);
...
BeanUtils now comes packaged (into jars) in two different ways:
commons-beanutils.jar)commons-beanutils-core.jar (core classes)commons-beanutils-bean-collections.jar
(additional dependency on commons-collections 3.0)
Those who want it all should grab the all-in-one commons-beanutils.jar
(and remember to add any optional dependencies needed)
whereas those who need minimal dependencies should use commons-beanutils-core.jar
plus any optional jars they plan to use.
The binary distribution contains all these jars.