The types of literal expressions supported are strings, dates, numeric values (int, real, and hex), boolean and null. String are delimited by single quotes. To put a single quote itself in a string use the backslash character. The following listing shows simple usage of literals. Typically they would not be used in isolation like this, but as part of a more complex expression, for example using a literal on one side of a logical comparison operator.

string helloWorld = (string) ExpressionEvaluator.GetValue(null, "'Hello World'"); // evals to "Hello World"

string tonyPizza  = (string) ExpressionEvaluator.GetValue(null, "'Tony\\'s Pizza'"); // evals to "Tony's Pizza"

double avogadrosNumber = (double) ExpressionEvaluator.GetValue(null, "6.0221415E+23");

int maxValue = (int)  ExpressionEvaluator.GetValue(null, "0x7FFFFFFF");  // evals to 2147483647

DateTime birthday = (DateTime) ExpressionEvaluator.GetValue(null, "date('1974/08/24')");

DateTime exactBirthday =
  (DateTime) ExpressionEvaluator.GetValue(null, " date('19740824T131030', 'yyyyMMddTHHmmss')");

bool trueValue = (bool) ExpressionEvaluator.GetValue(null, "true");

object nullValue = ExpressionEvaluator.GetValue(null, "null");

Note that the extra backslash character in Tony's Pizza is to satisfy C# escape syntax. Numbers support the use of the negative sign, exponential notation, and decimal points. By default real numbers are parsed using Double.Parse unless the format character "M" or "F" is supplied, in which case Decimal.Parse and Single.Parse would be used respectfully. As shown above, if two arguments are given to the date literal then DateTime.ParseExact will be used. Note that all parse methods of classes that are used internally reference the CultureInfo.InvariantCulture.

As shown in the previous example in Section 11.2, “Evaluating Expressions”, navigating through properties is easy, just use a period to indicate a nested property value. The instances of Inventor class, pupin and tesla, were populated with data listed in section Section 11.4, “Classes used in the examples”. To navigate "down" and get Tesla's year of birth and Pupin's city of birth the following expressions are used

int year = (int) ExpressionEvaluator.GetValue(tesla, "DOB.Year"));  // 1856

string city = (string) ExpressionEvaluator.GetValue(pupin, "PlaCeOfBirTh.CiTy");  // "Idvor"

For the sharp-eyed, that isn't a typo in the property name for place of birth. The expression uses mixed cases to demonstrate that the evaluation is case insensitive.

The contents of arrays and lists are obtained using square bracket notation.

// Inventions Array
string invention = (string) ExpressionEvaluator.GetValue(tesla, "Inventions[3]"); // "Induction motor"

// Members List
string name = (string) ExpressionEvaluator.GetValue(ieee, "Members[0].Name"); // "Nikola Tesla"

// List and Array navigation
string invention = (string) ExpressionEvaluator.GetValue(ieee, "Members[0].Inventions[6]") // "Wireless communication"

The contents of dictionaries are obtained by specifying the literal key value within the brackets. In this case, because keys for the Officers dictionary are strings, we can specify string literal.

// Officer's Dictionary
Inventor pupin = (Inventor) ExpressionEvaluator.GetValue(ieee, "Officers['president']";

string city = (string) ExpressionEvaluator.GetValue(ieee, "Officers['president'].PlaceOfBirth.City"); // "Idvor"

ExpressionEvaluator.SetValue(ieee, "Officers['advisors'][0].PlaceOfBirth.Country", "Croatia");

You may also specify non literal values in place of the quoted literal values by using another expression inside the square brackets such as variable names or static properties/methods on other types. These features are discussed in other sections.

Indexers are similarly referenced using square brackets. The following is a small example that shows the use of indexers. Multidimensional indexers are also supported.

public class Bar
{
    private int[] numbers = new int[] {1, 2, 3};

    public int this[int index]
    {
        get { return numbers[index];}
        set { numbers[index] = value; }
    }
}

Bar b = new Bar();

int val = (int) ExpressionEvaluator.GetValue(bar, "[1]") // evaluated to 2

ExpressionEvaluator.SetValue(bar, "[1]", 3);  // set value to 3

{1, 2, 3, 4, 5}
{'abc', 'xyz'}

new int[] {1, 2, 3, 4, 5}
new string[] {'abc', 'xyz'}

#{'key1' : 'Value 1', 'today' : DateTime.Today}
#{1 : 'January', 2 : 'February', 3 : 'March', ...}

Methods are invoked using typical C# programming syntax. You may also invoke methods on literals.

//string literal
char[] chars = (char[]) ExpressionEvaluator.GetValue(null, "'test'.ToCharArray(1, 2)"))  // 't','e'

//date literal
int year = (int) ExpressionEvaluator.GetValue(null, "date('1974/08/24').AddYears(31).Year") // 2005

// object usage, calculate age of tesla navigating from the IEEE society.

ExpressionEvaluator.GetValue(ieee, "Members[0].GetAge(date('2005-01-01')") // 149 (eww..a big anniversary is coming up ;)

Setting of a property is done by using the assignment operator. This would typically be done within a call to GetValue since in the simple case SetValue offers the same functionality. Assignment in this manner is useful when combining multiple operators in an expression list, discussed in the next section. Some examples of assignment are shown below

Inventor inventor = new Inventor();
String aleks = (String) ExpressionEvaluator.GetValue(inventor, "Name = 'Aleksandar Seovic'");
DateTime dt = (DateTime) ExpressionEvaluator.GetValue(inventor, "DOB = date('1974-08-24')");

//Set the vice president of the society
Inventor tesla = (Inventor) ExpressionEvaluator.GetValue(ieee, "Officers['vp'] = Members[0]");

Multiple expressions can be evaluated against the same context object by separating them with a semicolon and enclosing the entire expression within parentheses. The value returned is the value of the last expression in the list. Examples of this are shown below

//Perform property assignments and then return Name property.

String pupin = (String) ExpressionEvaluator.GetValue(ieee.Members,
  "( [1].PlaceOfBirth.City = 'Beograd'; [1].PlaceOfBirth.Country = 'Serbia'; [1].Name )"));

// pupin = "Mihajlo Pupin"

In many cases, you can reference types by simply specifying type name:

ExpressionEvaluator.GetValue(null, "1 is int")

ExpressionEvaluator.GetValue(null, "DateTime.Today")

ExpressionEvaluator.GetValue(null, "new string[] {'abc', 'efg'}")

This is possible for all standard types from mscorlib, as well as for any other type that is registered with the TypeRegistry as described in the next section.

For all other types, you need to use special T(typeName) expression:

Type dateType = (Type) ExpressionEvaluator.GetValue(null, "T(System.DateTime)")

Type evalType = (Type) ExpressionEvaluator.GetValue(null, "T(Spring.Expressions.ExpressionEvaluator, Spring.Core)")

bool trueValue = (bool) ExpressionEvaluator.GetValue(tesla, "T(System.DateTime) == DOB.GetType()")

	Note
	The implementation delegates to Spring's ObjectUtils.ResolveType method for the actual type resolution, which means that the types used within expressions are resolved in the exactly the same way as the types specified in Spring configuration files.

To refer to a type within an expression that is not in the mscorlib you need to register it with the TypeRegistry. This will allow you to refer to a shorthand name of the type within your expressions. This is commonly used in expression that use the new operator or refer to a static properties of an object. Example usage is shown below.

TypeRegistry.RegisterType("Society", typeof(Society));

Inventor pupin = (Inventor) ExpressionEvaluator.GetValue(ieee, "Officers[Society.President]");

Alternatively, you can register types using typeAliases configuration section.

Constructors can be invoked using the new operator. For classes outside mscorlib you will need to register your types so they can be resolved. Examples of using constructors are shown below:

// simple ctor
DateTime dt = (DateTime) ExpressionEvaluator.GetValue(null, "new DateTime(1974, 8, 24)");

// Register Inventor type then create new inventor instance within Add method inside an expression list.
// Then return the new count of the Members collection.

TypeRegistry.RegisterType(typeof(Inventor));
int three = (int) ExpressionEvaluator.GetValue(ieee.Members, "{ Add(new Inventor('Aleksandar Seovic', date('1974-08-24'), 'Serbian')); Count}"));

As a convenience, Spring.NET also allows you to define named constructor arguments, which are used to set object's properties after instantiation, similar to the way standard .NET attributes work. For example, you could create an instance of the Inventor class and set its Inventions property in a single statement:

Inventor aleks = (Inventor) ExpressionEvaluator.GetValue(null, "new Inventor('Aleksandar Seovic', date('1974-08-24'), 'Serbian', Inventions = {'SPELL'})");

The only rule you have to follow is that named arguments should be specified after standard constructor arguments, just like in the .NET attributes.

While we are on the subject, Spring.NET Expression Language also provides a convenient syntax for .NET attribute instance creation. Instead of using standard constructor syntax, you can use a somewhat shorter and more familiar syntax to create an instance of a .NET attribute class:

WebMethodAttribute webMethod = (WebMethodAttribute) ExpressionEvaluator.GetValue(null, "@[WebMethod(true, CacheDuration = 60, Description = 'My Web Method')]");

As you can see, with the exception of the @ prefix, syntax is exactly the same as in C#.

Slightly different syntax is not the only thing that differentiates an attribute expression from a standard constructor invocation expression. In addition to that, attribute expression uses slightly different type resolution mechanism and will attempt to load both the specified type name and the specified type name with an Attribute suffix, just like the C# compiler.

Variables can referenced in the expression using the syntax #variableName. The variables are passed in and out of the expression using the dictionary parameter in ExpressionEvaluator's GetValue or SetValue methods.

public static object GetValue(object root, string expression, IDictionary variables)

public static void SetValue(object root, string expression, IDictionary variables, object newValue)

The variable name is the key value of the dictionary. Example usage is shown below;

IDictionary vars = new Hashtable();
vars["newName"] = "Mike Tesla";
ExpressionEvaluator.GetValue(tesla, "Name = #newName", vars));

You can also use the dictionary as a place to store values of the object as they are evaluated inside the expression. For example to change Tesla's first name back again and keep the old value;

ExpressionEvaluator.GetValue(tesla, "{ #oldName = Name; Name = 'Nikola Tesla' }", vars);
String oldName = (String)vars["oldName"]; // Mike Tesla

Variable names can also be used inside indexers or maps instead of literal values. For example;

vars["prez"] = "president";
Inventor pupin = (Inventor) ExpressionEvaluator.GetValue(ieee, "Officers[#prez]", vars);

// sets the name of the president and returns its instance
ExpressionEvaluator.GetValue(ieee, "Officers['president'].( #this.Name = 'Nikola Tesla'; #this )")

// removes president from the Officers dictionary and returns removed instance
ExpressionEvaluator.GetValue(ieee, "Officers['president'].( #root.Officers.Remove('president'); #this )")

You can use the ternary operator for performing if-then-else conditional logic inside the expression. A minimal example is;

String aTrueString  = (String) ExpressionEvaluator.GetValue(null, "false ? 'trueExp' : 'falseExp'") // trueExp

In this case, the boolean false results in returning the string value 'trueExp'. A less artificial example is shown below

ExpressionEvaluator.SetValue(ieee, "Name", "IEEE");
IDictionary vars = new Hashtable();
vars["queryName"] = "Nikola Tesla";

string expression = @"IsMember(#queryName)
                      ? #queryName + ' is a member of the ' + Name + ' Society'
                      : #queryName + ' is not a member of the ' + Name + ' Society'";

String queryResultString = (String) ExpressionEvaluator.GetValue(ieee, expression, vars));

// queryResultString = "Nikola Tesla is a member of the IEEE Society"

List projection and selection are very powerful expression language features that allow you to transform the source list into another list by either projecting across its "columns", or selecting from its "rows". In other words, projection can be thought of as a column selector in a SQL SELECT statement, while selection would be comparable to the WHERE clause.

For example, let's say that we need a list of the cities where our inventors were born. This could be easily obtained by projecting on the PlaceOfBirth.City property:

IList placesOfBirth = (IList) ExpressionEvaluator.GetValue(ieee, "Members.!{PlaceOfBirth.City}")  // { 'Smiljan', 'Idvor' }

Or we can get the list of officers' names:

IList officersNames = (IList) ExpressionEvaluator.GetValue(ieee, "Officers.Values.!{Name}")  // { 'Nikola Tesla', 'Mihajlo Pupin' }

As you can see from the examples, projection uses !{projectionExpression} syntax and will return a new list of the same length as the original list but typically with the elements of a different type.

On the other hand, selection, which uses ?{projectionExpression} syntax, will filter the list and return a new list containing a subset of the original element list. For example, selection would allow us to easily get a list of Serbian inventors:

IList serbianInventors = (IList) ExpressionEvaluator.GetValue(ieee, "Members.?{Nationality == 'Serbian'}")  // { tesla, pupin }

Or to get a list of inventors that invented sonar:

IList sonarInventors = (IList) ExpressionEvaluator.GetValue(ieee, "Members.?{'Sonar' in Inventions}")  // { pupin }

Or we can combine selection and projection to get a list of sonar inventors' names:

IList sonarInventorsNames = (IList) ExpressionEvaluator.GetValue(ieee, "Members.?{'Sonar' in Inventions}.!{Name}")  // { 'Mihajlo Pupin' }

As a convenience, Spring.NET Expression Language also supports a special syntax for selecting the first or last match. Unlike regular selection, which will return an empty list if no matches are found, first or last match selection expression will either return an instance of the matched element, or null if no matching elements were found. In order to return a first match you should prefix your selection expression with ^{ instead of ?{, and to return last match you should use ${ prefix:

ExpressionEvaluator.GetValue(ieee, "Members.^{Nationality == 'Serbian'}.Name")  // 'Nikola Tesla'
ExpressionEvaluator.GetValue(ieee, "Members.${Nationality == 'Serbian'}.Name")  // 'Mihajlo Pupin'

Notice that we access the Name property directly on the selection result, because an actual matched instance is returned by the first and last match expression instead of a filtered list.

In addition to list projection and selection, Spring.NET Expression Language also supports several collection processors, such as distinct, nonNull and sort, as well as a number of commonly used aggregators, such as max, min, count, sum and average.

The difference between processors and aggregators is that processors return a new or transformed collection, while aggregators return a single value. Other than that, they are very similar -- both processors and aggregators are invoked on a collection node using standard method invocation expression syntax, which makes them very simple to use and allows easy chaining of multiple processors.

ExpressionEvaluator.GetValue(null, "{1, 5, -3}.count()")  // 3
ExpressionEvaluator.GetValue(null, "count()")  // 0

ExpressionEvaluator.GetValue(null, "{1, 5, -3, 10}.sum()")  // 13 (int)
ExpressionEvaluator.GetValue(null, "{5, 5.8, 12.2, 1}.sum()")  // 24.0 (double)

ExpressionEvaluator.GetValue(null, "{1, 5, -4, 10}.average()")  // 3
ExpressionEvaluator.GetValue(null, "{1, 5, -2, 10}.average()")  // 3.5

ExpressionEvaluator.GetValue(null, "{1, 5, -3, 10}.min()")  // -3
ExpressionEvaluator.GetValue(null, "{'abc', 'efg', 'xyz'}.min()")  // 'abc'

ExpressionEvaluator.GetValue(null, "{1, 5, -3, 10}.max()")  // 10
ExpressionEvaluator.GetValue(null, "{'abc', 'efg', 'xyz'}.max()")  // 'xyz'

ExpressionEvaluator.GetValue(null, "{ 'abc', 'xyz', null, 'abc', 'def', null}.nonNull()")  // { 'abc', 'xyz', 'abc', 'def' }
ExpressionEvaluator.GetValue(null, "{ 'abc', 'xyz', null, 'abc', 'def', null}.nonNull().distinct().sort()")  // { 'abc', 'def', 'xyz' }

ExpressionEvaluator.GetValue(null, "{ 'abc', 'xyz', 'abc', 'def', null, 'def' }.distinct(true).sort()")  // { null, 'abc', 'def', 'xyz' }
ExpressionEvaluator.GetValue(null, "{ 'abc', 'xyz', 'abc', 'def', null, 'def' }.distinct(false).sort()")  // { 'abc', 'def', 'xyz' }

ExpressionEvaluator.GetValue(null, "{1.2, 5.5, -3.3}.sort()")  // { -3.3, 1.2, 5.5 }
ExpressionEvaluator.GetValue(null, "{ 'abc', 'xyz', 'abc', 'def', null, 'def' }.sort()")  // { null, 'abc', 'abc', 'def', 'def', 'xyz' }

object[] arr = new object[] { "0", 1, 1.1m, "1.1", 1.1f };
decimal[] result = (decimal[]) ExpressionEvaluator.GetValue(arr, "convert(decimal)");

object[] arr = new object[] { "0", 1, 2.1m, "3", 4.1f };
object[] result = new ArrayList( (ICollection) ExpressionEvaluator.GetValue(arr, "reverse()") ).ToArray(); // { 4.1f, "3", 2.1m, 1, "0" }            

// orderBy expression
IExpression exp = Expression.Parse("orderBy('ToString()')");
object[] input = new object[] { 'b', 1, 2.0, "a" };
object[] ordered = exp.GetValue(input);  // { 1, 2.0, "a", 'b' }


// SpEL lambda expressions
IExpression exp = Expression.Parse("orderBy({|a,b| $a.ToString().CompareTo($b.ToString())})");
object[] input = new object[] { 'b', 1, 2.0, "a" };
object[] ordered = exp.GetValue(input);  // { 1, 2.0, "a", 'b' }

Hashtable vars = new Hashtable();
Expression.RegisterFunction( "compare", "{|a,b| $a.ToString().CompareTo($b.ToString())}", vars);
exp = Expression.Parse("orderBy(#compare)");
ordered = exp.GetValue(input, vars);  // { 1, 2.0, "a", 'b' }

// .NET delegate
private delegate int CompareCallback(object x, object y);
private int CompareObjects(object x, object y)
{
  if (x == y) return 0;
  return x.ToString().CompareTo(""+y);
}

Hashtable vars = new Hashtable();
vars["compare"] = new CompareCallback(CompareObjects);

IExpression exp = Expression.Parse("orderBy(#compare)");
object[] input = new object[] { 'b', 1, 2.0, "a" };
object[] ordered = exp.GetValue(input);  // { 1, 2.0, "a", 'b' }

        public class IntEvenSumCollectionProcessor : ICollectionProcessor
        {
            public object Process(ICollection source, object[] args)
            {
                object total = 0d;
                foreach (object item in source)
                {
                    if (item != null)
                    {
                        if (NumberUtils.IsInteger(item))
                        {
                            if ((int)item % 2 == 0)
                            {
                                total = NumberUtils.Add(total, item);
                            }
                        }
                        else
                        {
                            throw new ArgumentException("Sum can only be calculated for a collection of numeric values.");
                        }
                    }
                }

                return total;
            }
        }


        public void DoWork()
        {
            Hashtable vars = new Hashtable();
            vars["EvenSum"] = new IntEvenSumCollectionProcessor();
            int result = (int)ExpressionEvaluator.GetValue(null, "{1, 2, 3, 4}.EvenSum()", vars));  // 6
        }

Expressions can refer to objects that are declared in Spring's application context using the syntax @(contextName:objectName). If no contextName is specified the default root context name (Spring.RootContext) is used. Using the application context defined in the MovieFinder example from Chapter 41, IoC Quickstarts, the following expression returns the number of movies directed by Roberto Benigni.

public static void Main()
{
  . . .

// Retrieve context defined in the spring/context section of
// the standard .NET configuration file.
IApplicationContext ctx = ContextRegistry.GetContext();

int numMovies = (int) ExpressionEvaluator.GetValue(null,
                      "@(MyMovieLister).MoviesDirectedBy('Roberto Benigni').Length");

  . . .
}

The variable numMovies is evaluated to 2 in this example.

A somewhat advanced, but a very powerful feature of Spring.NET Expression Language are lambda expressions. Lambda expressions allow you to define inline functions, which can then be used within your expressions just like any other function or method. You may also use .NET delegates as described in the next section.

The syntax for defining lambda expressions is:

#functionName = {|argList| functionBody }

For example, you could define a max function and call it like this:

ExpressionEvaluator.GetValue(null, "(#max = {|x,y| $x > $y ? $x : $y }; #max(5,25))", new Hashtable())  // 25

As you can see, any arguments defined for the expression can be referenced within the function body using a local variable syntax, $varName. Invocation of the function defined using lambda expression is as simple as specifying the comma-separated list of function arguments in parentheses, after the function name.

Lambda expressions can be recursive, which means that you can invoke the function within its own body:

ExpressionEvaluator.GetValue(null, "(#fact = {|n| $n <= 1 ? 1 : $n * #fact($n-1) }; #fact(5))", new Hashtable())  // 120

Notice that in both examples above we had to specify a variables parameter for the GetValue method. This is because lambda expressions are actually nothing more than parameterized variables and we need variables dictionary in order to store them. If you don't specify a valid IDictionary instance for the variables parameter, you will get a runtime exception.

Also, in both examples above we used an expression list in order to define and invoke a function in a single expression. However, more likely than not, you will want to define your functions once and then use them within as many expressions as you need. Spring.NET provides an easy way to pre-register your lambda expressions by exposing a static Expression.RegisterFunction method, which takes function name, lambda expression and variables dictionary to register function in as parameters:

IDictionary vars = new Hashtable();
Expression.RegisterFunction("sqrt", "{|n| Math.Sqrt($n)}", vars);
Expression.RegisterFunction("fact", "{|n| $n <= 1 ? 1 : $n * #fact($n-1)}", vars);

Once the function registration is done, you can simply evaluate an expression that uses these functions, making sure that the vars dictionary is passed as a parameter to expression evaluation engine:

ExpressionEvaluator.GetValue(null, "#fact(5)", vars)  // 120
ExpressionEvaluator.GetValue(null, "#sqrt(9)", vars)  // 3

Finally, because lambda expressions are treated as variables, they can be assigned to other variables or passed as parameters to other lambda expressions. In the following example we are defining a delegate function that accepts function f as the first argument and parameter n that will be passed to function f as the second. Then we invoke the functions registered in the previous example, as well as the lambda expression defined inline, through our delegate:

Expression.RegisterFunction("delegate", "{|f, n| $f($n) }", vars);
ExpressionEvaluator.GetValue(null, "#delegate(#sqrt, 4)", vars)  // 2
ExpressionEvaluator.GetValue(null, "#delegate(#fact, 5)", vars)  // 120
ExpressionEvaluator.GetValue(null, "#delegate({|n| $n ^ 2 }, 5)", vars)  // 25

While this particular example is not particularly useful, it does demonstrate that lambda expressions are indeed treated as nothing more than parameterized variables, which is important to remember.

Delegate expressions allow you to refer to .NET delegates which can then be used within your expressions just like any other function or method.

For example, you can define a max delegate and call it like this

private delegate double DoubleFunctionTwoArgs(double arg1, double arg2);

private double Max(double arg1, double arg2)
{
  return Math.Max(arg1, arg2);
}


public void DoWork()
{
  Hashtable vars = new Hashtable();
  vars["max"] = new DoubleFunctionTwoArgs(Max);
  double result = (double) ExpressionEvaluator.GetValue(null, "#max(5,25)", vars);  // 25
}

If you do not specify a root object, i.e. pass in null, then the expressions evaluated either have to be literal values, i.e. ExpressionEvaluator.GetValue(null, "2 + 3.14"), refer to classes that have static methods or properties, i.e. ExpressionEvaluator.GetValue(null, "DateTime.Today"), create new instances of objects, i.e. ExpressionEvaluator.GetValue(null, "new DateTime(2004, 8, 14)") or refer to other objects such as those in the variable dictionary or in the IoC container. The latter two usages will be discussed later.