JAVA - 表达式解析和评估库

Question

I'm looking for a JAVA library to parse & evaluate expression. I searched and tried some libraries like Apache's JEXL and Jeval, but they are not exactly what I need.

My requirements:

1. Support all value types (ie int,double,boolean,String, etc.)
2. Support all known mathematical & logical operators (+,-,*,<,<=,etc.)
3. Support variables (without any special notation - for example in Jeval variable a should be written like #{a} - not good enough for me)
4. Support custom functions - with type enforcement and validation

Any recommendations?

Answer 1

Try Janino . It's a runtime in-memory compiler that can be used as an expression evaluator. Maybe it is the right thing for you.

Answer 2

Like suggested, you can use JavaScript. But also you could take a look at Spring EL which has support for your requirements.

Answer 3

You can try mXparser - it supports significant part of your requirements:

1. It is based on double, so int is supported, additionally boolean is supported as true = 1 and false = 0. Unfortunately strings are not supported.

Boolean example:

import org.mariuszgromada.math.mxparser.*;
...
...
Constant T = new Constant("T = 1");
Constant F = new Constant("F = 0");
Expression e = new Expression("T && (F || (F && T))", T, F);
System.out.println(e.getExpressionString() + " = " + e.calculate());

Result:

T && (F || (F && T)) = 0.0

2. mXparser has broad support for operators, functions , etc.. Check mXparser math collection . What is nice you can use help functionality inside the library.

Example:

import org.mariuszgromada.math.mxparser.*;
...
...
mXparser.consolePrintHelp("operator");

Result:

Help content: 

    2. +                   <Operator>              addition
    3. -                   <Operator>              subtraction
    4. *                   <Operator>              multiplication
    5. /                   <Operator>              division
    6. ^                   <Operator>              exponentiation
    7. !                   <Operator>              factorial
    8. #                   <Operator>              modulo function
    9. &                   <Boolean Operator>      logical conjunction (AND)
   10. &&                  <Boolean Operator>      logical conjunction (AND)
   11. /\                  <Boolean Operator>      logical conjunction (AND)
   12. ~&                  <Boolean Operator>      NAND - Sheffer stroke
   13. ~&&                 <Boolean Operator>      NAND - Sheffer stroke
   14. ~/\                 <Boolean Operator>      NAND - Sheffer stroke
   15. |                   <Boolean Operator>      logical disjunction (OR)
   16. ||                  <Boolean Operator>      logical disjunction (OR)
   17. \/                  <Boolean Operator>      logical disjunction (OR)
   18. ~|                  <Boolean Operator>      logical NOR
   19. ~||                 <Boolean Operator>      logical NOR
   20. ~\/                 <Boolean Operator>      logical NOR
   21. (+)                 <Boolean Operator>      exclusive or (XOR)
   22. -->                 <Boolean Operator>      implication (IMP)
   23. <--                 <Boolean Operator>      converse implication (CIMP)
   24. -/>                 <Boolean Operator>      material nonimplication (NIMP)
   25. </-                 <Boolean Operator>      converse nonimplication (CNIMP)
   26. <->                 <Boolean Operator>      logical biconditional (EQV)
   27. ~                   <Boolean Operator>      negation
   28. ¬                   <Boolean Operator>      negation
  162. add                 <Variadic Function>     (2.4) Summation operator add(a1,a2,a3,...,an)
  168. sum                 <Calculus Operator>     summation operator (SIGMA) sum(i, from, to, f(i,...))
  169. prod                <Calculus Operator>     product operator (PI) prod(i, from, to, f(i,...))
  170. int                 <Calculus Operator>     definite integral operator ( int(f(x,...), x, a, b) )
  171. der                 <Calculus Operator>     derivative operator ( der(f(x,...), x) ) 
  172. der-                <Calculus Operator>     left derivative operator ( der-(f(x,...), x) ) 
  173. der+                <Calculus Operator>     right derivative operator ( der+(f(x,...), x) ) 
  174. dern                <Calculus Operator>     n-th derivative operator ( dern(f(x,...), x) ) 
  175. diff                <Calculus Operator>     forward difference operator
  176. difb                <Calculus Operator>     backward difference operator
  177. avg                 <Calculus Operator>     (2.4) Average operator avg(i, from, to, f(i,...))
  178. vari                <Calculus Operator>     (2.4) Bias-corrected sample variance operator vari(i, from, to, f(i,...))
  179. stdi                <Calculus Operator>     (2.4) Bias-corrected sample standard deviation operator stdi(i, from, to, f(i,...))
  180. mini                <Calculus Operator>     (2.4) Minimum value mini(i, from, to, f(i,...))
  181. maxi                <Calculus Operator>     (2.4) Maximum value maxi(i, from, to, f(i,...))
  182. solve               <Calculus Operator>     (4.0) f(x) = 0 equation solving, function root finding: solve( f(x,...), x, a, b )
  301. @~                  <Bitwise Operator>      (4.0) Bitwise unary complement
  302. @&                  <Bitwise Operator>      (4.0) Bitwise AND
  303. @^                  <Bitwise Operator>      (4.0) Bitwise exclusive OR
  304. @|                  <Bitwise Operator>      (4.0) Bitwise inclusive OR
  305. @<<                 <Bitwise Operator>      (4.0) Signed left shift
  306. @>>                 <Bitwise Operator>      (4.0) Signed right shift

3. User defined variables and user defined constants are created without any special form.

Example:

import org.mariuszgromada.math.mxparser.*;
...
...
Argument x = new Argument("x = 10");
Constant y = new Constant("y = 2");
Expression e = new Expression("x/y", x, y);
System.out.println(e.getExpressionString() + " = " + e.calculate());

Result:

x/y = 5.0

Additionally please check: a)Tutorial - User defined arguments , b)Tutorial - User defined constants .

4. User defined functions are fully supported.

Example 1 - body defined in run-time:

import org.mariuszgromada.math.mxparser.*;
...
...
Function f = new Function("f(x,y) = x*y");
Expression e = new Expression("20-f(2,5)",f);
System.out.println(e.getExpressionString() + " = " + e.calculate());

Result 1

20-f(2,5) = 10.0

Example 2 - body extended via your own implementation:

import org.mariuszgromada.math.mxparser.*;
...
...
/*
 * Implementing FunctionExtension interface
 */
public class Addition implements FunctionExtension {
   double x;
   double y;
   public Addition() {
      x = Double.NaN;
      y = Double.NaN;
   }
   public Addition(double x, double y) {
      this.x = x;
      this.y = y;
   }
   public int getParametersNumber() {
      return 2;
   }
   public void setParameterValue(int argumentIndex, double argumentValue) {
      if (argumentIndex == 0) x = argumentValue;
      if (argumentIndex == 1) y = argumentValue;
   }
   public double calculate(double... params) {
      return x+y;
   }
   public FunctionExtension clone() {
      return new Addition(x, y);
   }   
}

/*
* Creating extended function
*/
Function f = new Function("f", new Addition());
mXparser.consolePrintln("f.calculate(1,2) = " + f.calculate(1,2) );
/*
* Using extended function in expression
*/
Expression e = new Expression("f(2,3)", f);
System.out.println(e.getExpressionString() + " = " + e.calculate() );

Result 2:

f.calculate(1,2) = 3.0
f(2,3) = 5.0

Additionally it is worth to follow the whole mXparser Tutorial .

Found recently - in case you would like to try the syntax (and see the advanced use case) you can download the Scalar Calculator app that is powered by mXparser.

Best regards

Answer 4

Here would be a couple of workaround solutions that you could choose from, if you don't find an actual Java expression evaluation library:

• Evaluate your expressions using XPath.
  • Pros: XPath knows logical operators, and you can implement variables and custom functions using Xalan's extensions
  • Cons: XPath has fewer types than Java
• Evaluate your expressions using JavaScript.
  • Pros: Javascript is very flexible and will still be suitable when your requirements tighten. You can implement variables and custom functions using Javascript as well
  • Cons: Javascript has fewer types than Java
• Evaluate your expressions using JSP's expression language (eg with JUEL )