How to compare objects by multiple fields

Viewed 506166

Assume you have some objects which have several fields they can be compared by:

public class Person {

    private String firstName;
    private String lastName;
    private String age;

    /* Constructors */

    /* Methods */

}

So in this example, when you ask if:

a.compareTo(b) > 0

you might be asking if a's last name comes before b's, or if a is older than b, etc...

What is the cleanest way to enable multiple comparison between these kinds of objects without adding unnecessary clutter or overhead?

  • java.lang.Comparable interface allows comparison by one field only
  • Adding numerous compare methods (i.e. compareByFirstName(), compareByAge(), etc...) is cluttered in my opinion.

So what is the best way to go about this?

23 Answers

You should implement Comparable <Person>. Assuming all fields will not be null (for simplicity sake), that age is an int, and compare ranking is first, last, age, the compareTo method is quite simple:

public int compareTo(Person other) {
    int i = firstName.compareTo(other.firstName);
    if (i != 0) return i;

    i = lastName.compareTo(other.lastName);
    if (i != 0) return i;

    return Integer.compare(age, other.age);
}

(from Ways to sort lists of objects in Java based on multiple fields)

Working code in this gist

Using Java 8 lambda's (added April 10, 2019)

Java 8 solves this nicely by lambda's (though Guava and Apache Commons might still offer more flexibility):

Collections.sort(reportList, Comparator.comparing(Report::getReportKey)
            .thenComparing(Report::getStudentNumber)
            .thenComparing(Report::getSchool));

Thanks to @gaoagong's answer below.

Messy and convoluted: Sorting by hand

Collections.sort(pizzas, new Comparator<Pizza>() {  
    @Override  
    public int compare(Pizza p1, Pizza p2) {  
        int sizeCmp = p1.size.compareTo(p2.size);  
        if (sizeCmp != 0) {  
            return sizeCmp;  
        }  
        int nrOfToppingsCmp = p1.nrOfToppings.compareTo(p2.nrOfToppings);  
        if (nrOfToppingsCmp != 0) {  
            return nrOfToppingsCmp;  
        }  
        return p1.name.compareTo(p2.name);  
    }  
});  

This requires a lot of typing, maintenance and is error prone.

The reflective way: Sorting with BeanComparator

ComparatorChain chain = new ComparatorChain(Arrays.asList(
   new BeanComparator("size"), 
   new BeanComparator("nrOfToppings"), 
   new BeanComparator("name")));

Collections.sort(pizzas, chain);  

Obviously this is more concise, but even more error prone as you lose your direct reference to the fields by using Strings instead (no typesafety, auto-refactorings). Now if a field is renamed, the compiler won’t even report a problem. Moreover, because this solution uses reflection, the sorting is much slower.

Getting there: Sorting with Google Guava’s ComparisonChain

Collections.sort(pizzas, new Comparator<Pizza>() {  
    @Override  
    public int compare(Pizza p1, Pizza p2) {  
        return ComparisonChain.start().compare(p1.size, p2.size).compare(p1.nrOfToppings, p2.nrOfToppings).compare(p1.name, p2.name).result();  
        // or in case the fields can be null:  
        /* 
        return ComparisonChain.start() 
           .compare(p1.size, p2.size, Ordering.natural().nullsLast()) 
           .compare(p1.nrOfToppings, p2.nrOfToppings, Ordering.natural().nullsLast()) 
           .compare(p1.name, p2.name, Ordering.natural().nullsLast()) 
           .result(); 
        */  
    }  
});  

This is much better, but requires some boiler plate code for the most common use case: null-values should be valued less by default. For null-fields, you have to provide an extra directive to Guava what to do in that case. This is a flexible mechanism if you want to do something specific, but often you want the default case (ie. 1, a, b, z, null).

Sorting with Apache Commons CompareToBuilder

Collections.sort(pizzas, new Comparator<Pizza>() {  
    @Override  
    public int compare(Pizza p1, Pizza p2) {  
        return new CompareToBuilder().append(p1.size, p2.size).append(p1.nrOfToppings, p2.nrOfToppings).append(p1.name, p2.name).toComparison();  
    }  
});  

Like Guava’s ComparisonChain, this library class sorts easily on multiple fields, but also defines default behavior for null values (ie. 1, a, b, z, null). However, you can’t specify anything else either, unless you provide your own Comparator.

Thus

Ultimately it comes down to flavor and the need for flexibility (Guava’s ComparisonChain) vs. concise code (Apache’s CompareToBuilder).

Bonus method

I found a nice solution that combines multiple comparators in order of priority on CodeReview in a MultiComparator:

class MultiComparator<T> implements Comparator<T> {
    private final List<Comparator<T>> comparators;

    public MultiComparator(List<Comparator<? super T>> comparators) {
        this.comparators = comparators;
    }

    public MultiComparator(Comparator<? super T>... comparators) {
        this(Arrays.asList(comparators));
    }

    public int compare(T o1, T o2) {
        for (Comparator<T> c : comparators) {
            int result = c.compare(o1, o2);
            if (result != 0) {
                return result;
            }
        }
        return 0;
    }

    public static <T> void sort(List<T> list, Comparator<? super T>... comparators) {
        Collections.sort(list, new MultiComparator<T>(comparators));
    }
}

Ofcourse Apache Commons Collections has a util for this already:

ComparatorUtils.chainedComparator(comparatorCollection)

Collections.sort(list, ComparatorUtils.chainedComparator(comparators));

You can implement a Comparator which compares two Person objects, and you can examine as many of the fields as you like. You can put in a variable in your comparator that tells it which field to compare to, although it would probably be simpler to just write multiple comparators.

Instead of comparison methods you may want to just define several types of "Comparator" subclasses inside the Person class. That way you can pass them into standard Collections sorting methods.

//Following is the example in jdk 1.8
package com;
import java.util.ArrayList;
import java.util.Comparator;
import java.util.List;

class User {
    private String firstName;
    private String lastName;
    private Integer age;

    public Integer getAge() {
        return age;
    }

    public User setAge(Integer age) {
        this.age = age;
        return this;
    }

    public String getFirstName() {
        return firstName;
    }

    public User setFirstName(String firstName) {
        this.firstName = firstName;
        return this;
    }

    public String getLastName() {
        return lastName;
    }

    public User setLastName(String lastName) {
        this.lastName = lastName;
        return this;
    }

}

public class MultiFieldsComparision {

    public static void main(String[] args) {
        List<User> users = new ArrayList<User>();

        User u1 = new User().setFirstName("Pawan").setLastName("Singh").setAge(38);
        User u2 = new User().setFirstName("Pawan").setLastName("Payal").setAge(37);
        User u3 = new User().setFirstName("Anuj").setLastName("Kumar").setAge(60);
        User u4 = new User().setFirstName("Anuj").setLastName("Kumar").setAge(43);
        User u5 = new User().setFirstName("Pawan").setLastName("Chamoli").setAge(44);
        User u6 = new User().setFirstName("Pawan").setLastName("Singh").setAge(5);

        users.add(u1);
        users.add(u2);
        users.add(u3);
        users.add(u4);
        users.add(u5);
        users.add(u6);

        System.out.println("****** Before Sorting ******");

        users.forEach(user -> {
            System.out.println(user.getFirstName() + " , " + user.getLastName() + " , " + user.getAge());
        });

        System.out.println("****** Aftre Sorting ******");

        users.sort(
                Comparator.comparing(User::getFirstName).thenComparing(User::getLastName).thenComparing(User::getAge));

        users.forEach(user -> {
            System.out.println(user.getFirstName() + " , " + user.getLastName() + " , " + user.getAge());
        });

    }

}

Code implementation of the same is here if we have to sort the Person object based on multiple fields.

import java.util.ArrayList;
import java.util.Collections;
import java.util.Comparator;

public class Person {

private String firstName;
private String lastName;
private int age;

public String getFirstName() {
    return firstName;
}

public void setFirstName(String firstName) {
    this.firstName = firstName;
}

public String getLastName() {
    return lastName;
}

public void setLastName(String lastName) {
    this.lastName = lastName;
}

public int getAge() {
    return age;
}

public void setAge(int age) {
    this.age = age;
}

public Person(String firstName, String lastName, int age) {
    this.firstName = firstName;
    this.lastName = lastName;
    this.age = age;
}


static class PersonSortingComparator implements Comparator<Person> {

    @Override
    public int compare(Person person1, Person person2) {

        // for first name comparison
        int firstNameCompare = person1.getFirstName().compareTo(person2.getFirstName());

        // for last name comparison
        int lastNameCompare = person1.getLastName().compareTo(person2.getLastName());

        // for last name comparison
        int ageCompare = person1.getAge() - person2.getAge();

        // Now comparing
        if (firstNameCompare == 0) {
            if (lastNameCompare == 0) {
                return ageCompare;
            }
            return lastNameCompare;
        }
        return firstNameCompare;
    }
}

public static void main(String[] args) {
    Person person1 = new Person("Ajay", "Kumar", 27);
    Person person2 = new Person("Ajay","Gupta", 23);
    Person person3 = new Person("Ajay","Kumar", 22);


    ArrayList<Person> persons = new ArrayList<>();
    persons.add(person1);
    persons.add(person2);
    persons.add(person3);


    System.out.println("Before Sorting:\n");
    for (Person person : persons) {
        System.out.println(person.firstName + " " + person.lastName + " " + person.age);
    }

    Collections.sort(persons, new PersonSortingComparator());

    System.out.println("After Sorting:\n");
    for (Person person : persons) {
        System.out.println(person.firstName + " " + person.lastName + " " + person.age);
    }
}

}

I think it'd be more confusing if your comparison algorithm were "clever". I'd go with the numerous comparison methods you suggested.

The only exception for me would be equality. For unit testing, it's been useful to me to override the .Equals (in .net) in order to determine if several fields are equal between two objects (and not that the references are equal).

If there are multiple ways a user might order person, you could also have multiple Comparators setup as constants somewhere. Most of the sort operations and sorted collections take a comparator as a parameter.

Better late than never - if you're looking for unnecessary clutter or overhead then it's hard to beat the following in terms of least code/fast execution at the same time.

Data class:

public class MyData {
    int id;
    boolean relevant;
    String name;
    float value;
}

Comparator:

public class MultiFieldComparator implements Comparator<MyData> {
    @Override
    public int compare(MyData dataA, MyData dataB) {
        int result;
        if((result = Integer.compare(dataA.id, dataB.id)) == 0 &&
           (result = Boolean.compare(dataA.relevant, dataB.relevant)) == 0 &&
           (result = dataA.name.compareTo(dataB.name)) == 0)
            result = Float.compare(dataA.value, dataB.value);
        return result;
    }
}

If you are just looking to sort a collection by a custom order then the following is even cleaner:

myDataList.sort((dataA, dataB) -> {
    int result;
    if((result = Integer.compare(dataA.id, dataB.id)) == 0 &&
       (result = Boolean.compare(dataA.relevant, dataB.relevant)) == 0 &&
       (result = dataA.name.compareTo(dataB.name)) == 0)
        result = Float.compare(dataA.value, dataB.value);
    return result;
});

Java 8 through lambda way we can compare by method reference. Student POJO

public class Student {
int id;
String firstName;
String lastName;
String subject;

public Student(int id, String firstName, String lastName, String subject) {
    this.id = id;
    this.firstName = firstName;
    this.lastName = lastName;
    this.subject = subject;
}
enter code here

Now we can sort based on

1. id->FirstName->LastName->Subject 2. Subject->id->FirstName->LastName

We will use Comparator in array Stream

public class TestComprator {
public static void main(String[] args) {
    Student s1= new Student(108, "James", "Testo", "Physics");
    Student s2= new Student(101, "Fundu", "Barito", "Chem");
    Student s3= new Student(105, "Sindhu", "Sharan", "Math");
    Student s4= new Student(98, "Rechel", "Stephen", "Physics");
    System.out.printf("----------id->FirstName->LastName->Subject-------------");
    Arrays.asList(s1,s2,s3,s4).stream()
            .sorted(Comparator.comparing(Student::getId)
                    .thenComparing(Student::getFirstName)
                .thenComparing(Student::getLastName)
                .thenComparing(Student::getSubject))
            .forEach(System.out::println);

    System.out.printf("----Subject->id->FirstName->LastName ------\n");
    Arrays.asList(s1,s2,s3,s4).stream()
            .sorted(Comparator. comparing(Student::getSubject)
                    .thenComparing(Student::getId)
                    .thenComparing(Student::getFirstName)
                    .thenComparing(Student::getLastName)
                   )
            .forEach(System.out::println);
}

}

Output:-

`----------id->FirstName->LastName->Subject-------------
Student{id=98, firstName='Rechel', lastName='Stephen', subject='Physics'}
Student{id=101, firstName='Fundu', lastName='Barito', subject='Chem'}
Student{id=105, firstName='Sindhu', lastName='Sharan', subject='Math'}
Student{id=108, firstName='James', lastName='Testo', subject='Physics'}
 ----Subject->id->FirstName->LastName ------
Student{id=101, firstName='Fundu', lastName='Barito', subject='Chem'}
Student{id=105, firstName='Sindhu', lastName='Sharan', subject='Math'}
Student{id=98, firstName='Rechel', lastName='Stephen', subject='Physics'}
Student{id=108, firstName='James', lastName='Testo', subject='Physics'}

If you implement the Comparable interface, you'll want to choose one simple property to order by. This is known as natural ordering. Think of it as the default. It's always used when no specific comparator is supplied. Usually this is name, but your use case may call for something different. You are free to use any number of other Comparators you can supply to various collections APIs to override the natural ordering.

Also note that typically if a.compareTo(b) == 0, then a.equals(b) == true. It's ok if not but there are side effects to be aware of. See the excellent javadocs on the Comparable interface and you'll find lots of great information on this.

It is easy to compare two objects with hashcode method in java`

public class Sample{

  String a=null;
  String b=null;

  public Sample(){
      a="s";
      b="a";
  }
  public Sample(String a,String b){
      this.a=a;
      this.b=b;
  }
  public static void main(String args[]){
      Sample f=new Sample("b","12");
      Sample s=new Sample("b","12");
      //will return true
      System.out.println((s.a.hashCode()+s.b.hashCode())==(f.a.hashCode()+f.b.hashCode()));

      //will return false
      Sample f=new Sample("b","12");
      Sample s=new Sample("b","13");
      System.out.println((s.a.hashCode()+s.b.hashCode())==(f.a.hashCode()+f.b.hashCode()));

}
Related