How do I sort a list by different parameters at different timed
I think your enum approach is basically sound, but the switch statements really need a more object oriented approach. Consider:
enum PersonComparator implements Comparator<Person> {
ID_SORT {
public int compare(Person o1, Person o2) {
return Integer.valueOf(o1.getId()).compareTo(o2.getId());
}},
NAME_SORT {
public int compare(Person o1, Person o2) {
return o1.getFullName().compareTo(o2.getFullName());
}};
public static Comparator<Person> decending(final Comparator<Person> other) {
return new Comparator<Person>() {
public int compare(Person o1, Person o2) {
return -1 * other.compare(o1, o2);
}
};
}
public static Comparator<Person> getComparator(final PersonComparator... multipleOptions) {
return new Comparator<Person>() {
public int compare(Person o1, Person o2) {
for (PersonComparator option : multipleOptions) {
int result = option.compare(o1, o2);
if (result != 0) {
return result;
}
}
return 0;
}
};
}
}
An example of usage (with a static import).
public static void main(String[] args) {
List<Person> list = null;
Collections.sort(list, decending(getComparator(NAME_SORT, ID_SORT)));
}
You can create comparators for each of properties you might want to sort and then try "comparator chaining" :-) like this:
public class ChainedComparator<T> implements Comparator<T> {
private List<Comparator<T>> simpleComparators;
public ChainedComparator(Comparator<T>... simpleComparators) {
this.simpleComparators = Arrays.asList(simpleComparators);
}
public int compare(T o1, T o2) {
for (Comparator<T> comparator : simpleComparators) {
int result = comparator.compare(o1, o2);
if (result != 0) {
return result;
}
}
return 0;
}
}
One way is to create a Comparator
that takes as arguments a list of properties to sort by, as this example shows.
public class Person {
private int id;
private String name, address;
public static Comparator<Person> getComparator(SortParameter... sortParameters) {
return new PersonComparator(sortParameters);
}
public enum SortParameter {
ID_ASCENDING, ID_DESCENDING, NAME_ASCENDING,
NAME_DESCENDING, ADDRESS_ASCENDING, ADDRESS_DESCENDING
}
private static class PersonComparator implements Comparator<Person> {
private SortParameter[] parameters;
private PersonComparator(SortParameter[] parameters) {
this.parameters = parameters;
}
public int compare(Person o1, Person o2) {
int comparison;
for (SortParameter parameter : parameters) {
switch (parameter) {
case ID_ASCENDING:
comparison = o1.id - o2.id;
if (comparison != 0) return comparison;
break;
case ID_DESCENDING:
comparison = o2.id - o1.id;
if (comparison != 0) return comparison;
break;
case NAME_ASCENDING:
comparison = o1.name.compareTo(o2.name);
if (comparison != 0) return comparison;
break;
case NAME_DESCENDING:
comparison = o2.name.compareTo(o1.name);
if (comparison != 0) return comparison;
break;
case ADDRESS_ASCENDING:
comparison = o1.address.compareTo(o2.address);
if (comparison != 0) return comparison;
break;
case ADDRESS_DESCENDING:
comparison = o2.address.compareTo(o1.address);
if (comparison != 0) return comparison;
break;
}
}
return 0;
}
}
}
It can then be used in code for instance like this:
cp = Person.getComparator(Person.SortParameter.ADDRESS_ASCENDING,
Person.SortParameter.NAME_DESCENDING);
Collections.sort(personList, cp);
One approach would be to compose Comparator
s. This could be a library method (I'm sure it exists somewhere out there).
public static <T> Comparator<T> compose(
final Comparator<? super T> primary,
final Comparator<? super T> secondary
) {
return new Comparator<T>() {
public int compare(T a, T b) {
int result = primary.compare(a, b);
return result==0 ? secondary.compare(a, b) : result;
}
[...]
};
}
Use:
Collections.sort(people, compose(nameComparator, addressComparator));
Alternatively, note that Collections.sort
is a stable sort. If performance isn't absolutely crucial, you sort be the secondary order before the primary.
Collections.sort(people, addressComparator);
Collections.sort(people, nameComparator);