How are Hash Map values stored (in case the values are duplicates)?

Question

Does a Java HashMap save duplicate values as unique or does it save them individually?

Let's imagine I have the following key-value pairs value:

A -> "a very long string...."
B -> "another very long string...."
C -> "the same very long string from A key.... (but not same string instance)"

Will it be saved as 3 keys and 3 values, or as 3 keys and 2 values (as value for A is exactly the same as value for C)

My concern is about sizing as the values I am going to save are quite big and make no sense duplicate them.

You can consider making `StringBuilder` instances of the strings and store them. That way you are storing the references only.. — prasad_, Jan 21 '19 at 13:46

score 3 · Accepted Answer · answered Jan 21 '19 at 12:40

A HashMap can contain duplicate values, so in your example there would be 3 keys and 3 values in the Map.

However, if you are concerned about large instances being stored multiple times as values in your Map, you should store references to the same instance.

For example:

String val = "a very long String............";
String val2 = "a second long String.........";
map.put(1,val);
map.put(2,val2);
map.put(3,val);

There are only two large String instances. The Map contains 3 references to those Strings (two references to the first String and one reference to the second String), but references occupy a small amount of storage.

Andrew Tobilko · Answer 2 · 2019-01-21T13:27:16.187

2

It doesn't keep values, it holds references to values, so the problem boils down to the question "Are "a very long string...." and "the same very long string from A key...." the same object in memory?"

If they aren't compile-time constants and are obtained/constructed at runtime, they are different objects. String interning, along with String#intern, can help you here but has its limitations mentioned by @bratkartoffel.

What is Java String interning?
When should we use intern method of String on String literals
Performance penalty of String.intern()

edited Jan 21 '19 at 13:27

answered Jan 21 '19 at 12:50

Andrew Tobilko

48,120
14
91
142

1

Please, do NOT use intern() on very large string. This will quickly lead to problems as the instances are saved in a separate space (Permanent Generation, not heap). That space is limited and intended for the classes, methods and internal JVM stuff. – bratkartoffel Jan 21 '19 at 13:18

score 2 · Answer 3 · answered Jan 21 '19 at 13:05

But this does not concern HashMap, I mean it is not specific to it. Having a value as a String implies two other different things too - such as String interning or the fact that Strings are immutable. Also a HashMap does not copy the value (whatever that might mean), it only stores references to other objects - what you do with those Objects is entirely up to you. If you change an Object outside of the HashMap and then query that Map against that particular associated key, you will see the update.

score 1 · Answer 4 · answered Jan 21 '19 at 12:39

1

The Java HashMap does not allow duplicate keys, but it does allow duplicate values. So it would save them as 3 keys and 3 values.

In order for it to not save the value twice, you would have to check for it (for example by using the .containsValue() method).

answered Jan 21 '19 at 12:39

Niby

474
1
6
20

I suppose there is no `contains()` method declared in `Map` – Gyro Gearless Jan 21 '19 at 12:53
1

Take your pick from ```containsKey()``` and ```containsValue()```. – Jan 21 '19 at 14:07

How are Hash Map values stored (in case the values are duplicates)?

4 Answers4