Can an array contain integers and floats

Somebody asked me: can an array in java contain integers and floats? She got that question from a teacher.

Now my an开发者_高级运维swer was: yes, since you can declare an array of objects and store integers and floats in it.

But now I'm wondering if that is correct, since technically when you store Integer and Float objects in an array, it kind of does contain both types, but if you would "ask" the array he would tell you he contains Objects, and if I don't do bookkeeping or class checks there's no way to tell that there are integers and floats in the array.

On the other hand I still feel it might be the right answer since theoretically the array contains objects of both types.

So I'm asking for a smart opinion: if you were asked (in an interview, a test whatever) wether in java an array can contain integers and floats, yes or no? What would you answer?

A int of float does not fit into a Object[] array. However, by autoboxing java will put a Float or Integer into the array instead.

Both Float and Integer extend Number. So you can even make a array of numbers Number[]

Also, you can put a int into a float[], but java will then cast the int into a float. The other way around is also possible, but precision will be lost. (edit: Even from int->float precision can be lost. float->int may lose information about the overall magnitude of the value).

The conclusion would depend on the question. For primitive datatypes a array can not contain the other datatype. If you use a Object array (Integer, Float, Number) the answer would be yes.

This isn't a question that has a clear-cut "yes" or "no" answer.

I can see three distinct ways in which the question can be answered in the affirmative:

1. the array could be of type Number[] and could contain (references to) Float and Integer objects;
2. the array could be of type double[] and could contain floats and ints cast to double (N.B. both casts are lossless);
3. the array could be of type int[] and could store ints as-as and floats converted to int using Float.floatToRawIntBits.

In case 3 (and arguably in case 2) you'd also need a parallel array that would record the type of the value stored in each element of the main array. Depending on the assumptions built into the question, this may well disqualify these as suitable answers.

If this question came up in an interview setting, I would outline the possibilities to the interviewer and ask which of the three interpretations, if any, they were looking for. If necessary, I would then elaborate further.

If you want to use a primitive array, you can use a double[] as double can store all possible int and float values. The only information lost is whether the number was originally an int or a float (there are 33 million integers which can be either)

To store an int or float in a double[]

double[] d = 
int i = 
float f = 
d[0] = i;
d[1] = f;

To retrieve an int or float value.

int i = (int) d[0];
float f = (float) d[1];

If you use an Object[] to store any combination of Objects including Integer and Float.

Interview questions can catch out interviewees by asking them questions they have never thought about, usually because there was no good reason to do this (possibly ever) If the question sounds odd, perhaps it should. ;)

In java there are only two data types - objects (everything that extends class Object) and primitives (int, byte, char and so on). Also, every primitive have its object sibling - for example int and java.lang.Integer

Technically, array can contain only objects.

But in java 5.0 it's possible to skip primitive-to-object conversion thanks to 'autoboxing' functions - basically it's replacing each call like
int myInt = 0; array[0] = myInt;
with
array[0] = new Integer(myInt);

This replacement is done automatically, however internally (in runtime) java machine will have array with objects (Integer), not with primitives (int), and this can affect performance of array operations.