Float-type Array Index (Error:Illegal use of floating point)

[raysa]

Hi,

I need to do a very simple task like:

Code: Select all

y[0.5] = 1;
y[1.5] = 2;
y[2.5] = 3;

but I always get "Illegal use of floating point" error message.
It seems like the array index should only contain a solid integer.

Is there any way to do that type of array assignment
(float-type of array index)?
Any help would be very much appreciated.

Thank you,

raysa

[A1]

Is there any way to do that type of array assignment
(float-type of array index)?

It is not possible in any p-language till now.

[mf]

For arrays, expression v[n] has the same meaning as *(v+n).
That is, v is interpreted as a pointer to the beginning of array, this pointer is incremented by n and dereferenced. Of course it's nonsensical to increment a pointer by a fraction.

What you probably need is STL's map or hash_map class. Declare v as:
map<double, int> v;
and your code will work. Don't forget to add: #include <map>.

[Moha]

Another solution is:

Code: Select all

double array[10000];
#define I(a) (((int)(a))*2+((a)-(int)(a))>0)

for accessing the array you can use array[I(1.5)]. but please pay attention that you should know the exact number of digits after decimal point.

[raysa]

Thank's A1, mf and Moha for the insights!

@Moha
Can you tell us what your code do?
I'm applying your code in my listing:

Code: Select all

scanf ("%f",y[I(1.5)]);

but it got "Subscripting missing: ]" and "Expression syntax" error.

I can't see any typo in the syntax. Any thought?

[chunyi81]

raysa:

It should be

Code: Select all

scanf("%lf",&y[I(1.5)]);

if y is a double array of course.

You forgot the '&'. Also check your y array declaration.

And I think moha's code is actually a simple hash function which computes the integer index for a floating point value.

[raysa]

Nevermind. It's my mistake.
I was putting a semicolon behind the #define syntax.

Anyway, please tell us what your code does...
What part in your code is to modify the number of digits after decimal point?

Thank's,

raysa

[raysa]

This is weird...
I can retrive y[1.5] value using y[1.8] call.

Code: Select all

printf ("y = "); scanf ("%lf",&y[I(1.5)]);
printf ("y = %lf\n", y[I(1.8)]);

How to fix this modifying Moha's hash?

[A1]

I don't understand what Moha try to say !
His I() function always return true or false(1\0)!
so how it can be used to iterate in a 10000 size array!!

[Moha]

Ok, The I(a) is suitable for all numbers which are in the form of n*0.5.
If you want to use it for all fractional numbers with one digit after decimal place, you should rewite it in this form:

Code: Select all

#define I(a) ((int)((a)*10))

These functions are a one to one mapping.
for the first hash function 1.8 is mapped to place that 1.5 is stored.
I did a very nasty mistake in my previous Hash function it should changed into

Code: Select all

#define I(a) (((int)(a))*2+(((a)-(int)(a))>0)) 

[Martin Macko]

#define I(a) (((int)(a))*2+(((a)-(int)(a))>0))

If you are using only numbers of the form n*0.5, where n>=0, you can simplfy it and write just the following, can't you?

#define I(a) ((int)((a)*2))

[raysa]

it's all OK now.
Thank you everyone.. Especially you, Moha!
Your code WORKS like charm!