I tried to use a boolean operator for an array operation :

a[] = b[] < c[];

It compiles but seems to only fill a[] with the result of b[0] < c[0].
Is there any "rational" reason to that?
And is there any way to use boolean operator for array operations?

On Thursday, 14 March 2013 at 13:58:45 UTC, n00b wrote:
> I tried to use a boolean operator for an array operation :
>
> a[] = b[] < c[];
>
> It compiles but seems to only fill a[] with the result of b[0] < c[0].
> Is there any "rational" reason to that?
> And is there any way to use boolean operator for array operations?

As far as I can tell, this is a bug.

n00b:

> Is there any "rational" reason to that?

It's not implemented (and it's a bug that it returns something different). Take a look in Bugzilla if it's already there.


> And is there any way to use boolean operator for array operations?

I think the only supported boolean vec operation is assignment.

Bye,
bearophile

> Take a look in Bugzilla if it's already there.

Code for a bug report:


import core.stdc.stdio: printf;
void main() {
    int[3] a = [1, 5, 1];
    int[3] b = [1, 1, 5];
    int[3] c;
    c[] = a[] < b[];
    printf("%d %d %d", c[0], c[1], c[2]);
}


Output:
0 0 0


Expected output:
0 0 1

Bye,
bearophile

This has already been reported
http://d.puremagic.com/issues/show_bug.cgi?id=5636

On Thursday, 14 March 2013 at 13:58:45 UTC, n00b wrote:
> I tried to use a boolean operator for an array operation :
>
> a[] = b[] < c[];
>
> It compiles but seems to only fill a[] with the result of b[0] < c[0].
> Is there any "rational" reason to that?
> And is there any way to use boolean operator for array operations?

Maybe this way:

int a[] = [1,2,3];
int b[] = [3,2,1];

a.zip(b).map!"a[0]<a[1]".writeln;
return;

On Thursday, 14 March 2013 at 13:58:45 UTC, n00b wrote:
> I tried to use a boolean operator for an array operation :
>
> a[] = b[] < c[];
>
> It compiles but seems to only fill a[] with the result of b[0] < c[0].
> Is there any "rational" reason to that?

Yes i think there is a rational reason. Check this:

int a[] = [1,0,0];
int b[] = [0,1,0];
int c[] = [1,1,0];

bool ab = a[] < b[]; // False.  a[0] > b[0]
bool ac = a[] < c[]; // True.  a[0] == c[0] but a[1] < c[1];

writeln(ab);
writeln(ac);
c[] = ab; // <-- assign!

You see? Your code do this:

a[] = (b[] < c[]);

On 03/14/2013 10:45 AM, Andrea Fontana wrote:

> You see? Your code do this:
>
> a[] = (b[] < c[]);

I thought the same thing at first but note the brackets after b and c. Those should make this an array-wise operation.

For all elements of a to be the same value, one would not write the brackets:

    a[] = b < c;

Some of the array-wise operations are confusing. :/

Ali

On Thursday, 14 March 2013 at 17:56:07 UTC, Ali Çehreli wrote:
> On 03/14/2013 10:45 AM, Andrea Fontana wrote:
>
> > You see? Your code do this:
> >
> > a[] = (b[] < c[]);
>
> I thought the same thing at first but note the brackets after b and c. Those should make this an array-wise operation.
>
> For all elements of a to be the same value, one would not write the brackets:
>
>     a[] = b < c;
>
> Some of the array-wise operations are confusing. :/
>
> Ali

Doesn't a[] means copy of a?
So copyofa[] < copyofb[] == bool

I don't think a boolean operator can be array-wise... I always think X < Y return a single bool, indipendently from X or Y type (class, array or what else). Math operations is a different topic in my mind.

Andrea Fontana:

> I always think X < Y return a single bool,

Having a practice with NumPy and the like, I think X < Y as returning as many bools as the length of X and Y:

>>> from numpy import *
>>> a = array([1, 5, 1])
>>> b = array([1, 1, 5])
>>> a < b
array([False, False,  True], dtype=bool)


Programmers with experience of Matlab, Matcad, etc, feel the same.

Bye,
bearophile