[Fortran] Rounding up Random Numbers

[Lukejambo]

Hi,

So I'm writing a programme in Fortran95 atm and I want to produce an array of 1s and 0s.

I've used a random number and random seed generator to produce 10 numbers between 0 and 1 and I want to use a NINT statement to round these random numbers to 0 or 1.

However when I try this the rounded numbers are very large which I know shouldn't happen...

Any help will be much appreciated.

 

[DaveC426913]

Perhaps you could show your work, so we can critique it.

 

[Lukejambo]

Thanks for the quick reply Dave! Here is the code:

program random
implicit none
integer :: n
real :: a(10)
integer :: i(10)

n=10
call random_seed()
call random_number(a)

write(6,*) (a(n),n=1,10)
write(6,*) ''

i(n) = NINT(a(n))

write(6,*) (i(n),n=1,10)

end program random

 

[DaveC426913]

And the output?

If you would put in some debugging lines, of particular interest would be the values of a(n) just before the NINT function.

 

[Mark44]

Thanks for the quick reply Dave! Here is the code:

program random
implicit none
integer :: n
real :: a(10)
integer :: i(10)

n=10
call random_seed()
call random_number(a)

write(6,*) (a(n),n=1,10)
write(6,*) ''

i(n) = NINT(a(n))

write(6,*) (i(n),n=1,10)

end program random

I think it's this line: i(n) = NINT(a(n)). I believe that all this is doing is setting i(10) to the rounded value in a(10). i(1), i(2), ..., i(9) will have garbage values.
If you want all 10 elements of your i array to have values, do this:

do n = 1, 10
   i(n) = NINT(a(n))
end do

 

[DrClaude]

I think it's this line: i(n) = NINT(a(n)). I believe that all this is doing is setting i(10) to the rounded value in a(10). i(1), i(2), ..., i(9) will have garbage values.

It's even worse than that, as n has a value of 11 after the loop in write(6,*) (a(n),n=1,10). The original code could produce a segmentation fault.

If you want all 10 elements of your i array to have values, do this:

do n = 1, 10
   i(n) = NINT(a(n))
end do

In Fortran 95, the loop can be implicit:

   i = NINT(a)

 

[Lukejambo]

Thanks very much for your help, I've managed to produce a matrix of ones and zeroes which are all randomly placed.

 

[DaveC426913]

It would be awesome if you posted your result, so that others may learn.

 

[Lukejambo]

Here is the code:! This programme produces a 10x10 matrix of 0's and 1's using the random number and random seed function.

program randomex
implicit none
integer :: n
real :: a(100)
integer :: i(100)
open(unit=20, file='randomex')n=100
call random_seed()
call random_number(a)

! Below writes the matrix of random numbers, all between 0 and 1.

write(6,'(10f5.3)') (a(n),n=1,100)
write(20,'(10f5.3)') (a(n),n=1,100)
write(6,*) ''

! Rounding these random numbers up/down.

i = NINT(a)

! Writing the matrix to screen.

write(6,'(10i2)') (i(n),n=1,100)
write(20,'(10i2)') (i(n),n=1,100)
write(6,*) ''end program randomex

 

[Lukejambo]

And the output:0.7180.3280.1310.5310.1930.5790.1660.4610.7620.613
0.7380.1280.4020.4000.2200.9330.0800.7640.9990.505
0.6310.6970.8230.9630.6980.4480.9360.3010.9390.489
0.9860.4930.3490.4180.1700.8120.8390.0900.4720.117
0.4600.9490.5750.1400.0460.8190.8350.4690.0500.043
0.6670.3730.0860.7360.2130.4890.0480.2610.5900.725
0.3230.7440.3710.1320.5550.6510.4830.4160.3020.064
0.8560.3270.6190.4030.6840.0930.0150.6130.3940.412
0.1760.7080.7980.8920.9990.4480.9090.7350.3680.634
0.0250.6520.9440.8540.3770.2880.5530.1050.0340.424

1 0 0 1 0 1 0 0 1 1
1 0 0 0 0 1 0 1 1 1
1 1 1 1 1 0 1 0 1 0
1 0 0 0 0 1 1 0 0 0
0 1 1 0 0 1 1 0 0 0
1 0 0 1 0 0 0 0 1 1
0 1 0 0 1 1 0 0 0 0
1 0 1 0 1 0 0 1 0 0
0 1 1 1 1 0 1 1 0 1
0 1 1 1 0 0 1 0 0 0

 

[Mark44]

@Lukejambo, do you still have a question? Your output suggests that your program is working as intended, although the output of the unrounded numbers is harder to read than it should be. For the real numbers, you are printing them using this write statement

write(6,'(10i2)') (i(n),n=1,100)

Here you are using a format edit descriptor of 10i2 to print them in fields of 10 integers (i2), using a width of two columns for each. Since each number takes up more than two columns, the compiler overrides your I2 format descriptor.

A better choice would be to print them using a real format descriptor such as F10.8 for each number, followed by a space, X.

I think this would work, but I don't currently have a Fortran compiler to test it on.

write(6, '5(F10.2, X)') (i(n),n=1,100)

 

[DrClaude]

For the real numbers, you are printing them using this write statement

write(6,'(10i2)') (i(n),n=1,100)

You got the wrong line there, he is using

write(6,'(10f5.3)') (a(n),n=1,100)

But you are correct that it would be better to add a space

write(6,'(10(f5.3,x))') (a(n),n=1,100)

 

[Mark44]

Right you are @DrClaude, I was thinking the last two write statements created the output he posted -- the first for the raw numbers, and the second for the rounded ints.