FORTRAN 90 help- matrix-mul-add- input from user defined file name

1. muzzy29

3
Hi everyone

I am completely new to the Fortran. I have to make a program in fortran 90 to either add or multiply the matrix on user demand and ask user to enter the file name inorder to read the desired matrices.

I came up with the the following code which does not read the program and produces runtime error before the open statement.
And i am gonna make subroutines for each case(add and multiply)
Can anyone help me identify the right way of making the source code in order to read the user defined input from input file and put the matrix on tothe screen as it is defined in the input file.
Following the formatting to input matrix

rowsa, colsa
2, 5, 6
8, 9, 7

Also fortran only does 10 * 10 matrix so how do i make sure it reads every entry of the program properly.

! Ask user for the desired operation
WRITE(*,*) "Enter the desired operation to perform as sum or multiply "

! Ask user to open file for the first matrix
WRITE(*,*) ""
WRITE(*,*) ""

WRITE(*,*) "Enter file name for the first matrix"

OPEN (UNIT=20, file='inputA', STATUS='OLD')

! Read dimension of matrix A

Do i= 1,rowsA

Do j=1,colsA

WRITE (*,*) "The first matrix is ",A

End do

End do

close (20)

! Ask user to open file for the Second matrix
WRITE(*,*) ""
WRITE(*,*) ""

WRITE(*,*) "Enter file name for the second matrix"

OPEN (UNIT=30, file='inputB', STATUS='OLD')

! Read dimension of matrix A

Do i= 1,rowsB

Do j=1,colsB

WRITE (*,*) "The second matrix is ",B

End do

End do

close (30)

Thanks very much.

Muzzy

2. gsal

963
Code (Text):

integer rowsA, colsA, rowsB, colsB, i, j
real A(100,100)
real B(100,100)
real C(100,100)
character operation, inputA*30, inputB*30

! Ask user for the desired operation
WRITE(*,*) "Enter the desired operation to perform as sum or multiply "

! Ask user to open file for the first matrix
WRITE(*,*) ""
WRITE(*,*) ""
WRITE(*,*) "Enter file name for the first matrix"
OPEN (UNIT=20, file=inputA, STATUS='OLD')

! Read dimension of matrix A
WRITE (*,*) "The first matrix is "
Do i= 1,rowsA
write(*,"(100F7.2,$)") (A(i,j),j=1,colsA) End do close (20) ! Ask user to open file for the Second matrix WRITE(*,*) "" WRITE(*,*) "" WRITE(*,*) "Enter file name for the second matrix" Read (*,*) inputB OPEN (UNIT=30, file=inputB, STATUS='OLD') ! Read dimension of matrix A READ (30,*) rowsB,colsB ! Read each matrix(array) elements WRITE (*,*) "The second matrix is " Do i= 1,rowsB READ (30,*) (B(i,j),j=1,colsB) write(*,"(100F7.2,$)") (B(i,j),j=1,colsB)
End do
close (30)

WRITE(*,*) ""
WRITE(*,*) ""
WRITE(*,*) "The result is:"
WRITE(*,*) ""
If (operation == '*') then
if (colsA.NE.rowsB) then
write(*,*) "Sorry, A*B cannot be performed;"
else
C = matmul(A(1:rowsA,1:colsA),B(1:rowsB,1:colsB))
Do i= 1,rowsA
WRITE(*,'(100F7.2)') (C(i,j),j=1,colsB)
End do
end if
else if (operation == '+') then
if (rowsA.NE.rowsB.or.colsA.NE.colsB) then
write(*,*) "Sorry, A+B cannot be performed;"
else
C = A(1:rowsA,1:colsA)+B(1:rowsB,1:colsB)
Do i= 1,rowsA
WRITE(*,'(100F7.2)') (C(i,j),j=1,colsB)
End do
end if
end if

3. muzzy29

3
THank you SO much it was really helpful :)

Just had a issue with multiplication part that seems like it just multiplies 1st row and 2nd row of A with first column of B and writing out third column of A as third column of result (matrix C) and 2nd column of matrix B in the 2nd column of result (matrix C). by adjusting the dimensions.

I am not completely aware of the loop that you coded up so I am not sure how would i fix that, if you can explain the formatting part while writing the result matrix and the looping part, I Will appreciate that.

Thank you again for your generous help :)

4. gsal

963
please post again and include (quote) the precise lines you want me to explain...and I'll be happy to do so.

5. muzzy29

3
I basically dont understand the loops which i included below.I do not understand that how fortran reads these loops and i want to write flowchart so if you can tell me how these loops works and the formatting statement that you used in write statement. Also there is some issue with 2nd loop (multiplication loop) and what is matmul? did not I need to make matrix C entries 0 before I add a*b entries in matrix C ?

"Do i= 1,rowsB
write(*,"(100F7.2,\$)") (B(i,j),j=1,colsB)
End do"

"C = matmul(A(1:rowsA,1:colsA),B(1:rowsB,1:colsB))
Do i= 1,rowsA
WRITE(*,'(100F7.2)') (C(i,j),j=1,colsB)
End do"

"C = A(1:rowsA,1:colsA)+B(1:rowsB,1:colsB)
Do i= 1,rowsA
WRITE(*,'(100F7.2)') (C(i,j),j=1,colsB)
End do "

Thank you

6. gsal

963
The DO-loops that you refer to take advantage of a Fortran feature called "implicit do-loop"...this do-loop does not need "do"/"end-do" enclosing statements; instead, it simply happens within a single set of parenthesis.

The construct "(B(i,j),j=1,colsB)" leaves i alone, but varies j from 1 to colsB.

The line
reads an entire row of matrix B; more precisely, it reads the i-th row as the value for i remains constant but j is varied from 1 to colsB.

The loop
Do i= 1,rowsB
End do
would seem equivalent to a double loop like this:
Do i= 1,rowsB
Do j= 1,colsB
End do
End do
but will not work as desired due to the way you want to format your files.

When reading sequential file like you are doing, the READ() command retrieves the next line in the file and reads from it as many values as requested; and so, the explicit double-do-loop above would only read the first value of every row and soon run out of file...for it to work, you would need to write your input file with one value per line, which looses it visual correspondance to a nice looking matrix.

The WRITE statements have also been written to take advantage of the implicit do-loop feature...because we do not know how many values per line we need to write, I have simply put a multiplier of 100 for the F7.2 specification...it will only use as many as needed.

The multiplication and addition operations as shown are taking advantage of Fortran-90 intrinsic features...just like the scalar addition c=a+b (c=3+7, for example) works in any other language, the array addition C=A+B works in Fortran-90 (a-la Matlab); the same goes for the intrinsic function 'matmul' (matrix multiplication)..it just does the double-loop for you, no need to do anything else.

The reason why I had to specify the dimensions of the matrices during addition and multiplication is because the matrices were declared larger than necessary since we did not know in advnace the necessary dimensions...it is possible to delay the creation of such matrices with "allocatable" but I did not want to introduce that at this time.

Don't be shy about "asking" questions to the comipler...by that, I mean, experiment with simple programs that test one feature at a time, compile then, and see how it behaves.

7. YunzhE

2
hey i tried the code when
"C = A(1:rowsA,1:colsA)+B(1:rowsB,1:colsB)
Do i= 1,rowsA
WRITE(*,'(100F7.2)') (C(i,j),j=1,colsB)
End do "

it count correctly in order when matrix 3,3 add together

but i cant find the correct answer for the multiplication with this, same matrix 3,3 multiply together

"C = matmul(A(1:rowsA,1:colsA),B(1:rowsB,1:colsB))
Do i= 1,rowsA
WRITE(*,'(100F7.2)') (C(i,j),j=1,colsB)
End do"

8. gsal

963
can you show me what matrices you are using and what results you are expecting and what results you are getting?

did you figure how to enter input to the program? the first answer is either the character '+' or the character '*', right? Then, the file names of the matrices...

Please list the files you are using, etc.

9. YunzhE

2
example:
file 1
3,3
2,1,2
3,1,0
1,1,1
file 2
3,3
2,2,2
2,2,2
2,2,2
yes i can do the addition part
4,3,4
5,3,2
3,3,3

but when the multiplication part it shows the answer is
4,2,4
6,2,0
2,2,2