CODEWITHSUNDEEP
formatfortran
zon kabouter
zon kabouter

Reputation: 13

Format in the READ statement and real number rounding

I am trying to use the format statement in FORTRAN 77 in order to define to precision with which I want to read a number.

Here are the lines of my code: 

double precision x,y,z   
open(u,FILE='points.dat',STATUS='OLD')    
do 10 i=1,n    
read(u,100) x,y,z    
write(*,*),'check', x,y,z    
enddo   
100 format(3(2x,D14.2))     
close(u)

I tried to read these numbers: 1.0542095e-03 -3.4176819e-02 -1.3796906e-01 with different format and the results are:

format(3(2x,E14.2)):  check  0.0010542095 -0.034176819 -0.13796906    
format(3(2x,E14.7)):  check  0.0010542095 -0.034176819 -0.13796906     
format(3(2x,D14.7)):  check  0.0010542095 -0.034176819 -0.13796906

I would have expected the first line to give 0.001000000 and the second line to give 0.001054209 but it seems that format statement has no effect in the reading or maybe I did not use it properly. Anyone could explain me the effect of the format in the read function and how to use it properly?

Upvotes: 1

Views: 423

Answers (1)

Vladimir F Героям слава
Vladimir F Героям слава

Reputation: 60008

Don't use the .2 part of the edit descriptor when reading. Never. It does very bad and confusing things. Always put there .0.

What it actually does is that it does nothing at all when the number being read contains a decimal point. So far so it is good. But if the number does not contain a decimal point it will add one according to your number.

So if you read

'1'

using

read '(E10.2)', x

x will become 0.01.

If you read '1.0' using the same, it becomes 1.0.

It is so confusing that I highly suggest to use .0 in the real edit descriptors under all circumstances.

For rounding your numbers you could, for example, do

read '(E10.0)', x
x = aint(x*100)/100

Upvotes: 5

Related Questions