[Tutor] reading Fortran unformatted files with Python

[Gabriele Brambilla]

Hi,

sorry for the confusion I understood that the Real(8) I'm using correspond
to dtype float64 in Python.
With the second method it seems to work but I get a wrong number of
elements.

They should be grouped by 21 float64 but I don't find a number of data =
21*m where m is an integer number.

I think the problem is that Fortran write 4bytes at the beginning and at
the end of each file.
I found a way to remove the first 4 bytes but not the last.

Do you have any suggestion?

Thanks


On Mon, Jul 13, 2015 at 2:08 PM, Gabriele Brambilla <
gb.gabrielebrambilla at gmail.com> wrote:

> The second method seem to work.
> But how can I know which dtype in Python corresponds to REAL in fortran?
>
> Thanks
>
>
>
> On Mon, Jul 13, 2015 at 11:29 AM, Laura Creighton <lac at openend.se> wrote:
>
>> In a message of Mon, 13 Jul 2015 11:14:36 -0400, Gabriele Brambilla
>> writes:
>> >Hi,
>> >
>> >I have problems reading unformatted fortran output (binary) with python.
>> >
>> >I have a code in fortran where I write data on a file inside a cycle:
>> >
>> >write(11) x,y,z,BA01(i,j,k,1),BA01(i,j,k,2),1
>> >BB01(i,j,k,1),BB01(i,j,k,2),2           BE01(i,j,k,1),3
>> >EC01(i,j,k,1),EC01(i,j,k,2),4           ED01(i,j,k,1),ED01(i,j,k,2),5
>> >         EF01(i,j,k,1),6           rGH01(i,j,k,1),rGH01(i,j,k,2),7
>> >      rGI01(i,j,k,1),rGI01(i,j,k,2),8           rGJ01(i,j,k,1),1
>> >    rGL(i,j,k,2),rGM(i,j,k,2),rGN(i,j,k,2)
>> >
>> >How can I read this 21 real line by line in Python? How can I decode this
>> >unit specifier 11? All the numbers are defined as REAL.
>> >
>> >thanks
>> >
>> >Gabriele
>> >_______________________________________________
>> >Tutor maillist  -  Tutor at python.org
>> >To unsubscribe or change subscription options:
>> >https://mail.python.org/mailman/listinfo/tutor
>>
>> You want to use this package:
>> https://pypi.python.org/pypi/fortranfile
>>
>> You need to also load numpy for this to work.
>>
>> Some people cannot get this to work, and like this way of doing
>> things better:
>>
>> https://physics.ucf.edu/~dle/blog.php?id=1
>>
>> Again, needs numpy.  Write back if you need a numpy-free solution (i.e.
>> you are on Jython).
>>
>> Laura
>>
>
>