scipy.io.FortranFile#

class scipy.io.FortranFile(filename, mode='r', header_dtype=<class 'numpy.uint32'>)[source]#

A file object for unformatted sequential files from Fortran code.

Parameters:
filenamefile or str

Open file object or filename.

mode{‘r’, ‘w’}, optional

Read-write mode, default is ‘r’.

header_dtypedtype, optional

Data type of the header. Size and endiness must match the input/output file.

Notes

These files are broken up into records of unspecified types. The size of each record is given at the start (although the size of this header is not standard) and the data is written onto disk without any formatting. Fortran compilers supporting the BACKSPACE statement will write a second copy of the size to facilitate backwards seeking.

This class only supports files written with both sizes for the record. It also does not support the subrecords used in Intel and gfortran compilers for records which are greater than 2GB with a 4-byte header.

An example of an unformatted sequential file in Fortran would be written as:

OPEN(1, FILE=myfilename, FORM='unformatted')

WRITE(1) myvariable

Since this is a non-standard file format, whose contents depend on the compiler and the endianness of the machine, caution is advised. Files from gfortran 4.8.0 and gfortran 4.1.2 on x86_64 are known to work.

Consider using Fortran direct-access files or files from the newer Stream I/O, which can be easily read by numpy.fromfile.

Examples

To create an unformatted sequential Fortran file:

>>> from scipy.io import FortranFile
>>> import numpy as np
>>> f = FortranFile('test.unf', 'w')
>>> f.write_record(np.array([1,2,3,4,5], dtype=np.int32))
>>> f.write_record(np.linspace(0,1,20).reshape((5,4)).T)
>>> f.close()

To read this file:

>>> f = FortranFile('test.unf', 'r')
>>> print(f.read_ints(np.int32))
[1 2 3 4 5]
>>> print(f.read_reals(float).reshape((5,4), order="F"))
[[0.         0.05263158 0.10526316 0.15789474]
 [0.21052632 0.26315789 0.31578947 0.36842105]
 [0.42105263 0.47368421 0.52631579 0.57894737]
 [0.63157895 0.68421053 0.73684211 0.78947368]
 [0.84210526 0.89473684 0.94736842 1.        ]]
>>> f.close()

Or, in Fortran:

integer :: a(5), i
double precision :: b(5,4)
open(1, file='test.unf', form='unformatted')
read(1) a
read(1) b
close(1)
write(*,*) a
do i = 1, 5
    write(*,*) b(i,:)
end do

Methods

close()

Closes the file.

read_ints([dtype])

Reads a record of a given type from the file, defaulting to an integer type (INTEGER*4 in Fortran).

read_reals([dtype])

Reads a record of a given type from the file, defaulting to a floating point number (real*8 in Fortran).

read_record(*dtypes, **kwargs)

Reads a record of a given type from the file.

write_record(*items)

Write a record (including sizes) to the file.