scipy.io._fortran:FortranFile.read_record

Notes

If the record contains a multidimensional array, you can specify the size in the dtype. For example

INTEGER var(5,4)

can be read with

read_record('(4,5)i4').T

Note that this function does not assume the file data is in Fortran column major order, so you need to (i) swap the order of dimensions when reading and (ii) transpose the resulting array.

Alternatively, you can read the data as a 1-D array and handle the ordering yourself. For example

read_record('i4').reshape(5, 4, order='F')

For records that contain several variables or mixed types (as opposed to single scalar or array types), give them as separate arguments

double precision :: a
integer :: b
write(1) a, b

record = f.read_record('<f4', '<i4')
a = record[0]  # first number
b = record[1]  # second number

and if any of the variables are arrays, the shape can be specified as the third item in the relevant dtype

double precision :: a
integer :: b(3,4)
write(1) a, b

record = f.read_record('<f4', np.dtype(('<i4', (4, 3))))
a = record[0]
b = record[1].T

NumPy also supports a short syntax for this kind of type

record = f.read_record('<f4', '(3,3)<i4')