Can NumPy directly load data into a Fortran ordered array?

Question

I often work with data stored in text files with [x, y, z] like format. When loading the data into a NumPy array, it is convenient to maintain the ordering from the text file, where each column is a different element, x, y, or z. The downside to this is with the C ordering NumPy uses as the default, operating on all the x-values mean accessing non-contiguous memory blocks.

To load the data into a Fortran ordered array, I can use

data = numpy.asfortranarray(numpy.loadtxt('data.txt'))

but is there a way to do that in one line? Looking at the loadtxt documentation, this method does not seem to provide that functionality. Is there an another load function that does?

Do you realize that loadtxt can allocate memory for a row but not for a column beforehand? So the best it could do is read the array in C order and then convert it to F order. — Stop harming Monica
– Stop harming Monica, Commented Dec 9, 2016 at 18:34

hpaulj · Accepted Answer · 2016-12-09 18:35:05Z

4

Both loadtxt and genfromtxt read the file line by line and collect results in a list of lists (or list of tuples). At the end they convert this to an array. Roughly:

rows = []
for line in f.readline():
    values = [float(i) for i  in line.split(delimiter)]
    rows.append(values)
data = np.array(rows)

You could write your own reader! Or just accept this extra bit of post processing. It doesn't look hard or expensive.

answered Dec 9, 2016 at 18:35

hpaulj

233k14 gold badges260 silver badges392 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Steven C. Howell Over a year ago

That makes sense. Essentially, because the file is stored in rows, to get a column-major array, Fortran ordering, the file has to be read in using row-major ordering, C ordering, then transformed.

Collectives™ on Stack Overflow

Can NumPy directly load data into a Fortran ordered array?

1 Answer 1

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related