saving a set of arrays from python output to csv file

Question

for tstep in arange (1,500,1):
    Intensity=np.sum(Itarray*detarray)
    print tstep*t*10**9, "\t", Intensity.real

For the above program how do I save the two arrays tstep*t*10**9 and Intensity.real into a csv file as two colums with a tab so that I get all the values as the loop goes from 1 to 500

Answer 1

Instead of using a for-loop, use

tsteps = np.arange(1,500,1, dtype='int64')*t*10**9

to build a NumPy array. Note carefully that NumPy arrays have a dtype. The dtype determines the range of numbers representable by the elements in the array. For example, with dtype int64, an array can represent all integers between

In [35]: np.iinfo('int64').min, np.iinfo('int64').max
Out[35]: (-9223372036854775808L, 9223372036854775807L)

For the sake of speed, NumPy does not check for arithmetic overflow. If 499*t*10**9 falls outside this range, then the array will contain wrong numbers. So the onus is on you to choose the right dtype to avoid arithmetic overflow.

Note also that if t is a float, then np.arange(1,500,1, dtype='int64')*t will get upcasted to dtype float64, whose range of representable values lie between

In [34]: np.finfo('float64').min, np.finfo('float64').max
Out[34]: (-1.7976931348623157e+308, 1.7976931348623157e+308)

---

np.sum(Itarray*detarray) does not depend on tstep, so it can be pulled outside the for-loop. Or, since we are not using a for-loop, it just needs to be computed once.

---

Finally, form a 2D array (using np.column_stack), and save it to a file with np.savetxt:

import numpy as np
tsteps = np.arange(1,500,1, dtype='int64')*t*10**9
Intensity = np.sum(Itarray*detarray)
np.savetxt(filename, np.column_stack(tsteps, Intensity.real), delimiter='\t')

Answer 2

As far as I understand, it's as simple as;

myfile = open("test.out", "w")

for tstep in arange (1,500,1):
    Intensity=np.sum(Itarray*detarray)
    myfile.write(str(tstep*t*10**9) + '\t' + str(Intensity.real) + '\n')

myfile.close()

Answer 3

You can use the csv module. Your question is a little bit unclear. I am assuming that Intensity is an array with 500 entries.

Intensity = np.sum(Itarray*detarray)
w = csv.writer(open('output.csv','w'), delimiter='\t')
w.writerows([('%e' % ((i+1)*t*10**9),'%f' % b) for i,b in enumerate(Intensity.real)])