Read a file consisting two columns of pure(?) double into a complex NumPy array - numpy

Question

Almost the same questions were asked by the following:
Numpy read complex numbers from text file
Writing and reading complex numbers using numpy.savetxt and numpy.loadtxt
loading complex numbers with numpy.loadtxt
Reading complex data into numpy array
However, the above involved slightly different input format, e.g. parentheses, than the file content herein.
Consider a file named example containing two columns of pure(?) double:
0.8355544313622164 0
1.199174279986189 0
1.417275292218002 0
I am able to generate a numpy array of np.complex64 by doing the following:
data = np.loadtxt("./example", dtype=np.float64, delimiter='\t')
complexData = data.T[0] + 1j*data.T[1]
Printing complexData now gives:
[ 0.83555443+0.j 1.19917428+0.j 1.41727529+0.j ... ]
Is it possible to reduce the above approach into a neater one?
For example, changing data type to np.complex64 raises TypeError:
data = np.loadtxt("./example", dtype=np.complex64, delimiter='\t')

Answer 1

Instead of converting the real array to complex with
complexData = data.T[0] + 1j*data.T[1]
you can create a complex view of the data:
complexData = data.view(np.complex128)
Then data and complexData share the underlying array of floating point numbers, but complexData interprets those values as complex numbers.
complexData will be an array with shape (n, 1). To get rid of the extraneous second dimension, you can use
complexData = data.view(np.complex128)[:, 0]
You could do the conversion immediately upon reading the data. For example, my sample file called "real.txt" is
0.8355544313622164 0
1.199174279986189 0
1.417275292218002 0
3.141592653589793 -1
and it is not tab-delimited, so I'll use the default delimiter. To read the data as complex:
In [18]: z = np.loadtxt('real.txt').view(np.complex128)[:, 0]
In [19]: z
Out[19]: array([0.83555443+0.j, 1.19917428+0.j, 1.41727529+0.j, 3.14159265-1.j])