I am trying to print two numpy arrays in relation to their order. I am trying to do this function without a for loop. I do not know the approach I could take to doing this. I essentially want to turn the vanilla Python version into numpy form.
a = np.arange(0, 15, 1)
b = np.arange(0, 150, 10)
Vanilla Python version
for i in range(len(a)):
print("when a is", a[i], "b is", b[i])
Expected output:
when a is 1 b is 10
when a is 2 b is 20
when a is 3 b is 30
...
I'm written this long winded answer because I suspect you don't quite understand the value of "no loop" when working with numpy.
So you have 2 arrays of the same length:
In [256]: a = np.arange(0,15,1)
...: b = np.arange(0,150,10)
In [257]: a
Out[257]: array([ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14])
In [258]: b
Out[258]:
array([ 0, 10, 20, 30, 40, 50, 60, 70, 80, 90, 100, 110, 120,
130, 140])
In [259]: a.shape
Out[259]: (15,)
In [260]: b.shape
Out[260]: (15,)
A straight forward print of each pair - this would even be faster if a and b were lists instead of arrays, but the real time limit is the print, not the iteration. (I'm using the newish f formating for convenience.)
In [261]: for i,j in zip(a,b):print(f'when a is {i}, b is {j}')
when a is 0, b is 0
when a is 1, b is 10
when a is 2, b is 20
....
when a is 13, b is 130
when a is 14, b is 140
We can make a 2d array from the two arrays:
In [262]: np.array((a,b))
Out[262]:
array([[ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12,
13, 14],
[ 0, 10, 20, 30, 40, 50, 60, 70, 80, 90, 100, 110, 120,
130, 140]])
And if you prefer a column display:
In [263]: np.array((a,b)).T
Out[263]:
array([[ 0, 0],
[ 1, 10],
[ 2, 20],
[ 3, 30],
...
[ 13, 130],
[ 14, 140]])
That makes a (15,2) array. There are several other ways of constructing such an array.
While the general layout matches your print, including all the extra text will take more work. numpy isn't optimized for string work; it's a numeric tool. If you want the fancy formatting stick with the python loop.
np.savetxt writes a csv file. It iterates on the rows, and for each does a formatted write like:
In [268]: for row in Out[263]: print('When a is %d, b is %d'%tuple(row))
When a is 0, b is 0
When a is 1, b is 10
When a is 2, b is 20
When a is 3, b is 30
In [270]: np.savetxt('test.txt', Out[263], fmt='when a is %d, b is %d')
In [271]: cat test.txt
when a is 0, b is 0
when a is 1, b is 10
when a is 2, b is 20
when a is 3, b is 30
Create a new array with a and b and then use np.apply_along_axis:
a = np.arange(0,15,1)
b = np.arange(0,150,10)
c = np.array([a, b])
np.apply_along_axis(lambda x: print(f'when a is {x[0]} b is {x[1]}'), 0, c)
Output:
when a is 0 b is 0
when a is 1 b is 10
when a is 2 b is 20
when a is 3 b is 30
when a is 4 b is 40
...
Btw, I don't think print is the correct usage of this function - it returns a series of None's. If you want a series, of strings, remove the print.
What's wrong with a for loop exactly?
If you really really don't want to use a for loop then
import numpy as np
a = np.arange(0, 15, 1)
b = np.arange(0, 150, 10)
text_1 = np.array(['what a is']*len(a))
text_2 = np.array(['b is']*len(a))
print(np.array([text_1, a, text_2, b]).transpose())
will work, but with lots of quotation marks everywhere. Note that it will only work if a and b are of the same length. You can use np.set_printoptions to format exactly how it looks.
