"only integer scalar arrays can be converted to a scalar index"

Question

I am generating a numpy array with 1000000 random numbers between 1 and 6 and i would like to calculate the mean of the first 10, 100, 1000, ... I also want to plot the means on logarithmic scale. I mustn't use anything but Python with numpy and matplotlib. Why do i get this error? What have i done wrong?

This is my code:

throws=numpy.random.randint(1,7,(1000000))
print(throws[1:10])

x=np.logspace(1,6,6)
plt.plot(x, int(mean(throws[1:x])))
plt.semilogx()

Sorry for my bad English and the german variable names...

I would avoid using an umlaut in a variable name. Might cause trouble with the interpreter. — Stef, Nov 28 '20 at 15:18
Apparently the interpreter is complaining that you used something as an index that can't be used as an index. I'm guessing this comes from `wurfe[1:x]` with `x=np.logspace(1,6,6)`. — Stef, Nov 28 '20 at 15:20
i'm getting TypeError: only integer scalar arrays can be converted to a scalar index — AufsMaulwurf, Nov 28 '20 at 16:38

score 1 · Accepted Answer · answered Nov 28 '20 at 16:53

You're almost there! You just have to slice the würfe (dice throws) array and then apply mean to it, this is easiest with

würfe=numpy.random.randint(1,7,(1000000))
print(würfe[1:10])

x=np.logspace(1,6,6)
y=[np.mean(würfe[:int(x_)]) for x_ in x]  # <--- just add this line
plt.plot(x, y)
plt.semilogx()
plt.show()

x here is [10, 100, 1000, 10000, 100000, 1000000] and würfe[:int(x_)] converts x from float to int and uses it to slice the original array into the parts you want to take the mean of. The mean is then taken with a Python list comprehension.

Thank you sooo much, you made my day – AufsMaulwurf Nov 28 '20 at 19:11 — AufsMaulwurf, Nov 28 '20 at 19:11

Kahn · Answer 2 · 2020-11-28T19:00:05.830

0

Try Run:

import numpy
x=numpy.logspace(1,6,6)
print(x.dtype)

Shows

float64

So würfe[1:x] is using float64 array as index,certainly it can't be right.

Use

x=numpy.logspace(1,6,6).astype(int)

And also würfe[1:x] requires x to be a int but not an array.

edited Nov 28 '20 at 19:00

answered Nov 28 '20 at 15:23

Kahn

186
7

Thanks for your help, but it doesn't change anything... Do you have another advice/ idea where the problem is caused? – AufsMaulwurf Nov 28 '20 at 15:38
The problem occurs at the `würfe[1:x]`, because x is an array not a number (scalar). Try using each member of x individually: `[würfe[:x_] for x_ in x]` – xjcl Nov 28 '20 at 16:54

"only integer scalar arrays can be converted to a scalar index"

2 Answers2