np.array - too many values to unpack

Question

I am writing a script to save a part of an image ("submatrix" in the code, reported below) as a tif file. When I run the code, I got the following error:

height, width = np.array(submatrix.shape, dtype = float) / dpi
ValueError: too many values to unpack

If instead of the submatrix I use a random matrix like np.random.random((10, 10)), everything runs fine. Do you spot what I am doing wrong?

Using

import matplotlib.pyplot as plt
import math
import numpy as np

Here's the code section where the error is:

submatrix = im[x_min:x_max, y_min:y_max]
dpi = size_box
height, width = np.array(submatrix.shape, dtype = float) / dpi

Answer 1

The error message tells you what the problem is. There are too many values to unpack. Clearly submatrix.shape has length greater than 2.

I cannot tell why that is so, since I don't know what im is. But take a look at the following output from the interactive prompt:

>>> height, width = np.array([1,2], dtype = float)
>>> height, width = np.array([1,2,3], dtype = float)
Traceback (most recent call last):
  File "", line 1, in 
ValueError: too many values to unpack

Something like that is causing your problem with submatrix fulfilling the role that [1,2,3] does above.

If you strip away all the numpy, this is just a standard sequence unpacking error. The simplest example of that being:

>>> x, y = (1, 2, 3)
Traceback (most recent call last):
  File "", line 1, in 
ValueError: too many values to unpack

When you perform sequence unpacking, the sequences on left and right hand side of the assignment operator must have the same length.

Answer 2

Any time I run into this, I set the output to a single var and the inspect that var to see what's inside.

>>> blah = np.array(submatrix.shape, dtype = float) / dpi
>>> dir(blah)

98% of the time, I simply miscounted the array size. (i.e. X, Y, and Z) the other 2% is usually due to variable length arrays (which really suck).

Early on, I would never use sequence unpack on the same line as the declaration, just to make debugging easier. I then set a break point on the height, width line to see why it caused a problem.

blah = np.array(submatrix.shape, dtype = float) / dpi
height, width = blah