Set none if numpy array index does not exist

Question

I have a function within a Python (2.7) class that should retrieve the values of the 'cells' around it in a 2 dimensional numpy array. If the index is out of range, I would the value should be set as None.

I'm struggling to find a way to do this without writing 8 try/catch statements, or without using multiple if x else None statements as is in my code below. While they would both work, they don't seem very well structured, and I'm thinking there must be a simpler way to do this - I'm probably caught thinking about this in entirely the wrong way. Any help would be much appreciated.

# This will return a dictionary with the values of the surrounding points
def get_next_values(self, column, row):
    if not (column < self.COLUMNS and row < self.ROWS):
        print "Invalid row/column."
        return False

    nextHorizIsIndex = True if column < self.COLUMNS - 2 else False
    nextVertIsIndex = True if row < self.ROWS - 2 else False

    n = self.board[column, row-1] if column > 0 else None
    ne = self.board[column+1, row-1] if nextHorizIsIndex else None
    e = self.board[column+1, row] if nextHorizIsIndex else None
    se = self.board[column+1, row+1] if nextHorizIsIndex and nextVertIsIndex else None
    s = self.board[column, row+1] if nextVertIsIndex else None
    sw = self.board[column-1, row+1] if nextVertIsIndex else None
    w = self.board[column-1, row] if row > 0 else None
    nw = self.board[column-1, row-1] if 0 not in [row, column] else None

    # debug
    print n, ne, e, se, s, sw, w, nw

Watch out for the negative index wrap-around ... !

wim
– wim

2013-02-04 02:30:19 +00:00
Commented Feb 4, 2013 at 2:30 — wim
– wim, Commented Feb 4, 2013 at 2:30
It hadn't occurred to me, thanks!

Adam
– Adam

2013-02-04 02:46:56 +00:00
Commented Feb 4, 2013 at 2:46 — Adam
– Adam, Commented Feb 4, 2013 at 2:46

unutbu · Accepted Answer · 2013-02-04 02:51:05Z

6

Here is a standard trick: Create your board with an edge padded with the value None. Then you can access any inner 3x3 square and fill in the appropriate values with

nw, n, ne, w, _, e, sw, s, se = (self.board[column-1:column+2, row-1:row+2]).ravel()

For example,

import numpy as np

board = np.empty((10,10), dtype = 'object')
board[:,:] = None
board[1:9, 1:9] = np.arange(64).reshape(8,8)
print(board)
# [[None None None None None None None None None None]
#  [None 0 1 2 3 4 5 6 7 None]
#  [None 8 9 10 11 12 13 14 15 None]
#  [None 16 17 18 19 20 21 22 23 None]
#  [None 24 25 26 27 28 29 30 31 None]
#  [None 32 33 34 35 36 37 38 39 None]
#  [None 40 41 42 43 44 45 46 47 None]
#  [None 48 49 50 51 52 53 54 55 None]
#  [None 56 57 58 59 60 61 62 63 None]
#  [None None None None None None None None None None]]

column = 1
row = 1
nw, n, ne, w, _, e, sw, s, se = (board[column-1:column+2, row-1:row+2]).ravel()
print(nw, n, ne, w, _, e, sw, s, se)
# (None, None, None, None, 0, 1, None, 8, 9)

Note that

when you define the board this way, the first non-None index is now 1, not 0.
I think it is more typical to think of the first index as the row, and the second index as the column, because when you print(board) that is the way the values are formatted. So perhaps you want board[row-1:row+2, column-1:column+2] instead. Of course, you could define your own print_board function, and then be free to use whatever convention you like.

edited Feb 4, 2013 at 2:51

answered Feb 4, 2013 at 2:26

unutbu

886k197 gold badges1.9k silver badges1.7k bronze badges

Sign up to request clarification or add additional context in comments.

4 Comments

wim Over a year ago

A good trick, but they may have trouble to use None in an array with numeric dtype. May have to use dtype object, which isn't very "numpythonic", or use nan instead of None.

unutbu Over a year ago

True, nan might be a better choice if the dtype is float.

Adam Over a year ago

Thank you! This is really helpful - I wasn't aware of the board[:,:] = None syntax either. My dtype is int, so I'll give it a try padded with nan.

unutbu Over a year ago

nan is compatible with float dtypes, but not int. You'll need to choose an integer value for int. Would a negative number like -1 be okay?

Collectives™ on Stack Overflow

Set none if numpy array index does not exist

1 Answer 1

4 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

4 Comments

Your Answer

Sign up or log in

Post as a guest

Related