CODEWITHSUNDEEP
pythonarraysnumpy
Arpit Chinmay
Arpit Chinmay

Reputation: 333

NumPy array row wise and column wise slicing syntax

Why does NumPy allow array[row_index, ] but array[, col_index] is not valid and gives Syntax Error. e.g. if I want to traverse the array row wise NumPy.array[row_index, :] and NumPy.array[row_index, ] both give the same answer where as only NumPy.array[:, col_index] produces the result in the latter case. Is there any reason behind this that I'm missing?

Upvotes: 1

Views: 194

Answers (1)

hpaulj
hpaulj

Reputation: 231355

arr[idx,] is actually short for arr[(idx,)], passing a tuple to the __getitem__ method. In python a comma creates a tuple (in most circumstances). (1) is just 1, (1,) is a one element tuple, as is 1,.

arr[,idx] is gives a syntax error. That's the interpreter complaining, not numpy.

arr[3], arr[3,] and arr[3,:] are all the same for a 2d array. Trailing : are added as needed. Leading : have to be explicit.

Upvotes: 2

Related Questions