1

I'm struggling how to explain this in the question title so I'll try better here.

I have an array which contains 360 columns and 180 rows which represents each degree of latitude and longitude on earth. For clarity, my data looks like the following: 

longitudes = [0, 1, 2 ... 358, 359]
latitudes = [90, -89 ... -89, -90]
data = np.random.randint(5, (180, 360))

It looks like this, where longitude '0' is equal to 0E and '359' is equal to -1E.

As you can see, Africa is split at the start and end of the array and I would like to shift the array so that the first columns corresponds to -180E and the last columns corresponds to 180E, like the traditional world view (map) - this will make subsetting the array to 'cookie cut' Africa out.

How do I shift/transform my array so that the first and last column are adjacent?

(Extra: since you can think of my array as a cylinder, as geographically, the first and last rows are adjacent, but how do I change where the 'cylinder is cut' - the analogy I've been thinking of)

tda
  • 2,045
  • 1
  • 17
  • 42
  • Possible duplicate of [Efficient way to shift a list in python](https://stackoverflow.com/questions/2150108/efficient-way-to-shift-a-list-in-python) – Druta Ruslan Aug 06 '18 at 12:56
  • I've looked at the possible dupe and don't see this as a solution. I understand that it would work for my two lists, but not for my array. – tda Aug 06 '18 at 13:10

2 Answers2

1

Solution was to use numpy.roll().

I firstly found the index of the column corresponding to 180E using:

idx = (longitude.index(180))

Then, I 'rolled' the data based on this index to move 180E to the furthest left column using:

rolled_data = np.roll(data, idx, axis=1)

This successfully centred the 0E column to create my desired output.

tda
  • 2,045
  • 1
  • 17
  • 42
-1

please see this answer: Efficient way to shift a list in python

Seems to do the same as you want.

From the accepted answer:

Collections.deque

from collections import deque
items = deque([1, 2])
items.append(3) # deque == [1, 2, 3]
items.rotate(1) # The deque is now: [3, 1, 2]
items.rotate(-1) # Returns deque to original state: [1, 2, 3]
item = items.popleft() # deque == [2, 3]
VegardKT
  • 1,226
  • 10
  • 21