Unifying different 2d array layouts in python

Question

I'm working on an application that gets a list of x and y coordinates from various data sources. Depending on which data sources I need to go to (which is outside my control), there are multiple ways the data is returned.

an array of x and y values - [[x1,x2,x3...],[y1, y2, y3...]]
an array of x,y pairs - [[x1,y1],[x2,y2], ...]
a dictionary of x:y values - {x1:x2, x2:y2, x3:y3, ...}
a numpy array of x and y values - array([[x1,x2,x3...],[y1, y2, y3...]])
a numpy of x,y pairs - array([[x1,y1],[x2,y2], ...])

Using isinstance and len/shape I know I can solve this problem with a switch statement. I was wondering if there was a better way to convert these to a consistent format, or to read these in a consistent way.

AChampion · Accepted Answer · 2015-04-08 16:36:36Z

1

There is no way to treat these structures identically (they aren't all ducks).
However, it's fairly easy to convert into a consistent format, make your pick and convert into that format, e.g.

zip(*[[x1,x2,x3...],[y1,y2,y3...]]) == [(x1,y1),(x2,y2),(x3,y3)...]
zip(*[[x1,y1],[x2,y2],[x3,y3]...]) == [(x1,x2,x3...),(y1,y2,y3...)]
list({x1:x2, x2:y2, x3:y3, ...}.items()) == [(x1,y1),(x2,y2),(x3,y3)...]
dict([(x1,y1),(x2,y2),(x3,y3)...]) = {x1:x2, x2:y2, x3:y3, ...}

Numpy arrays can largely be treated the same as Python lists.

edited Apr 8, 2015 at 16:36

answered Apr 8, 2015 at 2:15

AChampion

30.5k4 gold badges63 silver badges79 bronze badges

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

Comments

hpaulj · Accepted Answer · 2015-04-08 07:14:17Z

Adding to achampion's conversions:

You can construct a dictionary from a list of pairs:

dict([[x1,y1],[x2,y2],...]) == {x1:x2, x2:y2, x3:y3, ...}

np.asarray can convert lists to arrays, and happily passes an array through unchanged. numpy functions often pass their inputs through asarray (or a variant) to insure that the inputs are arrays of known dimension.

And array transpose is the equivalent of that list zip(*...).

np.asarray([[[x1,x2,x3...],[y1,y2,y3...]]) == np.asarray([(x1,y1),(x2,y2),(x3,y3)...]).T

I don't think you can go from dictionary to array without an intermediary list. But a (n,2) array works just as well as a list of pairs as input to dict().

So it is easy to convert one layout to another, but I don't think there is one function that does it for all of your layouts.

Is any one of these layouts more convenient for further processing?

Pseudo code might be:

if x is a dictioary:
   x = convert_to_list(x)
x = asarray(x)
if x.shape[0]==2 and x.shape[1]>2:
    x = x.T
# [[1,2],[3,4]] is ambiguous

Collectives™ on Stack Overflow

Unifying different 2d array layouts in python

2 Answers 2

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related