Get the indexes that would sort an array

Using numpy's argsort.

word_arr = np.array( ['lobated', 'demured', 'fristed', 'aproned', 'sheened', 'emulged',
    'bestrid', 'mourned', 'upended', 'slashed'])

idx_sorted=  np.argsort(word_arr)

idx_sorted
array([3, 6, 1, 5, 2, 0, 7, 4, 9, 8])

Let's look at the first and last three elements:

print "First three :", word_arr[ idx_sorted[:3] ]
First three : ['aproned' 'bestrid' 'demured']

print "Last three :", word_arr[ idx_sorted[-3:] ] 
Last three : ['sheened' 'slashed' 'upended']

Index of min / max element

Using numpy's argmin.

Min:

In [4]: np.argmin(word_arr)
3

print word_arr[np.argmin(word_arr)]
aproned

Max:

np.argmax(word_arr)
8

print word_arr[np.argmax(word_arr)]
upended