Sort a string in lexicographic order python

Question

I want to sort a string to a list in lexicographic order as

str='aAaBbcCdE'

to

['A','a','a','B','b','C','c','d','E']

but sorted() gives me this output:

['A','B','C','E','a','a','b','c','d']

How can I sort lexicographically?

Solution 1

Do not use lambda functions when there's builtin ones for the job. Also never use the cmp argument of sorted because it's deprecated:

sorted(s, key=str.lower)

or

sorted(s, key=str.upper)

But that may not keep 'A' and 'a' in order, so:

sorted(sorted(s), key=str.upper)

that will and, by the nature of sorted the operation will be very fast for almost sorted lists (the second sorted).

Solution 2

You could use a 2-tuple for the key:

text='aAaBbcCdE'
sorted(text, key=lambda x: (str.lower(x), x))
# ['A', 'a', 'a', 'B', 'b', 'C', 'c', 'd', 'E']

The first element in the tuple, str.lower(x) is the primary key (making a come before B), while x itself breaks ties (making A come before a).

Solution 3

cmp was the old way of doing this, now deprecated, but for posterity:

s='aAaBbcCdE'
sorted(s, lambda x,y: cmp(x.lower(), y.lower()) or cmp(x,y))

Solution 4

data = input() data=list(data) data.sort()

Now the variable "data" will have lexicographically sorte`d input provided.

View more solutions

Comments

[JBernardo]

that's deprecated use of sorted.

[Russia Must Remove Putin]

@KarolyHorvath to be completely obvious, fix your answer and he won't have to.

[Karoly Horvath]

@AaronHall: which will make the comments misleading. no, thanks.

[Antti Haapala -- Слава Україні]

Actually, let's keep this answer here :D

[Abhijeet]

sorted(sorted(s), key=str.upper) stands-out.

[Scott Gartner]

This solved it for me, but I don't understand why returning x doesn't result in correct sorting, but returning (lower(x), x) does. Can anyone explain?

[Raffi]

@ScottGartner returning a 2-uple (lower(x), x) basically uses the already implemented lexicographical ordering of the tuples: all elements that are equal on the index 0 of the tuple will be sorted by the element of the index 1 of the tuple, all elements that are different on index 0 will be sorted as usual.

[Raffi]

This is the best answer by far!