如何在清单中计算。 列表混合在Python中

Question

My list is L = [[1,1], [2,2], [1,1], 1]

I want to count distinct elements in L in numpy. How??? Answer should be 3 after counting.

Answer 1

Here's one way. The trick is to convert iterables to hashable types (eg tuples). You can then convert to a set and calculate len .

What this won't do, of course, is differentiate between list and tuple elements in your list. But this may not be applicable in your use case.

from collections import Iterable

L = [[1,1], [2,2], [1,1], 1]

len(set(tuple(i) if isinstance(i, Iterable) else i for i in L))

# 3

Answer 2

Here is a way to find unqiues if items are at least sortable within their class. As Python3 mostly does not allow across type comparisons we sort by type first and then by value:

>>> from operator import itemgetter as item_, attrgetter as attr_
>>> from itertools import groupby
>>> 
>>> by_type = groupby(sorted(zip(map(attr_('__qualname__'), map(type, L)), L)), item_(0))
>>> by_type = {k: list(map(item_(0), groupby(map(item_(1), g)))) for k, g in by_type}
>>> by_type
{'int': [1], 'list': [[1, 1], [2, 2]]}
# total number of uniques
>>> sum(map(len, by_type.values()))
3

For those who don't like map here is a translation using comprehensions:

>>> by_type = groupby(sorted((type(i).__qualname__, i) for i in L), item_(0))
>>> by_type = {k: [gk for gk, gg in groupby(gval for gtp, gval in g)] for k, g in by_type}
>>> by_type
{'int': [1], 'list': [[1, 1], [2, 2]]}

Answer 3

first stringify each element of list

then find set of list of stringified elements

then get length of set

print(len(set([str(i) for i in L])))

>>> 3

also... your OP requested "in numpy" which presents impossibility as you can't have multiple data types in an array (lists and ints).

eta:

from comments, nice catch Paul Panzer; this should patch the issue while remaining concise one liner:

L1 = [[0], np.array([0])]
print (len(set([(str(i)+str(type(i))) for i in L1])))
>>> 2

L2 = [[0, 1], np.array([0, 1])]
print (len(set([(str(i)+str(type(i))) for i in L2])))
>>> 2