在 Python 中模拟像 Sqlite3 这样的元组排序

Question

I'm trying to sort a list of tuples where there are certain tuple indices that contain None. In this case, the tuples with None in a specific index location need to be given priority over the other non-None values in the list of tuples.

So for example, if I have a list which looks like this:

[
(3.4, 43, None),
(5345.6, 42, None),
(43.24, 25, None),
(323.4, 433, None),
(5.6, 42, None),
(4.4, 235, None),
(None, None, 'Hello'),
(5.6, None, None),
(None, 235, None),
]

And I want my final list to look like (basically doing a "SELECT * FROM table ORDER BY column1,column3;")

[
(None, 235, None),
(None, None, 'Hello'),
(3.4, 43, None),
(4.4, 235, None),
(5.6, 42, None),
(5.6, None, None),
(43.24, 25, None),
(323.4, 433, None),
(5345.6, 42, None),
] 

As of right now, I'm currently using

sortedSelectedResult = sorted(selectedResult,key=itemgetter(*indexSortOrderList))

where indexSortOrderList is a list which has a list of the index order by which the list of tuples need to be sorted. But in this method, Nones are randomly placed within the list. My current result looks like this, where indexSortOrder = [0,2] (ie) sort by first column, then the third column:

[(3.4, 43, None),
 (5345.6, 42, None),
 (43.24, 25, None),
 (323.4, 433, None),
 (5.6, 42, None),
 (4.4, 235, None),
 (None, None, 'Hello'),
 (5.6, None, None),
 (None, 235, None)]

I tried looking into operator.itemgetter and modifying the function to apply to my use case, but I'm having trouble with that as well..

Answer 1

You can add a custom key considering None as 0:

sorted(lst, key=lambda x: (isinstance(x[0], str), x[0] or 0,  isinstance(x[2], str) or x[2] or 0))

Code :

lst = [
(3.4, 43, None),
(5345.6, 42, None),
(43.24, 25, None),
(323.4, 433, None),
(5.6, 42, None),
(4.4, 235, None),
(None, None, 'hello'),
(5.6, None, None),
(None, 235, None),
]

print(sorted(lst, key=lambda x: (isinstance(x[0], str), x[0] or 0,  isinstance(x[2], str) or x[2] or 0)))

# [(None, 235, None),
#  (None, None, 'hello'),
#  (3.4, 43, None),
#  (4.4, 235, None),
#  (5.6, 42, None),
#  (5.6, None, None),
#  (43.24, 25, None),
#  (323.4, 433, None),
#  (5345.6, 42, None)]