Sorting a list of different values by a specific field

Question

I've got a list of different types that I've read from a spreadsheet.

>>>data
[1.0, u'1', u'COMPLETED', u'TONGUE', 30.0, 1.0, '2014/01/28 09:33:16', 30.0, 210.0, u'1 CW', 0.0, 2.0, 210.0, u'H1490']
[1.0, u'1', u'COMPLETED', u'TONGUE', 30.0, 1.0, '04/02/2014 09:14:36', 30.0, 230.0, u'2 CCW', 0.0, 7.0, 210.0, u'H990']
.......

and I would like to be sorted by date, i.e. sorted by the 6th field (data[1][6],data[2][6],etc...). How could I indicate Python to sort this list by field number 6?

This is very difficult without a consistent date format. Is "04/02/2014" later or earlier than "05/01/2014"? Depends on whether the latter is May 1st 2014 or the 5th of January 2014. — Kevin, Aug 22 '14 at 15:21
you are going to need to turn your date `str` into something that will sort correctly, like a `datetime`. With inconsistent date format maybe look at `dateutil` parser. — cmd, Aug 22 '14 at 15:23
@cmd What if I have all dates in the same format? %Y/%m/%d %H:%M:%S ? Strings could also be sorted right? — Manolete, Aug 22 '14 at 15:25
YYYYMMDD style dates can be sorted lexicographically, so in that case you don't need to do any parsing at all. — Kevin, Aug 22 '14 at 15:28
yes `str` can be sorted lexicographically as Kevin said, but if its not in most significant element order and each element `0` padded to the same number of digits, it will not turn out how you want. — cmd, Aug 22 '14 at 15:32
@cmd Yes, I've noticed. I think I would have less control on sorting operations — Manolete, Aug 22 '14 at 15:34

score 3 · Accepted Answer · edited Aug 22 '14 at 15:37

3

You can use sorted and specify that you want to sort by 6th field.

You can use dateutil.parser to parse the date:

import dateutil.parser
sorted(data, key=lambda x: dateutil.parser.parse(x[6]))

Or if you want to sort data in-place and do less look-ups:

from dateutil.parser import parse
data.sort(key=lambda x: parse(x[6]))

edited Aug 22 '14 at 15:37

FredrikHedman

1,223
7
14

answered Aug 22 '14 at 15:16

jh314

27,144
16
62
82

@Ffisegydd then write your own function that returns comparable date for your string and use it as a key. – x3al Aug 22 '14 at 15:23
@Ffisegydd I've updated my answer based on your comments. Thanks! – jh314 Aug 22 '14 at 15:25
@jh314 That was useful. I could either use a parser or sort them lexicographically as @_Kevin said before – Manolete Aug 22 '14 at 15:32
1

@Manolete Glad to know it was useful! – jh314 Aug 22 '14 at 15:44

Padraic Cunningham · Answer 2 · 2014-08-22T15:56:01.763

Use sorted using a lambda as a key to sort and dateutil:

from dateutil import parser

sorted(data,key = lambda x:parser.parse(x[6]))

If you don't have dateutil installed:

pip install python-dateutil

If you want to sort the list in place and avoid creating a new list:

data.sort(key=lambda x:parser.parse(x[6]))

If the dates are in one of two different formats you can use your own function:

def date_parse(x):
    date = None
    for form in ["%Y/%m/%d %H:%M:%S","%d/%m/%Y %H:%M:%S"]:
        try:
            date = datetime.strptime(x,form)
        except ValueError:
            pass
    return date

sorted(data,key = lambda x: date_parse(x[6]))

Sorting a list of different values by a specific field

2 Answers2