Le Fri, 9 Jan 2009 08:10:27 -0800 (PST),
culpritNr1 <ig2ar-s...@yahoo.co.uk> a écrit :
>
> Hello Denis and All,
>
> Your solution does show elegance. To remind people, it's this one:
>
> lol = [[1,2,3],[4,5,6],[7,8,9]]
> new_lol = [[a,b**3,c] for [a,b,c] in lol]
> print lol
> print new_lol
> ==>
> [[1, 2, 3], [4, 5, 6], [7, 8, 9]]
> [[1, 8, 3], [4, 125, 6], [7, 512, 9]]
>
> Now, as I said in my original post, I can't assume that all inner-lists will
> have the same number of elements. Common data looks like this
>
> LoL = [['chrX', '160944034', '160944035', 'gnfX.145.788', '63.60'],
> ['chrX', '161109992', '161109993', 'rs13484104', '63.60'],
> ['chrX', '161414112', '161414113', 'undetermined'],
> ['chrX', '161544071', '161544072', 'rs13484106', '63.60',
> 'RNA-BP', 'PLoS Biol 6(10): e255'],
> ['chrX', '162030736', '162030737', 'gnfX.146.867', '67.05'],
> ['chrX', '164171913', '164171914', 'gnfX.148.995', '70.45']]
>
> Do you still think that this problem can be elegantly tackled by list
> comprehensions?
The straightforward method is indeed to to loop with neccsary checks/trials
embedded in loop. Now,
there imo is more elegant method which is to define a custom int that
integrates such tests. And
then to call this custom conversion instead of built-in int. This lets open the
possibility of a
list comp expression. Below an example:
def possible_cube(val):
try:
return val ** 3
except:
return val
lol = [[1,2,'x'],[4,'y',6],['z',8,9]]
new_lol = [[possible_cube(val) for val in l] for l in lol]
print lol
print new_lol
==>
[[1, 2, 'x'], [4, 'y', 6], ['z', 8, 9]]
[[1, 8, 'x'], [64, 'y', 216], ['z', 512, 729]]
This may be longer because there is a custom function call in each loop. Now,
it's a question of
taste and priority.
denis
------
la vida e estranya
_______________________________________________
Tutor maillist - Tutor@python.org
http://mail.python.org/mailman/listinfo/tutor
