how can I sort a bunch of lists over multiple fields?

Steven Bethard steven.bethard at gmail.com
Thu Apr 28 14:43:19 EDT 2005


googleboy wrote:
> firstly,  I am trying hard to figure out how to create a new file with
> the list rather than print to standard out.  I haev done this:
> 
>     for book in books:
>         print book # just to be sure it works as I expect
>         sort1 = open(r'D:\path to\sort1.csv', 'w+')
>         print >> sort1, book
>         sort1.close()
> 
> and this creates the file as I expect, however it creates it populated
> with only the information of the final book in the sorted list.

You're reopening the file on each iteration of the loop.  I think you 
want to open it only once, before the loop, e.g.

sort1_file = open(r'D:\path to\sort1.csv', 'w+')
for book in books:
     sort1_file.write('%s\n' % book) # same as "print >> sort1, book"
sort1_file.close()

Note that the opening and closing of the file is outside the loop.

> Secondly,  I am wondering how I can get a search algorithm that will
> search by multiple fields here,  so that I can (as one example) sort
> the books out by author and then date,  to present a list of the book
> grouped by authors and having each group presented in a chronological
> order,   or by author and title, grouping all the books up into authors
> presenting each group alphabetically by title.  Or by publisher and
> date,  or by publisher and code....
> 
> I have tried things like
> 
> books.sort(key = operator.attrgetter("author"), key =
> operator.attrgetter("title") and
> books.sort(key = operator.attrgetter("author", "title")
> 
> but they both give errors.

The problem is that operator.attrgetter only accepts a single attribute. 
  Basically, attrgetter looks something like:

     def attrgetter(attr_name):
         def func(obj):
             return getattr(obj, attr_name)
         return func

So attrgetter can't really solve your problem.  However, you can create 
a similar function that should do the job.  Something like (untested):

     def get_key(*attr_names):
         def key(book):
             return [getattr(book, name) for name in attr_names)]
         return key

Then you should be able to do something like:

     books.sort(key=get_key("author", "title"))

The trick is that the inner function, 'key', looks up a sequence of 
attributes on the book object, instead of just a single attribute like 
attrgetter does.

HTH,

STeVe



More information about the Python-list mailing list