[2.5.1.1/dictionary] Change sorting order?

[Jean-Michel Pichavant]

Jean-Michel Pichavant wrote:
> Gilles Ganault wrote:
>> On Fri, 22 Jan 2010 14:09:43 +0100, Jean-Michel Pichavant
>> <jeanmichel at sequans.com> wrote:
>>  
>>> Sorry, the code I provided produce this output:
>>>
>>> ['1a', 'a', 'ac', 'av', 'b', 'c']
>>> ['a', 'ac', 'av', 'b', 'c', '1a']
>>> ['b', 'c', '1a', 'a', 'ac', 'av']
>>> ['c', '1a', 'a', 'ac', 'av', 'b']
>>> ['1a', 'a', 'ac', 'av', 'b', 'c']
>>>
>>> which is actually what you are searching for. I just messed up with 
>>> my ipython shell history :o)
>>>     
>>
>> Thanks for the help. I'm a Python newbie, and have a difficult time
>> understanding what the [] + [] line does :-/
>>
>> I'll simplify things by using a list instead of a dictionary:
>>
>> ============
>> connected = []
>> connected.append("0test")
>> connected.append("aa")
>> connected.append("bb")
>> connected.append("cc")
>>
>> for start in '1abcd':
>>     result = [name for name in connected if name[0] >= start] + [name
>> for name in connected if name[0] < start]
>>     print result
>> ============
>> C:\>test.py
>> ['aa', 'bb', 'cc', '0test']
>> ['aa', 'bb', 'cc', '0test']
>> ['bb', 'cc', '0test', 'aa']
>> ['cc', '0test', 'aa', 'bb']
>> ['0test', 'aa', 'bb', 'cc']
>> ============
>>
>> Pretty close to what I need to do but..
>>
>> 1. Why is the first iteration done twice?
>>
>> 2. How can I have just one line, save the character that I used as
>> starting point, increment it, and save it into a file so it can be
>> read the next time this program runs? For instance, let's say we used
>> "b" to start looking for items, I'll save "c" in a file for the next
>> time.
>>
>> Thank you.
>>   
>
> 1/ [] + [] is using 2 python lists comprehension, Google it for 
> details. It is quite difficult to read until you become familiar with 
> it. Once you get it, you can write magical stuff :o)
> Basically, list comrehension allows to map functions to list elements 
> and / or filter those elements.
>
> So what I'm using is the filter feature of list comprehension:
> [list of names for which the first char is greater that 'start'] + 
> [list of names for which the first char is less than 'start']
>
> 2/ Are you sure you want to do that ? looks like you are using a 
> hammer to smash a fly. As someone has suggested before, maybe you want 
> to pickup some starting index random.
>
> connected = ['aa', 'bb', 'cc', '0test']
>
> import random
>
> def getNewOrder(myList):
>     index = random.randint(0,len(myList)-1)
>     print index
>     return myList[index:] + myList[:index] # using slicing instead of 
> list comprehension (suggested by DaveA)
>
> print getNewOrder(connected)
>
> JM
>
>
Ok I realized that picking up a random index prevent from grouping names 
starting with the same letter (to ease visual lookup).
Then go for the random char, and use char comparison (my first example).

JM