[Tutor] Sorting numbers

[Jeff Shannon]

Adam Vardy wrote:

>Tuesday, February 4, 2003, 4:27:01 PM, you wrote:
>
>>> >>> rawdata
>>>['45K', '100K', '3.4Meg', '17K', '300K', '9.3Meg', '512', '23Meg']
>>> >>> srt = {}
>>> >>> for item in rawdata:
>>>...     num, suffix = splitsuffix(item)
>>>...     value = srt.get(suffix, [])
>>>...     value.append(num)
>>>...     srt[suffix] = value
>>>      
>>>
>
>Can you explain the last three lines here?
>  
>

Sure.  First let's look at the line 'value = srt.get(suffix, [])'.  We 
have a dictionary called srt, and we want to see if it has anything in 
the key that's held in suffix.  I could've said 'value = srt[suffix]', 
and then value would contain whatever had previously been stored in the 
dictionary with that key, but what happens if nothing had been 
previously stored?

 >>> srt = {}
 >>> value = srt['Meg']
Traceback (most recent call last):
  File "<interactive input>", line 1, in ?
KeyError: Meg
 >>>

We get an exception.  Now, since I'm starting with an empty dictionary, 
I know that I'm going to try looking up nonexistent keys at least a few 
times.  I *could* use a try/except block, and respond to the exception, 
but there's an easier way.  Dictionaries have a convenient get() method, 
which will look up a key like normal, but if that key is not found then 
it will return a default value.  

 >>> value = srt.get('Meg', [])
 >>> value
[]
 >>> value = srt.get('Meg', [])
 >>> value
['2.4', '9.6']
 >>>

This basically asks the dictionary, "Give me whatever you've got for 
'Meg', but if you don't have anything, then give me an empty list instead."

Now 'value' is a list containing whatever had been previously been 
stored in the dictionary with that suffix -- or an empty list if nothing 
had been stored previously.  The next line, 'value.append(num)', tacks 
the current numeric value onto the end of that list.  Finally, the last 
line 'srt[suffix] = value' takes the modified list and stores that list 
back in the dictionary, under the same key as before.

The overall effect of these three lines, then, is that the list that the 
dictionary has stored under a given suffix has the current value of num 
tacked on the end, with new lists being created as necessary when new 
suffixes are found.  As the original list of raw data is iterated 
through, it's sorted into a series of lists depending on what the 
non-numeric suffix on the string is.  Eventually, all of the items that 
had ended in 'Meg' are stored in srt['Meg'], all of the numbers that had 
ended in 'K' are stored in srt['K'], and so on.  

Hope that helps.  If there's anything else you're not clear on, feel 
free to ask.

Jeff Shannon
Technician/Programmer
Credit International