Getting lists of values from a CSV

Question

I have a function that takes a column title, and a response.body from a urllib GET (I already know the body contains text/csv), and iterates through the data to build a list of values to be returned. My question to the gurus here: have I written this in the cleanest, most efficient way possible? Can you suggest any improvements?

def _get_values_from_csv(self, column_title, response_body):
    """retrieves specified values found in the csv body returned from GET
    @requires: csv
    @param column_title: the name of the column for which we'll build a list of return values.
    @param response_body: the raw GET output, which should contain the csv data
    @return: list of elements from the column specified.
    @note: the return values have duplicates removed. This could pose a problem, if you are looking for duplicates.
    I'm not sure how to deal with that issue."""
    dicts = [row for row in csv.DictReader(response_body.split("\r\n"))]
    results = {}
    for dic in dicts:
        for k, v in dic.iteritems():
            try:
                results[k] = results[k] + [v] #adds elements as list+list
            except: #first time through the iteritems loop.
                results[k] = [v]

    #one potential problem with this technique: handling duplicate rows
    #not sure what to do about it.
    return_list = list(set(results[column_title]))
    return_list.sort()
    return return_list

Answer 1

Here's a shorter function that does the same thing. It doesn't create lists for the columns you're not interested in.

def _get_values_from_csv(self, column_title, response_body):
    dicts = csv.DictReader(response_body.split("\r\n"))
    return sorted(set(d[column_title] for d in dicts))

Answer 2

My suggestions:

def _get_values_from_csv(self, column_title, response_body):
    # collect results in a set to eliminate duplicates
    results = set()

    # iterate the DictReader directly
    for dic in csv.DictReader(response_body.split("\r\n")):
        # only add the single column we are interested in
        results.add(dic[column_title])

    # turn the set into a list and sort it
    return_list = list(results)
    return_list.sort()
    return return_list