[Tutor] update list of dictionaries based on key

[greg whittier]

On Fri, Mar 13, 2009 at 11:09 AM, ski <norman at khine.net> wrote:
> Hello,
> Here is what I have so far:
>
>>>> mylist = [{'index': 0, 'title': 'Association of British Travel Agents',
>>>> 'selected': False, 'edit_row': '?edit_affiliation=0', 'affiliation': 'ABTA',
>>>> 'affiliation_no': u'G3903'}, {'index': 1, 'title': 'Appointed Agents of
>>>> IATA', 'selected': False, 'edit_row': '?edit_affiliation=1', 'affiliation':
>>>> 'IATA', 'affiliation_no': u'none'}, {'index': 0, 'title': 'Association of
>>>> Airline Cons.', 'selected': False, 'edit_row': '?edit_affiliation=0',
>>>> 'affiliation': 'AAC', 'affiliation_no': u'sss'}, {'index': 1, 'title':
>>>> 'Association of British Travel Agents', 'selected': False, 'edit_row':
>>>> '?edit_affiliation=1', 'affiliation': 'ABTA', 'affiliation_no': u'zser'}]
>
>>>> key = 'affiliation_no'
>
>>>> my_dup_keys = []
>>>> merged_list = []
>
>>>> mykeys = [item['affiliation'] for item in mylist]
>
>>>> for x in mykeys:
>        if mykeys.count(x) >= 2:
>                my_dup_keys.append(x)
>
>>>> for item in mylist:
>        if item['affiliation'] in set(my_dup_keys):
>                merged_list.append(item)
>
>
>>>> values = [d[key] for d in merged_list if d.has_key(key)]
>>>> for dict in merged_list:
>        dict[key] = values
>        mylist.append(dict)
>
>>>> print mylist
> [{'index': 0, 'title': 'Association of British Travel Agents', 'selected':
> False, 'edit_row': '?edit_affiliation=0', 'affiliation': 'ABTA',
> 'affiliation_no': [u'G3903', u'zser']}, {'index': 1, 'title': 'Appointed
> Agents of IATA', 'selected': False, 'edit_row': '?edit_affiliation=1',
> 'affiliation': 'IATA', 'affiliation_no': u'none'}, {'index': 0, 'title':
> 'Association of Airline Cons.', 'selected': False, 'edit_row':
> '?edit_affiliation=0', 'affiliation': 'AAC', 'affiliation_no': u'sss'},
> {'index': 1, 'title': 'Association of British Travel Agents', 'selected':
> False, 'edit_row': '?edit_affiliation=1', 'affiliation': 'ABTA',
> 'affiliation_no': [u'G3903', u'zser']}, {'index': 0, 'title': 'Association
> of British Travel Agents', 'selected': False, 'edit_row':
> '?edit_affiliation=0', 'affiliation': 'ABTA', 'affiliation_no': [u'G3903',
> u'zser']}, {'index': 1, 'title': 'Association of British Travel Agents',
> 'selected': False, 'edit_row': '?edit_affiliation=1', 'affiliation': 'ABTA',
> 'affiliation_no': [u'G3903', u'zser']}]
>
>
> This sort of works but I want to return a list that updates 'mylist' not
> append to it, so I will get:
>
> [{'index': 0, 'title': 'Association of British Travel Agents', 'selected':
> False, 'edit_row': '?edit_affiliation=0', 'affiliation': 'ABTA',
> 'affiliation_no': [u'G3903', u'zser']}, {'index': 1, 'title': 'Appointed
> Agents of IATA', 'selected': False, 'edit_row': '?edit_affiliation=1',
> 'affiliation': 'IATA', 'affiliation_no': u'none'}, {'index': 0, 'title':
> 'Association of Airline Cons.', 'selected': False, 'edit_row':
> '?edit_affiliation=0', 'affiliation': 'AAC', 'affiliation_no': u'sss'}]
>

It's a little hard to figure what you're getting at.  This looks like
a table represented by a list of dictionaries that you'd like to group
by "affiliation."

affiliation_nos = {}
for row in mylist:
    affiliation_nos.set_default(row['affiliation'],[]).append(row['affliation_no'])

will give you a list of affiliation_no's for each affiliation and if
you want a list of dicts, you could do

mynewlist = [dict(affiliation=affiliation, affiliation_nos =
affiliation_nos[affiliation]) for affiliation in
affiliation_nos.keys()]

I don't know what to do about all the other fields though.  In your
example input list you have two dictionaries with affiliation =
'ABTA'.  In your output you kept the one with index=0 and threw away
index=1.  (same for 'edit_row')  How do you determine which to keep?