使用python将其他两个键的并集追加到字典中的键

Question

This is my input :

ClientData = {
'ClientName1': {
           'aggregate_PageviewsByWeek': [('2013-01-06', 2),
                                       ('2013-02-03', 1),
                                       ('2013-02-10', 1),
                                       ('2013-02-24', 1),
                                       ('2013-03-03', 2),

           'aggregate_Pageviews_VisitsByWeek': [],
           'aggregate_VisitsByWeek': [('2013-03-03', 1),
                                         ('2013-05-12', 1)]

                                       },


'ClientName2': {
           'aggregate_PageviewsByWeek': [('2013-01-06', 2),
                                       ('2013-02-03', 1),
                                       ('2013-02-10', 1),
                                       ('2013-02-24', 1),
                                       ('2013-03-03', 2),
                                       ('2013-03-24', 1),
      ],
           'aggregate_Pageviews_VisitsByWeek': [],
           'aggregate_VisitsByWeek': [('2013-03-03', 1),
                                      ('2013-03-31', 1),
                                      ('2013-05-12', 1),
                                      ('2013-05-19', 2),
                                      ('2013-06-30', 2)]
                                       }

}

How can I append to the key 'aggregate_Pageviews_VisitsByWeek' the union of the 'aggregate_PageviewsByWeek' and 'aggregate_VisitsByWeek' based on the date key

the output will looks like something similar to this :

{
'ClientName1': {
           'aggregate_PageviewsByWeek': [('2013-01-06', 2),
                                       ('2013-02-03', 1),
                                       ('2013-02-10', 1),
                                       ('2013-02-24', 1),
                                       ('2013-03-03', 2)],

           'aggregate_Pageviews_VisitsByWeek': [

                                               ('2013-01-06', 2, 0),
                                               ('2013-02-03', 1, 0),
                                               ('2013-02-10', 1, ),
                                               ('2013-02-24', 1, 0),
                                               ('2013-03-03', 2, 1),
                                               ('2013-05-12', 0, 1)],
           'aggregate_VisitsByWeek': [('2013-03-03', 1),
                                         ('2013-05-12', 1)]

                                       },



'ClientName2': {
           'aggregate_PageviewsByWeek': [('2013-01-06', 2),
                                       ('2013-02-03', 1),
                                       ('2013-02-10', 1),
                                       ('2013-02-24', 1),
                                       ('2013-03-03', 2),
                                       ('2013-03-24', 1)],

           'aggregate_Pageviews_VisitsByWeek': [
                                       ('2013-01-06', 2, 0),
                                       ('2013-02-03', 1, 0),
                                       ('2013-02-10', 1, 0),
                                       ('2013-02-24', 1, 0),
                                       ('2013-03-03', 2, 1),
                                       ('2013-03-31', 1, 1),
                                       ('2013-05-12', 0, 1),
                                       ('2013-05-19', 0, 2),
                                       ('2013-06-30', 0, 2)],

           'aggregate_VisitsByWeek': [('2013-03-03', 1),
                                      ('2013-03-31', 1),
                                      ('2013-05-12', 1),
                                      ('2013-05-19', 2),
                                      ('2013-06-30', 2)]
                                       }

}

if the key "which the date in this case" is not on the other list I want to replace that value with 0 (Date, aggregate_PageviewsByWeek_Value, aggregate_VisitsByWeek_Value )

example :
aggregate_PageviewsByWeek : ('2013-01-06', 12) and aggregate_VisitsByWeek : (2013-01-13, 30)

the output will be :
aggregate_Pageviews_VisitsByWeek : [('2013-01-06', 12, 0), (2013-01-13, 0, 30)]

my goal of thsi question is to get the trends of page views and visits based on the date.

Thanks!

Answer 1

First, you need a function that merges a single client's entries.

There are two easy ways to merge parallel sequences that might each be missing some values: You can iterate the two in parallel, or you can build a dictionary (or sorted map) of keys, and just handle each sequence separately. You can see an example of the first, eg, here . But the second is simpler, at least in Python, so long as the keys are hashable. So:

def merge_client(client):
    merged = {}
    for day, views in client['aggregate_PageviewsByWeek']:
        merged[day] = [views, 0]
    for day, visits in client['aggregate_VisitsByWeek']:
        merged.setdefault(day, [0, 0])[1] = visits
    flattened = [tuple([key] + value) for key, value in merged.items()]
    client['aggregate_Pageviews_VisitsByWeek'] = sorted(flattened)

To make this algorithm to more than two entries, you'd use append —or, if there may be a huge number of entires, just use a dict instead of a list (so we don't have to fill in all those default 0's until the end).

Now we just need to call this on each client in the list:

for client in ClientData.values():
    merge_client(client)

Answer 2

Convert each list to dict, combine keys of these dicts, loop thru keys and generate another list, where each element is date, value from first dict or 0 and value from second dict or 0, it is better explained via code :)

def merge_lists(list1, list2):
    dict1 = dict(list1)
    dict2 = dict(list2)
    dates = list(set(dict1.keys())|set(dict2.keys()))
    dates.sort()
    merged_list = []
    for date in dates:
        item = [date]
        item.append(dict1.get(date,0))
        item.append(dict2.get(date,0))
        merged_list.append(item)

    return merged_list

merged_list = merge_lists([('2013-01-06', 2),
            ('2013-02-03', 1),
            ('2013-02-10', 1),
            ('2013-02-24', 1),
            ('2013-03-03', 2),
            ('2013-03-24', 1)],
            [('2013-03-03', 1),
            ('2013-03-31', 1),
            ('2013-05-12', 1),
            ('2013-05-19', 2),
            ('2013-06-30', 2)])


import pprint
pprint.pprint(merged_list)

output:

[['2013-01-06', 2, 0],
 ['2013-02-03', 1, 0],
 ['2013-02-10', 1, 0],
 ['2013-02-24', 1, 0],
 ['2013-03-03', 2, 1],
 ['2013-03-24', 1, 0],
 ['2013-03-31', 0, 1],
 ['2013-05-12', 0, 1],
 ['2013-05-19', 0, 2],
 ['2013-06-30', 0, 2]]

You can make it generic by merging any number of lists

def merge_lists(*lists):
    dicts = [dict(l) for l in lists]
    dates = set()
    for d in dicts:
        dates |= set(d.keys())
    dates = list(dates)
    dates.sort()
    merged_list = []
    for date in dates:
        item = [date]
        for d in dicts:
            item.append(d.get(date,0))
        merged_list.append(item)

    return merged_list