Iterating Through Pandas Groupby And Merging Dataframes

This seems like it should be straightforward but is stumping me. Really love being able to iterate through the groups of a groupby operation and I am getting the result I want from

Solution 1:

Why not just:

pd.concat(dfs, axis=1, join='outer')

Solution 2:

Because this had to work on a very large dataset I went with the following implementation which doesn't seem very elegant at all but performs pretty well on large datasets:

for name, groupin df.groupby('date', sort=False):
    dfs.append(pd.DataFrame(pd.DataFrame(list(chain.from_iterable(group['values'])),
                              columns=['key']).groupby('key').size(),
                               columns=[name.strftime('%Y-%m-%d')]).reset_index())

df2 = pd.concat(dfs, axis=1, join='outer')

df3 = pd.merge(pd.merge(pd.merge(pd.merge(pd.merge(pd.merge(pd.merge(pd.merge(df2.iloc[:, :2], 
        pd.DataFrame(list(set(chain.from_iterable(df['values']))), columns=['key']), how='right'),
            df2.iloc[:, 2:4], how='left'),
            df2.iloc[:, 4:6], how='left'),
            df2.iloc[:, 6:8], how='left'),
            df2.iloc[:, 8:10], how='left'),
            df2.iloc[:, 10:12], how='left'),
            df2.iloc[:, 12:14], how='left'),
            df2.iloc[:, 14:16], how='left').fillna(0).set_index('key').sort_index(axis=1)