python - 对值列表进行分组并执行汇总功能，例如 sum 和 count

Question

按节目分组，统计对应的 VideoID，对观看次数求和，对收入求和。我怎样才能做到这一点？

我的桌子看起来像

    show.    views.  revenue  VideoID
    batman.   100.     10.    v1
    batman.   200.     20.    v2
    joker.    100.     10.    v3
    joker.    300.     15.    v4
    superman. 500.     30.    v5

我的预期输出是

Show          total_views    total_revenue.    video_count
batman.           300.            30.           2
joker             400.            25            2
superman          500             30            1

我怎样才能做到这一点？

这是我迄今为止尝试过的，但输出错误

def grouping_series(df_series):
    t = defaultdict(list)
    gp = df_series.groupby('show')
    for i, k in gp:
        t['total_views'].append(k['views'].sum())
        t['total_revenue'].append(k['revenue'].sum())
        t['video_count'].append(k['VideoID'].count())
    return pd.DataFrame(t)

df = grouping_series(df_series)    

Answer 1

我们通常做agg

s=df.groupby('show').agg(total_views=('views', 'sum'), 
                         total_revenue=('revenue', 'sum'),
                         video_count=('VideoID', 'nunique')).reset_index()
        show  total_views  total_revenue  video_count
0    batman.        300.0           30.0            2
1     joker.        400.0           25.0            2
2  superman.        500.0           30.0            1

Answer 2

这是我的建议：

iimport pandas as pd

frame = {
    "show": ["batman", "batman", "joker", "joker", "superman"],
    "views": [100, 200, 100, 300, 500],
    "revenue": [10, 20, 10, 15, 30],
    "VideoID": ["v1", "v2", "v3", "v4", "v5"],
}

df = pd.DataFrame(frame)

aggretations = {"views": "sum", "revenue": "sum", "VideoID": "nunique"}

df.groupby(["show"]).agg(aggretations)