如何根据数据框值和字典键向数据框添加行

Question

我有一个 defaultdict 和一个数据框看起来像：

[('SERVER01', ['app01', 'app02']), ('SERVER02', ['app03','app04']), ('SERVER03', ['app05', 'app06','app07'])]

数据框：

Date           Server          Satut          Risk
------------------------------------------------------
01/01/2019     SERVER01         Up             High
08/01/2019     SERVER02         Down           Low
01/02/2019     SERVER03         Up             High
08/02/2019     SERVER01         Down           High
10/02/2019     SERVER01         Up             Low

我想要输出：

Date           Server/app      Satut          Risk
------------------------------------------------------
01/01/2019     SERVER01         Up             High
01/01/2019     app01            Up             High
01/01/2019     app02            Up             High
08/01/2019     SERVER02         Down           Low
08/01/2019     app03            Down           Low
08/01/2019     app04            Down           Low
01/02/2019     SERVER03         Up             High
01/02/2019     app05            Up             High
01/02/2019     app06            Up             High
01/02/2019     app07            Up             High
08/02/2019     SERVER01         Down           High
08/02/2019     app01            Down           High
08/02/2019     app02            Down           High
10/02/2019     SERVER01         Up             Low
10/02/2019     app01            Up             Low
10/02/2019     app02            Up             Low

所以我想将键与列服务器的值链接并复制该行然后用应用程序替换服务器

Answer 1

所以这是你的默认字典：

d = defaultdict(list,
                {'SERVER01': ['app01', 'app02'],
                 'SERVER02': ['app03', 'app04'],
                 'SERVER03': ['app05', 'app06', 'app07']})

app_df = pd.DataFrame()
for k in d:
    temp_df = pd.DataFrame(d[k], 
                          index=[k] * len(d[k])).reset_index()
    temp_df.columns = ['Server', 'App']
    app_df = pd.concat([app_df, temp_df])

# This will give you the Server and App dataframe
    Server      App
0   SERVER01    app01
1   SERVER01    app02
0   SERVER02    app03
1   SERVER02    app04
0   SERVER03    app05
1   SERVER03    app06
2   SERVER03    app07

这段代码可以满足你的需要：

# Iterate over the keys of the defaultdict and get the row from the dataframe corresponds to that server
for k in d:
    r = df_raw.loc[df_raw['Server'] == k].to_dict(orient = 'records')[0]

    # Iterate over the apps of that server
    for app in d[k]:
        # make a copy of that row
        new_row = r.copy()
        # Update the Server key value with the app value
        new_row['Server'] = app
        # Convert it to a dataframe
        df_temp = pd.DataFrame.from_dict(new_row, orient='index').T
        # Append it to the main dataframe
        df_raw = df_raw.append(df_temp)

# Merge the two dataframes together
df_raw.merge(app_df, on='Server', how='inner')

输出此数据帧：

    Date    Server  Statu    Risk   App
0   1/1/19  SERVER01    Up   High   app01
1   1/1/19  SERVER01    Up   High   app02
2   8/1/19  SERVER02    Down Low    app03
3   8/1/19  SERVER02    Down Low    app04
4   1/2/19  SERVER03    Up   High   app05
5   1/2/19  SERVER03    Up   High   app06
6   1/2/19  SERVER03    Up   High   app07

这可能有点矫枉过正，但这是我想出的第一件事，希望对您有所帮助！