Question

我有2个数据帧如下：

import pandas as pd
df = pd.DataFrame([[1,2,1],[4,15,11],[7,22,15]],columns=['Start','End','Duration'])

df2 = pd.DataFrame([[0,2,3],[1,2,2],[1,8,9],[2,1,1],[2,1,1]],columns=['Index','Data1','Data2'])

df
Out[6]: 
   Start  End  Duration
0      1    2         1
1      4   15        11
2      7   22        15

df2
Out[7]: 
   Index  Data1  Data2
0      0      2      3
1      1      2      2
2      1      8      9
3      2      1      1
4      2      1      1

我想要做的是使用以下标准创建第3个数据帧df3：

查看df2的专栏Index，

1）如果有1个唯一值，我取df [Index]的开始和结束并将其附加到df2，以便df3

df3
Out[7]: 
Index  Data1  Data2   Start   End     Duration 
    0      2      3       1     2            1

2）如果存在多个相等的值（例如，对于Index = 1），我取第一个的开始（即4）并且结束I加5 - ＆gt; 9和持续时间将是5

df3
Out[7]: 
Index  Data1  Data2   Start   End     Duration 
    0      2      3       1     2            1
    1      2      2       4     9            5

另一个

df3
Out[7]: 
Index  Data1  Data2   Start   End     Duration 
    0      2      3       1     2            1
    1      2      2       4     9            5
    1      8      9      10    15            5

最终的数据帧df3应为：

df3
Out[7]: 
Index  Data1  Data2   Start   End     Duration 
    0      2      3       1     2            1
    1      2      2       4     9            5
    1      8      9      10    15            5
    2      1      1       7    12            5
    2      1      1      13    18            5

这是我的尝试：

import pandas as pd
import numpy as np

df1 = pd.DataFrame([[1,2,1],[4,15,11],[7,22,15]],columns=['Start','End','Duration'])

df2 = pd.DataFrame([[0,2,3],[1,2,2],[1,8,9],[2,1,1],[2,1,1]],columns=['Index','Data1','Data2'])
df2.index = df2['Index']
df3 = pd.DataFrame()

for index in df1.index:


    current_idx_df1 = pd.DataFrame(df1.loc[index].values.reshape(1,-1),columns=[df1.keys()])

    if np.sum(df2.index==index)==1:
        current_idx_df2 = pd.DataFrame(df2.loc[index].values.reshape(1,-1),columns=[df2.keys()])
        df3 = df3.append(pd.concat([current_idx_df1, current_idx_df2 ], axis=1))
        df3.reset_index(drop =True,inplace =True)
    else:
        current_idx_df2 = df2.loc[index]
        df3_temp = pd.concat([current_idx_df1, pd.DataFrame(current_idx_df2.iloc[0].values.reshape(1,-1),columns=[current_idx_df2.keys()]) ], axis=1)
        df3_temp['End'] = df3_temp['Start']+5
        df3_temp['Duration'] = 5

        df3 = df3.append(df3_temp)
        df3.reset_index(drop =True,inplace =True)

        for index2 in range(1,np.sum(df2.index==index)):

            df3_temp = pd.concat([current_idx_df1, pd.DataFrame(current_idx_df2.iloc[index2].values.reshape(1,-1),columns=[current_idx_df2.keys()]) ], axis=1)
            df3_temp['Start'] = df3['End'].values[-1]+1
            df3_temp['End'] = df3['End'].values[-1]+1+5
            df3_temp['Duration'] = 5
            df3 = df3.append(df3_temp)
            df3.reset_index(drop =True,inplace =True)

合并2个pandas数据帧

0 个答案: