主页 / coding / 问题 / 79552670
Accepted
Whitebeard13
Asked: 2025-04-03 18:28:53 +0800 CST

将包含单个值的列转换为python pandas中的行

  • 772

考虑以下数据框示例：

maturity_date   simulation  simulated_price realized_price
30/06/2010      1           0.539333333     0.611
30/06/2010      2           0.544           0.611
30/06/2010      3           0.789666667     0.611
30/06/2010      4           0.190333333     0.611
30/06/2010      5           0.413666667     0.611

除了保留最后一列的值并连接之外，还有其他方法可以调整数据框以使最后一列变成行吗？

所需的输出如下：

maturity_date   simulation      simulated_price
30/06/2010      1               0.539333333     
30/06/2010      2               0.544           
30/06/2010      3               0.789666667     
30/06/2010      4               0.190333333     
30/06/2010      5               0.413666667     
30/06/2010      realized_price  0.611
python

1 个回答

  1. Best Answer

    也许更简单的是从最后一行处理字典，DataFrame.pop技巧是删除原始列realized_price：

    d = df.iloc[-1].to_dict()
d['simulated_price'] = d.pop('realized_price')
d['simulation'] = 'realized_price'
df.loc[len(df.pop('realized_price'))] = d

    选择：

    last = df.columns[-1]
d = df.iloc[-1].to_dict()
d['simulated_price'] = d.pop(last)
d['simulation'] = last
df.loc[len(df.pop(last))] = d

print (df)

  maturity_date      simulation  simulated_price
0    30/06/2010               1         0.539333
1    30/06/2010               2         0.544000
2    30/06/2010               3         0.789667
3    30/06/2010               4         0.190333
4    30/06/2010               5         0.413667
5    30/06/2010  realized_price         0.611000

    另一个想法是使用DataFrame.locDataFrame 的默认索引设置新行，通过选择中的最后一行DataFrame.iloc，rename然后重新附加新simulation值：realized_priceSeries.reindex

    s = (df.iloc[-1].drop(['simulated_price','simulation'])
                .rename({'realized_price':'simulated_price'})
                .reindex(df.columns[:-1], fill_value='realized_price'))

df.loc[len(df.pop('realized_price'))] = s

print (df)

  maturity_date      simulation  simulated_price
0    30/06/2010               1         0.539333
1    30/06/2010               2         0.544000
2    30/06/2010               3         0.789667
3    30/06/2010               4         0.190333
4    30/06/2010               5         0.413667
5    30/06/2010  realized_price         0.611000

    另一种方法是先重新分配列simulation，然后获取最后一行并处理Series：

    s = (df.assign(simulation='realized_price')
       .iloc[-1]
       .drop(['simulated_price'])
       .rename({'realized_price':'simulated_price'}))

df.loc[len(df.pop('realized_price'))] = s

print (df)


  maturity_date      simulation  simulated_price
0    30/06/2010               1         0.539333
1    30/06/2010               2         0.544000
2    30/06/2010               3         0.789667
3    30/06/2010               4         0.190333
4    30/06/2010               5         0.413667
5    30/06/2010  realized_price         0.611000

    另一个想法是concat：

    out = (pd.concat([df, 
                  df.iloc[[-1]]
                    .assign(simulation='realized_price',
                            simulated_price=df['realized_price'].iat[0])],
                 ignore_index=True)
        .drop('realized_price', axis=1))
print (out)
  maturity_date      simulation  simulated_price
0    30/06/2010               1         0.539333
1    30/06/2010               2         0.544000
2    30/06/2010               3         0.789667
3    30/06/2010               4         0.190333
4    30/06/2010               5         0.413667
5    30/06/2010  realized_price         0.611000
    • 2

相关问题