python - 如何根据条件将列中的值设置为另一列

Question

我有以下df：

subscription|amount| Total
a           |500   |
x           | 0    | 5000
x           |7500  | 5000
y           |7500  | 5000
y           | 0    | 5000
z           |7500  | 5000
z           | 0    | 5000
b           | 0    | 
b           |1000  |

我想将“总计”的值转移到“金额”中已经有一个值的“金额”列，当然还有“总计”。

我过滤了：

df.loc[(df['total'].notnull()) & (df['amount']!=0)]

但无法将值从总计（5000）转移到金额。

期望的输出：

subscription|amount| Total
a           |500   |
x           | 0    | 5000
x           |5000  | 5000
y           |5000  | 5000
y           | 0    | 5000
z           |5000  | 5000
z           | 0    | 5000
b           | 0    | 
b           |1000  |

Answer 1

如果将列更改为浮点数没有问题，还有更多解决方案 -DataFrame.loc或numpy.where：

mask = (df['Total'].notnull()) & (df['amount']!=0)

---

df.loc[mask, 'amount'] = df['Total']

df['amount'] = np.where(mask, df['Total'], df['amount'])

print (df)
  subscription  amount   Total
0            a   500.0     NaN
1            x     0.0  5000.0
2            x  5000.0  5000.0
3            y  5000.0  5000.0
4            y     0.0  5000.0
5            z  5000.0  5000.0
6            z     0.0  5000.0
7            b     0.0     NaN
8            b  1000.0     NaN

没有将整数列更改为浮点数或转换为with 的Series.mask解决方案：Series.whereintegersnp.where

df['amount'] = df['amount'].mask(mask, df['Total'])

df['amount'] = df['amount'].where(~mask, df['Total'])

df['amount'] = df['amount'].where(~mask, df['Total']).astype(int)

print (df)
  subscription  amount   Total
0            a     500     NaN
1            x       0  5000.0
2            x    5000  5000.0
3            y    5000  5000.0
4            y       0  5000.0
5            z    5000  5000.0
6            z       0  5000.0
7            b       0     NaN
8            b    1000     NaN