https://stackoverflow.com/questions/36928487/a-value-is-trying-to-be-set-on-a-copy-of-a-slice-from-a-dataframe
df = pd.DataFrame({'period': ['Q1', 'Q2', 'Q3', 'Q4', 'Q5'], 'qtr': [1, 2, 3, 4, 5]})
| period | qtr |
|---|---|
| Q1 | 1 |
| Q2 | 2 |
| Q3 | 3 |
| Q4 | 4 |
| Q5 | 5 |
1. map解决---这个还是有copywarning
quarter = {"Q1": "Mar", "Q2": "Jun", "Q3": "Sep", "Q4": "Dec"}
df['period'] = df['period'].astype(str).map(quarter)
| period | qtr |
|---|---|
| Mar | 1 |
| Jun | 2 |
| Sep | 3 |
| Dec | 4 |
| NaN | 5 |
2. assign
Unmapped values unchanged
df = df.assign(period=[quarter.get(q, q) for q in df.period])
| period | qtr |
|---|---|
| Mar | 1 |
| Jun | 2 |
| Sep | 3 |
| Dec | 4 |
| 5 | 5 |
Unmapped values get NaN
df = df.assign(period=df.period.map(quarter))
| period | qtr |
|---|---|
| Mar | 1 |
| Jun | 2 |
| Sep | 3 |
| Dec | 4 |
| NaN | 5 |
3. 单独赋值
df[df['A'] > 2]['B'] = new_val # new_val not set in df
The warning offers a suggestion to rewrite as follows:
df.loc[df['A'] > 2, 'B'] = new_val