Iterable TypeError遇到NaN值时

Question

我有一个包含NaN值的数据框架

df:
score home_odds draw_odds away_odds
0          1:0    59/100   263/100    231/50
1          2:1     24/25     53/20   237/100
2          0:2   221/100     93/50     67/50
3          1:1    259/50   251/100     16/25
.......
1970       NaN     36/25   197/100     47/25
1971       NaN       NaN       NaN       NaN
1972       NaN       NaN       NaN       NaN
1973       NaN       NaN       NaN       NaN
1974       NaN      17/5   263/100     39/50
1975       NaN     77/50     21/10       7/4
1976       NaN     19/50     86/25   691/100
1977       NaN       NaN       NaN       NaN

当我尝试用转换公式

将小数值转换为小数时例句:

123/100 = (123/100 + 1) = 2.23

333/100 = (333/100 +1) = 4.33

计算为分数值+ 1

当然，保留十进制值的代码:

def convert(s):
if '/' in s:  # is a fraction
num, den = s.split('/')
return 1 + (int(num) / int(den))
else:
return float(s)
df['home_odds'] = df['home_odds'].apply(convert)
df['away_odds'] = df['away_odds'].apply(convert)
df['draw_odds'] = df['draw_odds'].apply(convert)

我得到错误:

if '/' in s:  # is a fraction
TypeError: argument of type 'float' is not iterable

当我删除NaN值时，公式生效。

我想在数据框中保留NaN值，只转换非NaN值。

我该怎么做?

Answer 1

检查在函数开始时数据是否为Nan(这是float):

def convert(s):
if pd.isnull(s):
return s
if '/' in s:  # is a fraction
num, den = s.split('/')
return 1 + (int(num) / int(den))
else:
return float(s)

Answer 2

与pd.eval的一个hack。使用str.replace代替:

1. 59/100by1+59/100(结果是1.59)
2. NaNby1e999-1e999(结果是NaN)

df.update(df.filter(like='_odds')
.replace({r'(d+)/(d+)': r'1+1/2',
np.NaN: '1e999-1e999'}, regex=True)
.apply(pd.eval))

>>> df
score home_odds draw_odds away_odds
0      1:0      1.59      3.63      5.62
1      2:1      1.96      3.65      3.37
2      0:2      3.21      2.86      2.34
3      1:1      6.18      3.51      1.64
1970   NaN      2.44      2.97      2.88
1971   NaN       NaN       NaN       NaN
1972   NaN       NaN       NaN       NaN
1973   NaN       NaN       NaN       NaN
1974   NaN       4.4      3.63      1.78
1975   NaN      2.54       3.1      2.75
1976   NaN      1.38      4.44      7.91
1977   NaN       NaN       NaN       NaN