python - python 获取一个csv字段中2个数字的平均值

Question

我正在尝试清理 python(pandas)中的数据集(csv)

在“预计投资”列中，我有包含 2 个数字的数据。例如 30-35 我怎样才能得到这个平均值，以便该字段包含 32.5

Answer 1

我认为最好的是创建float列，而不是混合数字与字符串。

第一replace missing 为 NaN，然后 split ，转换为 float 并最后得到mean:

df = pd.DataFrame({'Projected investment':['missing','30-35','77']})
print (df)
  Projected investment
0              missing
1                30-35
2                   77

df['Projected investment'] = df['Projected investment'].replace('missing', np.nan) \
                                                       .str.split('-', expand=True) \
                                                       .astype(float) \
                                                       .mean(axis=1)
print (df)
   Projected investment
0                   NaN
1                  32.5
2                  77.0

print (df['Projected investment'].dtypes)
float64

如果需要缺少作为字符串:

def parse_number(x): 
    try: 
        return np.mean(np.array(str(x).split('-')).astype(float))
    except ValueError: 
        return x 

df['Projected investment'] = df['Projected investment'].map(parse_number) 
print (df)
  Projected investment
0              missing
1                 32.5
2                   77

print (df['Projected investment'].apply(type))
0              <class 'str'>
1    <class 'numpy.float64'>
2    <class 'numpy.float64'>
Name: Projected investment, dtype: object