检查dtype of series是否属于熊猫中的dtype列表

Question

我有一个熊猫系列。我想检查这个系列的dtype是否在一个dtype列表中。类似于：

series.dtype not in [pd.dtype('float64'), pd.dtype('float32')]

这给了我以下错误：

AttributeError: 'module' object has no attribute 'dtype'

我该怎么办？

Answer 1

您可以使用dtypes来检查Series是否不是float

print df
          a      b    c
0 -1.828219   True  1.0
1  0.681694  False  2.0
2 -2.360949   True  1.0
3  1.034397  False  2.0
4  1.073993   True  1.0
5  1.306872  False  2.0

print df.a.dtype
float32
print df.b.dtype
bool

print df.a.dtype not in [pd.np.dtype('float64'), pd.np.dtype('float32')] 
False           
print df.b.dtype not in [pd.np.dtype('float64'), pd.np.dtype('float32')] 
True

工作np.dtype也如提到的MaxU

print df.a.dtype not in [np.dtype('float64'), np.dtype('float32')] 
False              
print df.b.dtype not in [np.dtype('float64'), np.dtype('float32')] 
True

Answer 2

没有像pd.dtype()这样的方法/东西--有np.dtype()或pd.np.dtype()

In [86]: pd.dtype
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-86-dbe1c1048375> in <module>()
----> 1 pd.dtype

AttributeError: module 'pandas' has no attribute 'dtype'

In [87]: np.dtype
Out[87]: numpy.dtype

In [88]: pd.np.dtype
Out[88]: numpy.dtype

Answer 3

使用select_dtypes和pass include=['floating']

In [11]:
df = pd.DataFrame({'float':[0.0, 1.2, 5.5], 'int':[0,1,2], 'str':list('abc')})
df

Out[11]:
   float  int str
0    0.0    0   a
1    1.2    1   b
2    5.5    2   c

In [12]:
df.select_dtypes(include=['floating'])

Out[12]:
   float
0    0.0
1    1.2
2    5.5