pandas数据帧中值的矢量化查找：另一个例子

Question

我看了一个例子 Vectorized look-up of values in Pandas dataframe

但不知怎么的，我的问题有点不同，我找不到正确的方法，这是一个简单的问题

所以我有一个数据帧

PoliceStations_raw=pd.DataFrame(
[['BAYVIEW'   ,37.729732,-122.397981],
 ['CENTRAL'   ,37.798732,-122.409919],
 ['INGLESIDE' ,37.724676,-122.446215],
 ['MISSION'   ,37.762849,-122.422005],
 ['NORTHERN'  ,37.780186,-122.432467],
 ['PARK'      ,37.767797,-122.455287],
 ['RICHMOND'  ,37.779928,-122.464467],
 ['SOUTHERN'  ,37.772380,-122.389412],
 ['TARAVAL'   ,37.743733,-122.481500],
 ['TENDERLOIN',37.783674,-122.412899]],columns=['PdDistrict','XX','YY'])

我还定义了

^{pr2}$

然后我有另一个表df，它有一个列'PdDistrict'，它包含一个分类变量，可以采用'BAYVIEW'、'CENTRAL'等值。。。在

我想要一个专栏 df['XX']它将为每一行df返回警察局中相应的条目。。。在

我找不到正确的语法。。。谢谢你的帮助

如果可能的话，我更喜欢使用PoliceStations_raw（而不是转置表）的语法，因为我认为这个表更“自然”。。。在

我试过了，但没用

df_raw['value'] = PoliceStations.lookup('XX',df_raw['PdDistrict'])

--------------------------------------------------------------------------- ValueError Traceback (most recent call last) in () ----> 1 df_raw['value'] = PoliceStations.lookup('XX',df_raw['PdDistrict'])

/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/pandas/core/frame.pyc in lookup(self, row_labels, col_labels) 2641 n = len(row_labels) 2642 if n != len(col_labels): -> 2643 raise ValueError('Row labels must have same size as column labels') 2644 2645 thresh = 1000

ValueError: Row labels must have same size as column labels

虽然我认为我没有犯过标签错误

df_raw['PdDistrict'].cat.categories

Index([u'BAYVIEW', u'CENTRAL', u'INGLESIDE', u'MISSION', u'NORTHERN', u'PARK', u'RICHMOND', u'SOUTHERN', u'TARAVAL', u'TENDERLOIN'], dtype='object')

编辑：

我也在尝试以下方法：

PoliceStations_raw=pd.DataFrame(
[['BAYVIEW'   ,37.729732,-122.397981],
 ['CENTRAL'   ,37.798732,-122.409919],
 ['INGLESIDE' ,37.724676,-122.446215],
 ['MISSION'   ,37.762849,-122.422005],
 ['NORTHERN'  ,37.780186,-122.432467],
 ['PARK'      ,37.767797,-122.455287],
 ['RICHMOND'  ,37.779928,-122.464467],
 ['SOUTHERN'  ,37.772380,-122.389412],
 ['TARAVAL'   ,37.743733,-122.481500],
 ['TENDERLOIN',37.783674,-122.412899]],columns=['PdDistrict','XX','YY'])


df1=pd.DataFrame([[0,'CENTRAL'],[1,'TARAVAL'],[3,'CENTRAL'],[2,'BAYVIEW']])
df1.columns = ['Index','PdDistrict']


  Index PdDistrict
0   0   CENTRAL
1   1   TARAVAL
2   3   CENTRAL
3   2   BAYVIEW

尽管输入了sort=False，但返回的对象已经合并了表，但使用PdDistrict作为索引，并更改了原始左dataframe的行的顺序。在

救命啊！在

pd.merge(df1,PoliceStations_raw,sort=False)

把这个还给我

  Index PdDistrict  XX        YY
0   0   CENTRAL 37.798732   -122.409919
1   3   CENTRAL 37.798732   -122.409919
2   1   TARAVAL 37.743733   -122.481500
3   2   BAYVIEW 37.729732   -122.397981