Question

我有以下数据框df1：

    X           Y           Order_  NEW_ID
0   484970.4517 408844.0920 95083   1320437
1   478512.3233 415791.5395 96478   1320727
2   504516.3032 452923.4420 105246  1321260
3   485147.0529 428172.1055 99633   1320979

另一个，df2：

    Order_  Loc
0   83158   239,211
1   83159   239,212
2   83160   239,213
3   83161   239,214

我希望与第一列合并，以便Loc列添加正确的值df1。要进行合并，我使用map执行左合并，首先将Loc值转换为字符串：

df2['Loc'] = df2['Loc'].astype(str)
df1['Loc']=df1.Order_.map(df2.Loc)

结果很奇怪，Loc中显示的df1值属于NaN类型：

    X           Y           Order_  NEW_ID  Loc
0   484970.4517 408844.0920 95083   1320437 NaN
1   478512.3233 415791.5395 96478   1320727 NaN
2   504516.3032 452923.4420 105246  1321260 NaN
3   485147.0529 428172.1055 99633   1320979 NaN

虽然我希望它们是字符串并以239,211方式出现（包含逗号的字符串）。在调查Loc df2中的dtype时，我得到：

Order_     int64
Loc       object
dtype: object

我的问题：如何在对象之间执行类型更改，以便我能够有效地读取Loc值，并避免它们成为NaN }？

Answer 1

我认为如果需要Order_，我需要将int投射到dtypes：

df1['Order_'] = df1['Order_'].astype(int)

但也许问题是您需要按Series或dict制作地图，因此必须将Order_设置为索引：

d = df2.set_index('Order_')['Loc'].to_dict()
df1['Loc']= df1.Order_.map(d)

样品：

print (df1)
             X            Y  Order_   NEW_ID
0  484970.4517  408844.0920   95083  1320437
1  478512.3233  415791.5395   96478  1320727
2  504516.3032  452923.4420  105246  1321260
3  485147.0529  428172.1055   99633  1320979

print (df2)
   Order_      Loc
0   95083  239,211 <-first value was changed for align
1   83159  239,212
2   83160  239,213
3   83161  239,214

#check if same dtypes
print (df1['Order_'].dtypes)
int64
print (df2['Order_'].dtypes)
int64


d = df2.set_index('Order_')['Loc'].to_dict()
print (d)
{83160: '239,213', 83161: '239,214', 95083: '239,211', 83159: '239,212'}

df1['Loc']= df1.Order_.map(d)
print (df1)
             X            Y  Order_   NEW_ID      Loc
0  484970.4517  408844.0920   95083  1320437  239,211
1  478512.3233  415791.5395   96478  1320727      NaN
2  504516.3032  452923.4420  105246  1321260      NaN
3  485147.0529  428172.1055   99633  1320979      NaN

熊猫：使用`map`进行左合并返回NaN

1 个答案: