当连接列有轻微的拼写差异时，如何将列合并/添加到 pandas 中的数据框？

Question

所以我有这样一个数据框

   Rank        State/Union territory  NSDP Per Capita (Nominal)(2019–20)[1][2]  state_id
0     1                          Goa                                  466585.0      30.0
1     2                       Sikkim                                  425656.0      11.0
2     3                        Delhi                                  376143.0       NaN
3     4                   Chandigarh                                       NaN       4.0
4     5                      Haryana                                  247207.0       6.0
5     6                    Telangana                                  225756.0       0.0
6     7                    Karnataka                                  223246.0      29.0
7     8                       Kerala                                  221904.0      32.0
8     9                   Puducherry                                  220949.0      34.0
9    10  Andaman and Nicobar Islands                                  219842.0       NaN
10   11                   Tamil Nadu                                  218599.0      33.0
11   12                      Gujarat                                  216329.0      24.0
12   13                      Mizoram                                  204018.0      15.0
13   14                  Uttarakhand                                  202895.0       5.0
14   15                  Maharashtra                                  202130.0      27.0
15   16             Himachal Pradesh                                  190255.0       2.0
16   17               Andhra Pradesh                                  168480.0      28.0
17   18            Arunachal Pradesh                                  164615.0       NaN
18   19                       Punjab                                  161083.0       3.0
20   20                     Nagaland                                  130282.0      13.0
21   21                      Tripura                                  125630.0      16.0
22   22                    Rajasthan                                  115492.0       8.0
23   23                  West Bengal                                  115348.0      19.0
24   24                       Odisha                                   98896.0      21.0
25   25                 Chhattisgarh                                  105281.0      22.0
26   26            Jammu and Kashmir                                  102882.0       NaN
27   27               Madhya Pradesh                                  103288.0      23.0
28   28                    Meghalaya                                   92174.0      17.0
29   29                        Assam                                   90758.0      18.0
30   30                      Manipur                                   84746.0      14.0
31   31                    Jharkhand                                   79873.0      20.0
32   32                Uttar Pradesh                                   65704.0       9.0
33   33                        Bihar                                   46664.0      10.0

我的另一本词典有

{'Telangana': 0, 'Andaman & Nicobar Island': 35, 'Andhra Pradesh': 28, 'Arunanchal Pradesh': 12, 'Assam': 18, 'Bihar': 10, 'Chhattisgarh': 22, 'Daman
& Diu': 25, 'Goa': 30, 'Gujarat': 24, 'Haryana': 6, 'Himachal Pradesh': 2, 'Jammu & Kashmir': 1, 'Jharkhand': 20, 'Karnataka': 29, 'Kerala': 32, 'Lakshadweep': 31, 'Madhya Pradesh': 23, 'Maharashtra': 27, 'Manipur': 14, 'Chandigarh': 4, 'Puducherry': 34, 'Punjab': 3, 'Rajasthan': 8, 'Sikkim': 11, 'Tamil Nadu': 33, 'Tripura': 16, 'Uttar Pradesh': 9, 'Uttarakhand': 5, 'West Bengal': 19, 'Odisha': 21, 'Dadara & Nagar Havelli': 26, 'Meghalaya': 17, 'Mizoram': 15, 'Nagaland': 13, 'NCT of Delhi': 7}

所以你可能已经看到了问题，Andaman and Nicobar Islands 在两者中都存在但拼写不同，就像字典中的 ' Andaman & Nicobar Island' 一样。这使得最后一列 NaN
9 10 Andaman and Nicobar Islands 219842.0 NaN

如何将其与 difflib 库结合使用？

我试过了

df_19_20['State/Union territory'] = df_19_20['State/Union territory'].apply(get_close_matches(df_19_20['State/Union territory'], id_d.keys()))

和

df_19_20['State/Union territory'] = get_close_matches(df_19_20['State/Union territory'], id_d.keys())

有什么我想念的吗？如何处理列以获得最佳匹配？

Answer 1

问题出在 df.apply

的应用程序中

df.apply 需要被赋予一个函数，该函数从它被迭代的每一行中获取值。您还需要清理 get_close_matches 的 return，其中 return 是 list，因此您需要取第一个元素

df_19_20['State/Union territory'].apply(lambda x: get_close_matches(x, id_d.keys())[0])

应该可以

当连接列有轻微的拼写差异时，如何将列合并/添加到 pandas 中的数据框？

How to merge/ add columns to dataframes in pandas when the joining column has slight spelling differences?

python

dataframe

sequencematcher

pandas