python - 将序号添加到 Pandas 中的 groupby().head(n) 表达式

Question

我在 Pandas 中有一个表达式，我按国家/地区对前三个值进行排序:

Country              | Value
---------------------|------
Germany              | 102.1
Germany              | 90.3
Germany              | 44.6
Switzerland          | 59.9
Switzerland          | 35.3
Switzerland          | 21.6

...and so on

这是我使用 df.groupby("Country").head(3)[["Country", "Value"]] 获得的。现在，我想附加第三列，将国家内的排名与值相关联:

Country              | Value  | Rank
---------------------|--------|------
Germany              | 102.1  | 1
Germany              | 90.3   | 2
Germany              | 44.6   | 3
Switzerland          | 59.9   | 1
Switzerland          | 35.3   | 2
Switzerland          | 21.6   | 3

...and so on

我最好怎么做？

Answer 1

我相信你需要GroupBy.rank和 method='dense' 的排名总是按 Value 列的排序值在组之间增加 1，并转换为 integers:

df['Rank'] = df.groupby("Country")["Value"].rank(method='dense', ascending=False).astype(int)
print (df)
       Country  Value  Rank
0      Germany  102.1     1
1      Germany   90.3     2
2      Germany   44.6     3
3  Switzerland   59.9     1
4  Switzerland   35.3     2
5  Switzerland   21.6     3

如果需要计数器，那么最好使用 GroupBy.cumcount :

df['Rank1'] = df.groupby("Country").cumcount() + 1

差异在变化的数据中最明显:

print (df)
       Country  Value
0      Germany   90.3 second largest per group - 2
1      Germany  102.1 largest per group - 1
2      Germany   44.6 third largest per group - 3
3  Switzerland   21.6
4  Switzerland   35.3
5  Switzerland   59.9

df['Rank'] = df.groupby("Country")["Value"].rank(method='dense', ascending=False).astype(int)
df['Rank1'] = df.groupby("Country").cumcount() + 1

print (df)
       Country  Value  Rank  Rank1
0      Germany   90.3     2      1
1      Germany  102.1     1      2
2      Germany   44.6     3      3
3  Switzerland   21.6     3      1
4  Switzerland   35.3     2      2
5  Switzerland   59.9     1      3