Pandas数据框使用分区数据添加新的计算字段
Pandas数据框使用分区数据添加新的计算字段
提问于 2019-03-15 08:37:55
我正在尝试添加一个新的计算字段。我正在尝试Adding calculated column(s) to a dataframe in pandas的第二个最好的答案，因为在我看来它是最好的，因为它很整洁。请随时提供更好的替代方案。
无论哪种方式，我的初始代码如下：
import pandas as pd
#https://github.com/sivabalanb/Data-Analysis-with-Pandas-and-Python/blob/master/nba.csv
dt_nba = pd.read_csv("data//nba.csv")
#note this is just basic function. I want to pass partitioned data like team's average salary
def GetSalaryIncrement(val):
return val * 1.1
dt_nba["SalaryPlus10Percent"] = map(GetSalaryIncrement,dt_nba["Salary"])
dt_nba[["Name","Team","Salary","SalaryPlus10Percent"]][:5]然而，结果并不是我所期望的：
+----+---------------+----------------+--------------+--------------------------------+
| ID | Name | Team | Salary | SalaryPlus10Percent |
+----+---------------+----------------+--------------+--------------------------------+
| 0 | Avery Bradley | Boston Celtics | 7730337.0000 | <map object at 0x7fb819e9b7b8> |
| 1 | Jae Crowder | Boston Celtics | 6796117.0000 | <map object at 0x7fb819e9b7b8> |
| 2 | John Holland | Boston Celtics | nan | <map object at 0x7fb819e9b7b8> |
| 3 | R.J. Hunter | Boston Celtics | 1148640.0000 | <map object at 0x7fb819e9b7b8> |
| 4 | Jonas Jerebko | Boston Celtics | 5000000.0000 | <map object at 0x7fb819e9b7b8> |
+----+---------------+----------------+--------------+--------------------------------+特别是，我对传递“窗口/聚合数据”感兴趣，因为它应该优雅地忽略NAN值。
在T-SQL中，我可以这样做：
-- INCREASE EACH PLAYERS SALARY BY 10% OF AVERAGE SALARY OF THE TEAM
SELECT NewSalary= Salary + (.1 * AVG(Salary) OVER (PARTITION BY Team))
FROM nba_data如果可能的话，我想在Pandas中做到这一点。谢谢。
回答 1
Stack Overflow用户
回答已采纳
发布于 2019-03-15 08:44:05
我想你要找的是
dt_nba["Salary"]=dt_nba["Salary"].map(GetSalaryIncrement)此外，您还可以使用
GetSalaryIncrement(dt_nba["Salary"])dt_nba["Salary"].apply(GetSalaryIncrement) 到计算的INCREASE EACH PLAYERS SALARY BY 10% OF AVERAGE SALARY OF THE TEAM
dt_nba['Newsa']=dt_nba.groupby('Team')['Salary'].transform('mean')*0.1+dt_nba["Salary"]页面原文内容由Stack Overflow提供。腾讯云小微IT领域专用引擎提供翻译支持
原文链接：
https://stackoverflow.com/questions/55173994
复制相关文章
相似问题