Python 如何使用pandas中的另一列更新列

我正试图创建一个数据框架，跟踪2010-2016年间公立学校的开学数量

StatusType  County  2010 ...2016    OpenYear    ClosedYear
1   Closed  Alameda 0        0        2005        2015.0
2   Active  Alameda 0        0        2006         NaN
3   Closed  Alameda 0        0        2008        2015.0
4   Active  Alameda 0        0        2011         NaN
5   Active  Alameda 0        0        2011         NaN
6   Active  Alameda 0        0        2012         NaN
7   Closed  Alameda 0        0        1980        1989.0
8   Active  Alameda 0        0        1980         NaN
9   Active  Alameda 0        0        1980         NaN

我想更新2010-2016专栏，以跟踪每年开学的学校数量。例如，数据框架中的第一所学校于2005年开学，2015年关闭。迭代器应检查“ClosedYear”列，并将1添加到所有列的“rows”值<2015（20102011…，2014）。如果“ClosedYear”列显示“NaN”，则从“OpenYear”列中的年份开始，将1添加到所有列的“行”值>=“OpenYear”（例如：学校#4，列[20112012…，2016]+1和列[2010]无变化）

我正在考虑使用“apply”将函数应用于数据帧。但这可能不是解决问题的最有效方法。需要帮助找出如何使这项工作！谢谢

额外步骤： 完成计数后，我想按县对年份列进行分组。我倾向于使用“groupby”w/sum函数来汇总每个县每年的开放学校数量。如果有人能在回答上述问题的同时补充这一点，那将非常有帮助

预期产出：

StatusType       County 2010 ...2016    OpenYear    ClosedYear
    1   Closed  Alameda 1        0        2005        2015.0
    2   Active  Alameda 1        1        2006         NaN
    3   Closed  Alameda 1        0        2008        2015.0
    4   Active  Alameda 0        1        2011         NaN
    5   Active  Alameda 0        1        2011         NaN
    6   Active  Alameda 0        1        2012         NaN
    7   Closed  Alameda 0        0        1980        1989.0
    8   Active  Alameda 1        1        1980         NaN
    9   Active  Alameda 1        1        1980         NaN

---

我觉得应该有一种方法可以做到这一点，而不必使用

for loop

，但是，我想不出它是atm，所以我的解决方案是：

# Read Example data
from io import StringIO # This only works python 3+
df = pd.read_fwf(StringIO(
"""StatusType  County    OpenYear    ClosedYear
Closed      Alameda   2005        2015.0
Active      Alameda   2006         NaN
Closed      Alameda   2008        2015.0
Active      Alameda   2011         NaN
Active      Alameda   2011         NaN
Active      Alameda   2012         NaN
Closed      Alameda   1980        1989.0
Active      Alameda   1980         NaN
Active      Alameda   1980         NaN"""))

# For each year
for year in range(2010, 2016+1):
    # Create a column of 0s
    df[str(year)] = 0
    # Where the year is between OpenYear and ClosedYear (or closed year is NaN) set it to 1
    df.loc[(df['OpenYear'] <= year) & (pd.isna(df['ClosedYear']) | (df['ClosedYear'] >= year)), str(year)] = int(1)

print(df.to_string)

---

（注：我不太清楚你想用

groupby

做什么）

除非真的需要创建这些中间列，否则你可以直接用

groupby

和

.size

获取计数，这取决于你是否想包括截止年份，把不平等从

改为
你能增加一个你预期产出的例子吗？我相信你的
小组会把一所从
2012
开到
NaN
的学校算作历年都在开的学校。不用担心，我花了30次努力才把它们弄对。
df=df.groupby（['County']）.sum（）
pub\u s.loc['OpenYear']=year]），year]+=int（1）
  StatusType   County  OpenYear  ClosedYear  2010  2011  2012  2013  2014  2015  2016
0     Closed  Alameda      2005      2015.0     1     1     1     1     1     1     0
1     Active  Alameda      2006         NaN     1     1     1     1     1     1     1
2     Closed  Alameda      2008      2015.0     1     1     1     1     1     1     0
3     Active  Alameda      2011         NaN     0     1     1     1     1     1     1
4     Active  Alameda      2011         NaN     0     1     1     1     1     1     1
5     Active  Alameda      2012         NaN     0     0     1     1     1     1     1
6     Closed  Alameda      1980      1989.0     0     0     0     0     0     0     0
7     Active  Alameda      1980         NaN     1     1     1     1     1     1     1
8     Active  Alameda      1980         NaN     1     1     1     1     1     1     1

  StatusType   County  OpenYear  ClosedYear
1     Closed  Alameda      2005      2015.0
2     Active  Alameda      2006         NaN
3     Closed  Alameda      2008      2015.0
4     Active  Alameda      2011         NaN
5     Active  Alameda      2011         NaN
6     Active  Alameda      2012         NaN
7     Closed  Alameda      1980      1989.0
8     Active  Alameda      1980         NaN
9     Active  Alameda      1980         NaN

import pandas as pd
year_list = [2010, 2011, 2012, 2013, 2014, 2015, 2016]
df_list = []

for year in year_list:
    group = ((df.ClosedYear.isnull()) | (df.ClosedYear >= year)) & (df.OpenYear <= year)
    n_schools = df.groupby([group, df.County]).size()[True]
    df_list.append(pd.DataFrame({'n_schools':n_schools, 'year': year}))

ndf = pd.concat(df_list)
#         n_schools  year
#County                  
#Alameda          5  2010
#Alameda          7  2011
#Alameda          8  2012
#Alameda          8  2013
#Alameda          8  2014
#Alameda          8  2015
#Alameda          6  2016