问题描述
从此：
+------+------+--------------------------+-----------------+
| code | type | name | final_component |
+------+------+--------------------------+-----------------+
| C001 | ACT | Exhaust Blower Drive | |
| C001 | AL | | |
| C001 | AL | | |
| C001 | SET | Exhaust Blower Drive | |
| C001 | AL | | |
| C001 | AL | | |
| C001 | AL | | |
| C002 | ACT | Spray Pump Motor 1 Pump | |
| C002 | SET | Spray Pump Motor 1 Pump | |
| C003 | ACT | Spray Pump Motor 2 Pump | |
| C003 | SET | Spray Pump Motor 2 Pump | |
| C004 | ACT | Spray Pump Motor 3 Pump | |
| C004 | SET | Spray Pump Motor 3 Pump | |
+------+------+--------------------------+-----------------+
预期：
+------+------+--------------------------+--------------------------+
| code | type | name | final_component |
+------+------+--------------------------+--------------------------+
| C001 | ACT | Exhaust Blower Drive | Exhaust Blower Drive |
| C001 | AL | | Exhaust Blower Drive |
| C001 | AL | | Exhaust Blower Drive |
| C001 | SET | Exhaust Blower Drive | Exhaust Blower Drive |
| C001 | AL | | Exhaust Blower Drive |
| C001 | AL | | Exhaust Blower Drive |
| C001 | AL | | Exhaust Blower Drive |
| C002 | ACT | Spray Pump Motor 1 Pump | Spray Pump Motor 1 Pump |
| C002 | SET | Spray Pump Motor 1 Pump | Spray Pump Motor 1 Pump |
| C003 | ACT | Spray Pump Motor 2 Pump | Spray Pump Motor 2 Pump |
| C003 | SET | Spray Pump Motor 2 Pump | Spray Pump Motor 2 Pump |
| C004 | ACT | Spray Pump Motor 3 Pump | Spray Pump Motor 3 Pump |
| C004 | SET | Spray Pump Motor 3 Pump | Spray Pump Motor 3 Pump |
+------+------+--------------------------+--------------------------+
对于所有相同的代码，我必须将类型为“ SET”的名称值复制到final_component 像C001一样，“ SET”类型的名称是“排气鼓风机” 我必须将其复制到所有C001的final_component
for ind in dataframe.index:
if dataframe['final_component'][ind]!=None:
temp = dataframe['final_component'][ind]
temp_code = dataframe['code'][ind]
i = ind
while dataframe['code'][i] == temp_code:
dataframe['final_component'][ind] = temp
i+=1
我可以提出这个 但它陷入了while循环
解决方法
解决方案1：将数据按顺序分组
如果'name'字段中的数据已具有Null值，则可以执行一些简单的操作，例如ffill（）。 Pandas dataframe.ffill（）函数用于填充数据框中的缺失值。 “填充”代表“向前填充”，并将向前传播最后一个有效观察值。在这种情况下，它不会考虑code中的值。如果您也想考虑这一点，请查看解决方案2。
import pandas as pd
import numpy as np
a = {'code':['C001']*7+['C002']*2+['C003']*2+['C004']*2,'typ':['ACT','AL','SET','ACT','SET'],'name':['Exhaust Blower Drive',None,'Exhaust Blower Drive',np.nan,'Spray Pump Motor 1 Pump','Spray Pump Motor 2 Pump','Spray Pump Motor 3 Pump','Spray Pump Motor 3 Pump']}
df = pd.DataFrame(a)
#copy all the values from name to final_component' with ffill()
#it will fill the values where data does not exist
#this will work only if you think all values above are part of the same set
df['final_component'] = df['name'].ffill()
解决方案2：何时数据必须基于另一个列值
如果需要基于代码中的值进行填充，则可以使用以下解决方案。
您可以进行查找，然后更新值。尝试这样的事情。
import pandas as pd
import numpy as np
a = {'code':['C001']*7+['C002']*2+['C003']*2+['C004']*2,'Spray Pump Motor 3 Pump']}
df = pd.DataFrame(a)
#copy all the values from name to final_component' including nulls
df['final_component'] = df['name']
#create a sublist of items based on unique values in code
lookup = df[['code','final_component']].groupby('code').first()['final_component']
#identify all the null values that need to be replaced
noname=df['final_component'].isnull()
#replace all null values with correct value based on lookup
df['final_component'].loc[noname] = df.loc[noname].apply(lambda x: lookup[x['code']],axis=1)
print(df)
输出将如下所示：
code typ name final_component
0 C001 ACT Exhaust Blower Drive Exhaust Blower Drive
1 C001 AL NaN Exhaust Blower Drive
2 C001 AL NaN Exhaust Blower Drive
3 C001 SET Exhaust Blower Drive Exhaust Blower Drive
4 C001 AL NaN Exhaust Blower Drive
5 C001 AL NaN Exhaust Blower Drive
6 C001 AL NaN Exhaust Blower Drive
7 C002 ACT Spray Pump Motor 1 Pump Spray Pump Motor 1 Pump
8 C002 SET Spray Pump Motor 1 Pump Spray Pump Motor 1 Pump
9 C003 ACT Spray Pump Motor 2 Pump Spray Pump Motor 2 Pump
10 C003 SET Spray Pump Motor 2 Pump Spray Pump Motor 2 Pump
11 C004 ACT Spray Pump Motor 3 Pump Spray Pump Motor 3 Pump
12 C004 SET Spray Pump Motor 3 Pump Spray Pump Motor 3 Pump
这是一种方法。首先，重新创建数据框：
from io import StringIO
import pandas as pd
data = '''| code | type | name | final_component |
| C001 | ACT | Exhaust Blower Drive | |
| C001 | AL | | |
| C001 | AL | | |
| C001 | SET | Exhaust Blower Drive | |
| C001 | AL | | |
| C001 | AL | | |
| C001 | AL | | |
| C002 | ACT | Spray Pump Motor 1 Pump | |
| C002 | SET | Spray Pump Motor 1 Pump | |
| C003 | ACT | Spray Pump Motor 2 Pump | |
| C003 | SET | Spray Pump Motor 2 Pump | |
| C004 | ACT | Spray Pump Motor 3 Pump | |
| C004 | SET | Spray Pump Motor 3 Pump | |
'''
df = pd.read_csv(StringIO(data),sep='|',)
df = df.drop(columns=['Unnamed: 0','Unnamed: 5'])
现在，删除前导和尾随空格：
# remove leading / trailing spaces
df.columns = [c.strip() for c in df.columns]
for col in df.columns:
if df[col].dtype == object:
df[col] = df[col].str.strip()
并填充final_component：
# populate 'final component'
df['final_component'] = df['name']
现在用None替换空字符串并使用ffill()
# find final component that is empty string...
mask = df['final_component'] == ''
# ... and convert to None...
df.loc[mask,'final_component'] = None
# ...so we can use ffill()
df['final_component'] = df['final_component'].ffill()
print(df)
code type name final_component
0 C001 ACT Exhaust Blower Drive Exhaust Blower Drive
1 C001 AL Exhaust Blower Drive
2 C001 AL Exhaust Blower Drive
3 C001 SET Exhaust Blower Drive Exhaust Blower Drive
4 C001 AL Exhaust Blower Drive
5 C001 AL Exhaust Blower Drive
6 C001 AL Exhaust Blower Drive
7 C002 ACT Spray Pump Motor 1 Pump Spray Pump Motor 1 Pump
8 C002 SET Spray Pump Motor 1 Pump Spray Pump Motor 1 Pump
9 C003 ACT Spray Pump Motor 2 Pump Spray Pump Motor 2 Pump
10 C003 SET Spray Pump Motor 2 Pump Spray Pump Motor 2 Pump
11 C004 ACT Spray Pump Motor 3 Pump Spray Pump Motor 3 Pump
12 C004 SET Spray Pump Motor 3 Pump Spray Pump Motor 3 Pump