Python-基于2个标准（案例和时间序列）组合2个数据库

我想使用以下数据库（1）和（2）将两个数据帧组合在一起。 预期结果如（3）所示

1)

---

(二)

3） 决赛

以下是我正在寻找的标准：

1.）根据“案例”进行匹配

2.“Date_Y”必须=<“Date_X”，并且必须是“case”数据框中的最大日期

3.）显示与日期相对应的值

我试着寻找类似的代码，但找不到

提前感谢您的帮助。

以下是您想要的：

import numpy as np
import pandas as pd


df1 = pd.DataFrame({'case':[123,123,345,456],'date':['2019-12-21','2019-12-17','2019-12-21','2019-12-21']})
df2 = pd.DataFrame({'case':[123,123,123],'date':['2019-12-21','2019-12-18','2019-12-15'],'value':[0.4,0.5,1.2]})

df1['date'] = pd.to_datetime(df1['date'])
df2['date'] = pd.to_datetime(df2['date'])

datey = []
val = []
for i in range(len(df1)):
#checking case
    tmp=df2[df2['case']==df1['case'][i]]
#comparing date
    tmp = tmp[tmp['date']<=df1['date'][i]]
    if(len(tmp)>0):
        tmp = tmp.loc[tmp['date'].idxmax()]
#appending to lists
        datey.append(tmp.date)
        val.append(tmp.value)
    else:
        datey.append(None)
        val.append(None)

df1['date_y'] = datey
df1['value'] = val

print(df1)

您可以删除Nan值。

以下是您想要的：

import numpy as np
import pandas as pd


df1 = pd.DataFrame({'case':[123,123,345,456],'date':['2019-12-21','2019-12-17','2019-12-21','2019-12-21']})
df2 = pd.DataFrame({'case':[123,123,123],'date':['2019-12-21','2019-12-18','2019-12-15'],'value':[0.4,0.5,1.2]})

df1['date'] = pd.to_datetime(df1['date'])
df2['date'] = pd.to_datetime(df2['date'])

datey = []
val = []
for i in range(len(df1)):
#checking case
    tmp=df2[df2['case']==df1['case'][i]]
#comparing date
    tmp = tmp[tmp['date']<=df1['date'][i]]
    if(len(tmp)>0):
        tmp = tmp.loc[tmp['date'].idxmax()]
#appending to lists
        datey.append(tmp.date)
        val.append(tmp.value)
    else:
        datey.append(None)
        val.append(None)

df1['date_y'] = datey
df1['value'] = val

print(df1)

---

您可以删除Nan值。

是否使用熊猫？你必须这么做吗？你的数据集有多大？到目前为止你尝试了什么？检查我上传的解决方案。你使用熊猫吗？你必须这么做吗？你的数据集有多大？到目前为止你尝试了什么？检查我上传的解决方案。

Case  Date_X     Date_Y      Value 
123 2019-12-21 2019-12-20   0.4
123 2019-12-16 2019-12-14   1.2
234 2019-12-21
345 2019-12-21

import numpy as np
import pandas as pd


df1 = pd.DataFrame({'case':[123,123,345,456],'date':['2019-12-21','2019-12-17','2019-12-21','2019-12-21']})
df2 = pd.DataFrame({'case':[123,123,123],'date':['2019-12-21','2019-12-18','2019-12-15'],'value':[0.4,0.5,1.2]})

df1['date'] = pd.to_datetime(df1['date'])
df2['date'] = pd.to_datetime(df2['date'])

datey = []
val = []
for i in range(len(df1)):
#checking case
    tmp=df2[df2['case']==df1['case'][i]]
#comparing date
    tmp = tmp[tmp['date']<=df1['date'][i]]
    if(len(tmp)>0):
        tmp = tmp.loc[tmp['date'].idxmax()]
#appending to lists
        datey.append(tmp.date)
        val.append(tmp.value)
    else:
        datey.append(None)
        val.append(None)

df1['date_y'] = datey
df1['value'] = val

print(df1)

   case       date     date_y  value
0   123 2019-12-21 2019-12-21    0.4
1   123 2019-12-17 2019-12-15    1.2
2   345 2019-12-21        NaT    NaN
3   456 2019-12-21        NaT    NaN