python - 怎么对dataframe中筛选过的数据进行计算

Question

要求：对已生成的一组dataframe数据列，需要根据日期做条件判断，对不同日期范围的数据，以不同的规则进行数据类运算，运算结果添加至一个新的列。

...
a=pd.DataFrame(data1,columns=['end_date','province','price']) #数组a
b=pd.DataFrame(data2,columns=['province','index_2013','index_2014','index_2015','index_2016'])#数据b
c=pd.merge(a,b,on='province') #以province进行连接

if a['end_date']<= datetime.strptime('2013-12-31','%Y-%m-%d').date(): #判断日期小于13年

c['axp']=c['price']c['index_2013'] #数据c添加列，数据值=priceindex_2013
print c

对if语句行报错，报错如下：ValueError: The truth value of a Series is ambiguous. Use a.empty, a.bool(), a.item(), a.any() or a.all().

在报错行，条件后添加.any()或者.all()则报错消失，但得到的数据不是想要的结果如：if a['end_date'].any()<= datetime。。。 则不报错

Answer 1

You can use the apply function to perform horizontal or vertical calculations on the dataframe object
For example, suppose you do an if on the first column of your c and calculate the quotient of the last two columns:

def cal(x):
    if x[0] > 10:  # 对第一列筛选
        return x[-1] / x[-2]
        
c.apply(cal, axis=1)

At this point, you can get the calculation result that meets the filtering conditions, which is a Series object

Answer 2

This is a basic Python question.

if a['end_date']<= datetime.strptime('2013-12-31','%Y-%m-%d').date()

The vectorized calculation of a Series still returns the Series. The Series is an np array and does not provide a method to return a true value. Comparison is possible only after using .any(),.all().

Then study the Pandas documentation carefully.