python - use data from two or more columns when using .map to apply a function -

i want calculate difference between 2 dates in format yyyy-mm-dd. when write df["diff"] = df["date1"] - df["date2"] value 75 days can't sort.

how can write such map function calc_diff(date1,date2) where

def calc_diff(date1,date2): x = float(date2-date1) return x 

i.e. df["diff"] = df["date1"],df["date2"].map(calc_diff)

when write following, get:

df["until_payable"] = map(calc_diff, df["ex_div_date"], "date_payable")  def calc_diff(d1,d2):     y = pd.to_datetime(d1)     x = pd.to_datetime(d2)     return x-y 

error:

<ipython-input-167-b151c936730f> in calc_diff(d1, d2)      39     y = pd.to_datetime(d1)      40     x = pd.to_datetime(d2) ---> 41     return x-y      42       43 #convert data in columns data types  pandas/tslib.pyx in pandas.tslib._timestamp.__sub__ (pandas/tslib.c:17620)()  typeerror: descriptor '__sub__' requires 'datetime.datetime' object received 'str' 

if understood correctly solution problem is:

res = map(calc_diff, date1, date2) 

please note date1 , date2 should same length