How do I create a Python function that creates values based on previous row's values?-CodePudding

I have a Pandas dataframe with 3 columns: Date, Return, and Result that has stock gains % in the Return column. The dates are set to daily and the index is reset.

The dataframe has hundreds of rows. I'm trying to fill in the current value of a row by using the previous value in the results column. I can do this manually like this:

df["Result"].iloc[0] = 100 * (1   df["Return"].iloc[0])

df["Result"].iloc[1] = df["Result"].iloc[0] * (1   df["Return"].iloc[1])

enter image description here

The problem is when I try to make this into a function and I've also tried using a lambda function with no results. Other resources oversimplify the examples. Can anyone help here?

This is one of several different iterations of an attempted function that didn't work.

def result_calc(df, i):
  df["Result"].iloc[i] = df["Result"].iloc[i-1] * (1   df["Return"].iloc[i])

---------------------------------------------------------------------------

TypeError                                 Traceback (most recent call last)
\<ipython-input-410-c6a07bcb23a7\> in \<module\>
1 df1 = buy_hold_df.copy()
\----\> 2 df1.iloc\[1:\]\["Result"\] = df1.iloc\[1:\]\["Result"\].apply(result_calc(df1, 1))

\~\\Anaconda3\\envs\\FFN\\lib\\site-packages\\pandas\\core\\series.py in apply(self, func, convert_dtype, args, \*\*kwds)
4106             else:
4107                 values = self.astype(object).\_values
\-\> 4108                 mapped = lib.map_infer(values, f, convert=convert_dtype)
4109
4110         if len(mapped) and isinstance(mapped\[0\], Series):

pandas_libs\\lib.pyx in pandas.\_libs.lib.map_infer()

TypeError: 'NoneType' object is not callable

CodePudding user response：

IIUC:

# pseudo code
result[0] = 100 * (1   return[0])
result[1] = result[0] * (1   return[1]) = 100 * (1   return[0]) * (1   return[1])
...
return[n] = result[n-1] * (1   return[n])
          = 100 * (1   return[0]) * (1   return[1]) * ... * (1   return[n])
          = 100 * cumprod(1   return)

So:

>>> import pandas as pd
>>> 
>>> df = pd.DataFrame(dict(
...     Date=["2019-01-02", "2019-01-03", "2019-01-04", "2019-01-07", "2019-01-08"],
...     Return=[.035039, .001354, .025693, .018128, .012625]
... ))
>>> df["Result"] = 100 * (df.Return   1).cumprod()
>>> df
         Date    Return      Result
0  2019-01-02  0.035039  103.503900
1  2019-01-03  0.001354  103.644044
2  2019-01-04  0.025693  106.306971
3  2019-01-07  0.018128  108.234103
4  2019-01-08  0.012625  109.600559