This code works inside a for loop where 'i' is iterating it through the dataframe rows and occasionally this code line is used to reset all the rows preceeding row 'i', in column 'atr_Lts', so my cummulation calcs work as I require.
cummax.loc[:i, 'atr_Lts'] = cummax.loc[cummax.index[i], 'atr_Lts']
It works fine, but it is giving me the following warning, and I cannot find a better approach.
FutureWarning: Slicing a positional slice with .loc is not supported, and will raise TypeError in a future version. Use .loc with labels or .iloc with positions instead.
I need to be able to set all rows in column 'atr_Lts' up to the current iteration i, to the current 'atr_Lts' value. To get at that I need to use row selection (:i) and column label ('atr_Lts') but according to the warning I can not combine the approach.
In many searches this was the recommended method, but is going to be obselete so wondered what the better approach is.
EDIT: copy of example cummax dataframe print
Date Close atr atr_Lts atr_Sts atrts
2022-02-23 421.950012 NaN 371.614245 413.538855 413.538855
2022-02-24 428.299988 NaN 371.614245 413.538855 413.538855
2022-02-25 437.750000 NaN 371.614245 413.538855 413.538855
2022-02-28 436.630005 NaN 371.614245 413.538855 413.538855
2022-03-01 429.980011 NaN 371.614245 413.538855 413.538855
this is the print(cummax.info())
Data columns (total 5 columns):
# Column Non-Null Count Dtype
--- ------ -------------- -----
0 Close 125 non-null float64
1 atr 112 non-null float64
2 atr_Lts 125 non-null float64
3 atr_Sts 112 non-null float64
4 atrts 112 non-null float64
dtypes: float64(5)
CodePudding user response：
This warning is raised because .loc requires a label based index, you passed and integer based index: .loc[:i, <some columns>] (i is an int), therefore causing pandas to be confused.
To make this warning go away, you can either try:
- Ignore the warnings using the
warningsmodule. - Typing this instead:
cummax.iloc[:i].loc[:, 'atr_Lts'] = ....