If the value of a specific column from a previous date is greater than zero then True, less than or-CodePudding

,clock_now,competition,market_name,lay,per_day
0,2022-12-30,A,B,-1.0,-1.0
1,2022-12-31,A,B,1.28,0.28
2,2023-01-01,A,B,-1.0,-0.72
3,2023-01-02,A,B,1.0,0.28
4,2023-01-03,A,B,1.0,1.28
5,2023-01-04,A,B,-1.0,-1.72
6,2023-01-04,A,B,-1.0,-1.72
7,2023-01-04,A,B,-1.0,-1.72

The idea is to get the value of the per_day column of the previous date closest to the row being analyzed.

For example:

In the lines with the date 2023-01-04, check in any of the lines that have the date 2023-01-03 which is the value of the column per_day, if it is greater than zero, True, if it is less or equal to zero, False.

The list would look like this:

False
False
True
False
True
True
True
True

My attempt:

df.clock_now = pd.to_datetime(df.clock_now)
df['invest'] = np.where(df.loc[df.clock_now == df['clock_now'] - timedelta(days=1),'per_day'].values[0] > 0,True,False)

But they all return False and there is another problem, it is not sure that the date will always be 1 day ago, it could be 2 or more, so it would still be a failed option.

How should I proceed in this case?

CodePudding user response：

Use groupby to group together rows with the same date, aggregate with last to get last value, then shift by one to assign each date the value corresponding to previous date, then join back with original dataframe:

df.join(df.groupby('clock_now').last()['per_day'].shift(1).rename('prev_day') > 0, on='clock_now')

This will output

   Unnamed: 0   clock_now competition market_name   lay  per_day  prev_day
0           0  2022-12-30           A           B -1.00    -1.00     False
1           1  2022-12-31           A           B  1.28     0.28     False
2           2  2023-01-01           A           B -1.00    -0.72      True
3           3  2023-01-02           A           B  1.00     0.28     False
4           4  2023-01-03           A           B  1.00     1.28      True
5           5  2023-01-04           A           B -1.00    -1.72      True
6           6  2023-01-04           A           B -1.00    -1.72      True
7           7  2023-01-04           A           B -1.00    -1.72      True