熊猫如何根据条件填充单元格

Question

I am building a backtester for my algo where i want it to fill a new column with 0 or 1 if its a buy or sell signal

the BUY signal is generated:
1-if two cells in two different columns exceeded a certain threshold [ Q5> 7 & Q7>7 ] :
fill a new col ['BuySignal'] = 1
else condition not met in rule one and rule two below fill BuySignal = 0

2-if there is two buy signals and yet there is no sell signal after the first buy signal, dont fill BuySignal with 1, keep it 0 ( as you can see in row 6 in yellow.) we can only have one buy signal until the next sell signal

the SELL signal is generated:

1-sell if only one cell value exceeded a certain threshold [TCN > 8]
PS: if at the same time both buy and sell conditions are met I want to label only BuySignal as 1 and Sell = 0


what i am expecting is : buy at row 1, ignore buy at row 6, sell at row 10, buy again at row 13
input data for your ref: https://drive.google.com/file/d/16qqrcvxJuTgd41233RNjwhaNVNREqWWe/view?usp=sharing

Answer 1

First of all, I think there is an error in your output table - according to rule 2 for buy signal, I think the buy signal should be also 0 in rows 1 and 13, since there is no sell signal in rows 2 and 14. Here's what I got based on your description, not sure if this is exactly what you meant, but defo a good starting point. To simplify the problem, I created two BuySignal columns - first shows the state after applying the first rule and second shows the state after applying the second rule to the first. You can just drop the first column in your final solution.

# buy rule 1
data['BuySignal1'] = (data['Q5']>7) & (data['Q7']>7)

# sell rule 1
data['SellSignal'] = data['TCN']>8

# buy rule 2
data['BuySignal2'] = (data['BuySignal1']==1)&(data['SellSignal'].shift(-1)==1)

# PS basically - if BuySignal==1 then SellSignal==0
data.loc[data['BuySignal2']==1,'SellSignal'] = 0

# change booleans to types (just for display)
cols = ['BuySignal1','BuySignal2','SellSignal']
data[cols] = data[cols].astype(int)

And this is the output: