0

I know this has been asked a few times, but I cannot get to solve it. I'm terribly sorry. I have a dataframe 'stock_df' with 1 column with 'RSI_14' values. I'm a newbie, so in order to make it easier I initialize 3 columns with == 0 for 3 states. This auxiliary columns are to explain better the question:

  • More than 70 --> 'plus70' = 1
  • Less than 30 --> 'minus30' = 1
  • Between 70 and 30 --> 'between' = 1

The objective is to make a 'Signal' that is +1 when 'RSI_14' > 70 and -1 when 'RSI_14' < 30 (that's easy), but the tricky part is that when the state is 'between' 70 and 30 I need to put the former +1 or -1 state in 'Signal', and keep that number until the next +1 or -1 change of state and keep with that on and on... This shouldn't be that difficult with .shift(1) or .diff(1), but I don't get it.

This is the desired outcome: Outcome

I've tried np.where, but it's the last "stock_df['Signal'].shift(1)" that doesn't seem to work:

stock_df['Signal'] = np.where(stock_df['RSI_14'] > 70, 1, (np.where(stock_df['RSI_14'] < 30, -1, stock_df['Signal'].shift(1))))

I think the solution must be in "groupby" with "transform" but I've tried many different ways, but I'm quite clumsy... I really think is with groupby. I've checked A LOT of answers here, but I don't get to solve it. I'd really appreciate your help. Thanks

Enrique de Juan
  • 1
  • Welcome to stack overflow. Please see [How to make good pandas examples](https://stackoverflow.com/questions/20109391/how-to-make-good-reproducible-pandas-examples) and [edit] your question to include a [mcve] with sample input and expected output in the text of your question, not as an image or link. This will help us to help you, as we can't copy and paste an image to help you find a solution – G. Anderson Jan 27 '22 at 19:54

2 Answers2

1

I don't think you need to create any additional columns to achieve this. Just use:

condition_values = stock_df.RSI_14.values
signal = []
last = None
for item in condition_values:
    if item < 30:
        signal.append(-1)
        last = -1
    elif item > 70:
        signal.append(1)
        last = 1
    else:
        signal.append(last)

 df['signal'] = signal

Change the none in the beginning to your liking value.

Olca Orakcı
  • 372
  • 3
  • 12
0

The following code snippet could work. Assigning signal value to be equal to the previous row signal value (when between equals 1) should get the job done.

for i in range(len(stock_df)):
    if stock_df['between'][i] == 1:
        if i == 0:
            stock_df['Signal'][i] = 1
        else:
            stock_df['Signal'][i] = stock_df['Signal'][i - 1]
Nur Imtiazul Haque
  • 383
  • 1
  • 2
  • 10