python - Wrong result with moving average cross over

Question

I have a dataframe with the following values. I want to dedect dates, when the Low is crossing below the SMA8Low.

  df['SMA10High'] = df.loc[:,'High'].rolling(window=10).mean()
  df['SMA8Low'] = df.loc[:,'Low'].rolling(window=8).mean()
  df['yesterday'] = df['Low'].shift(1) >= df['SMA8Low'].shift(1) #yesterdays 
  LOW >= SMA8Low
  df['today'] = df['Low'] <= df['SMA8Low']  #Todays Low <= SMA8Low
  df.dropna(inplace=True) #droping NaN values

Dataframe result after the above code is the following:

Now the new column crossing should only be True, if yesterday is true and today is true.

I tried the following code:

df['crossover'] = np.where(df['yesterday'] == df['today'],True,False)
#df['crossover'] = np.where(df['today'] & (df['yesterday'] == 
df['today']),True,True)

But the result is wrong. Only the following dates should be True in the column crossover: 2021-01-08 and 2021-01-22

question from:https://stackoverflow.com/questions/65847773/wrong-result-with-moving-average-cross-over

Answer 1

Just add another condition to np.where:

df['crossover'] = np.where(df['today'] & (df['yesterday'] == df['today']),True,True)