How to sum rows from a CSV only when not all values are above zero and not all below zero?-CodePudding

For me to analyze when all 5 values are greater than zero, I use it like this:

import pandas as pd

df = pd.read_csv(csv_file)

print(df.loc[
    (df['pressao1'] > 0) & 
    (df['pressao2'] > 0) & 
    (df['pressao3'] > 0) & 
    (df['pressao4'] > 0) & 
    (df['pressao5'] > 0), 
    'PROFIT/LOSS'
    ].sum())

But what if I want to calculate values when not all values are greater than zero and not all values are less than zero? Only when mixed values exist.

Example:

1,2,3,4,5 → SKIP
-1,2,3,7,-1 → SUM
-1,-2,-3,-4,-5 → SKIP

I tried using not all and not all but it was not accepted:

print(df.loc[
    not all(value>0 for value in [df['pressao1'],df['pressao2'],df['pressao3'],df['pressao4'],df['pressao5']]) 
    and 
    not all(value<0 for value in [df['pressao1'],df['pressao2'],df['pressao3'],df['pressao4'],df['pressao5']]), 
    'PROFIT/LOSS'
    ].sum())

But it gives me the error:

ValueError: The truth value of a Series is ambiguous. Use a.empty, a.bool(), a.item(), a.any() or a.all().

CodePudding user response：

You can use (so we only run the filter once):

cols = df.filter(regex=r'^pressao[1-5]$')

Then use that to index into your DF:

wanted = df.loc[~cols.eq(0).any(1) & cols.lt(0).any(1) & cols.gt(0).any(1), 'PROFIT/LOSS'].sum()