CODEWITHSUNDEEP
pythonpandas
Rockport Redfish
Rockport Redfish

Reputation: 109

Python Pandas - If statement checking a True False Boolean Column

I have a Pandas df with a column of True False values. I am trying to construct an if statement that tests that column, but am not getting the desired result. I think I am using the .bool method incorrectly. The basic idea is to check if the value of the current row Col1is True, and if any of the three prior rows Col1 was False, return True in Col2

from pandas import DataFrame
 
names = {'col1': [False, False, False, False, False, True, True, 
     True, False, False]}
df = DataFrame(names, columns =['col1'])

if df.col1.bool == True:
   if df.col1.shift(1).bool == False:
    df['col2'] = True
  elif df.col1.shift(2).bool == False:
    df['col2'] = True
  elif df.col1.shift(3).bool == False:
    df['col2'] = True
  else:
    df['col2'] = False

df

Upvotes: 0

Views: 3010

Answers (3)

clcto
clcto

Reputation: 9648

The lines df['col2'] = True and df['col2'] = False set the whole column to True and False, respectively. Since you want element-wise operations, you need to use the overloaded bitwise operations & (for AND) and | (for OR).

Your new column should be true when the current value of col1 is True AND at least one of the previous 3 values are False, which could be encoded as:

df['col2'] = df.col1 & (
    (df.col1.shift(1) == False) |
    (df.col1.shift(2) == False) |
    (df.col1.shift(3) == False)
)

Be careful with the operator precedence when using the bitwise operations as they have lower priority than the comparison operators. This can often lead to subtle bugs producing the wrong results but do not cause errors. I advise to use extra parenthesis in these expressions.

Upvotes: 1

Naveed
Naveed

Reputation: 11650

here is one way to do it, using np.where and pd.rolling and by taking the sum of boolean values, which would be less than 3 unless all previous three values are true IIUC, the previous three, excluding the current row

df['col2']=np.where((df['col1'].shift(1).rolling(3).sum()<3) &(df['col1']==True),
                    True,
                    False)
df

    col1    col2
0   False   False
1   False   False
2   False   False
3   False   False
4   False   False
5   True    True
6   True    True
7   True    True
8   False   False
9   False   False

Upvotes: 1

Christian Trujillo
Christian Trujillo

Reputation: 522

from pandas import DataFrame

names = {'col1': [False, False, False, False, False, True, True, 
 True, False, False]}
df = DataFrame(names, columns =['col1'])
df['col2'] = False
df['col2'][df['col1']==False] = True
df['col2'][df['col1'].shift(1)==False] = True
df['col2'][df['col1'].shift(2)==False] = True
df['col2'][df['col1'].shift(3)==False] = True

df

or if you want to compact it a bit

from pandas import DataFrame

names = {'col1': [False, False, False, False, True, True, True, 
 True, False, False]}
df = DataFrame(names, columns =['col1'])
df['col2'] = False
df['col2'][(df['col1']==False) | (df['col1'].shift(1)==False) | (df['col1'].shift(2)==False) | (df['col1'].shift(3)==False)] = True

Upvotes: 0

Related Questions