CODEWITHSUNDEEP
pythonpandasdataframe
SLE
SLE

Reputation: 85

Replace cells in row if condition met

Is there a way to for Less, Middle Greater columns. If it is greater than 0 to replace the ROW with 1. So the top row would become?

                  Conc      Less    Middle   Greater
Date                                                
2005-03-02 00:00  10.3  0.000000  1         1

This is the original

                  Conc      Less    Middle   Greater
Date                                                
2005-03-02 00:00  10.3  0.000000  0.083333  0.916667

2005-03-02 01:00  14.1  0.000000  0.750000  0.250000

2005-03-02 02:00   7.0  0.000000  0.833333  0.166667

2005-03-02 03:00   7.0  0.000000  1.000000  0.000000

2005-03-02 04:00   7.2  0.000000  1.000000  0.000000

2005-03-02 06:00   6.6  0.333333  0.666667  0.000000

2005-03-02 07:00   6.6  0.416667  0.583333  0.000000

i've tried:

df.loc[df['Less']>0:]=1
df.loc[df['Less']==0:]=0

but that shows up in red and says False True (In the correct places) and: dtype: bool, None, None)

I also tried: looping through like:

for line in df['Less']:
    if df['Less'] >0:
        df['Less']=1

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

Upvotes: 1

Views: 1691

Answers (2)

A.Kot
A.Kot

Reputation: 7903

To do them all at once:

columns = ['Less', 'Middle', 'Greater']
df[columns] = np.where(df[columns] >0, 1 ,0)

Or to do it individually (Note double brackets for selecting columns):

df[['Less']] = np.where(df[['Less']] >0, 1 ,0)
df[['Middle']] = np.where(df[['Middle']] >0, 1 ,0)
df[['Greater']] = np.where(df[['Greater']] >0, 1 ,0)

Upvotes: 0

EdChum
EdChum

Reputation: 394041

You can use loc with boolean condition:

In [250]:
df.loc[df['Less'] > 0, 'Less'] = 1
df

Out[250]:
                     Conc  Less    Middle   Greater
Date                                               
2005-03-02 00:00:00  10.3   0.0  0.083333  0.916667
2005-03-02 01:00:00  14.1   0.0  0.750000  0.250000
2005-03-02 02:00:00   7.0   0.0  0.833333  0.166667
2005-03-02 03:00:00   7.0   0.0  1.000000  0.000000
2005-03-02 04:00:00   7.2   0.0  1.000000  0.000000
2005-03-02 06:00:00   6.6   1.0  0.666667  0.000000
2005-03-02 07:00:00   6.6   1.0  0.583333  0.000000

this df.loc[df['Less']>0:] is invalid syntax, you want to use a comma and pass the list of column names of interest

Your for loop version:

for line in df['Less']:
    if df['Less'] >0:
        df['Less']=1

is invalid because if doesn't understand how to interpret an array of boolean values hence the error, if you did if (df['Less'] >0).all() or if (df['Less'] >0).any() then it would be happy but it doesn't make sense anyway as you're iterating row-wise but then testing the entire df which is wasteful.

Upvotes: 2

Related Questions