在 Python 中更改 Pandas DataFrame 中的值

Question

I have a data set that has been loaded into a pandas DataFrame . When I print data.head() , it looks like this.

G1  G2  G3  absences  failures  studytime romantic internet
0   5   6   6         6         0          2       no       no
1   5   5   6         4         0          2       no      yes
2   7   8  10        10         3          2       no      yes
3  15  14  15         2         0          3      yes      yes
4   6  10  10         4         0          2       no       no

I am attempting to create a linear regression model and want to convert the yes' and no's to 1s and 0s in the romantic and internet columns.

The code I used:

df['romantic'].replace('yes', 0)
df['romantic'].replace('no', 1)
df['internet'].replace('yes', 0)
df['internet'].replace('no', 1)

Did not work :( It also did not display an error of any sort.

I tried to make a linear model with data = df[["G1", "G2", "G3", "absences", "failures", "studytime", "romantic", "internet"]] and it showed:

ValueError: could not convert string to float: 'yes'

Even though I thought I converted them. Please help, thanks...

Answer 1

To convert both your columns of interest, run:

df.romantic = (df.romantic == 'yes').astype(int)
df.internet = (df.internet == 'yes').astype(int)

Note also that you wrote convert the yes' and no's to 1s and 0s , so in your code sample you attempt to assing the values just the opposite way.

Answer 2

如果要将所有“是”替换为 0，将所有“否”替换为 1，请使用：

df.replace({'yes': 0, 'no': 1})

Answer 3

df.replace({'yes': 0, 'no': 1}, regex=True)

Try this. Replaces all occurences of 'yes' with 0 and all occurences of 'no' with 1.

Answer 4

mapper = {'yes':0,'no':1}
df.loc[:,'romantic'] = df['romantic'].map(mapper)
df.loc[:,'internet'] = df['internet'].map(mapper)

Use map function for this job

Answer 5

You need to assign it while replacing:

df = pd.DataFrame({'romantic':['no','no','no','yes','no'], 'internet':['no','yes','yes','yes','no']})
df

df['romantic'] = df['romantic'].replace('yes', 0)
df['romantic'] = df['romantic'].replace('no', 1)
df['internet'] = df['internet'].replace('yes', 0)
df['internet'] = df['internet'].replace('no', 1)

print(df)

   romantic  internet
0         1         1
1         1         0
2         1         0
3         0         0
4         1         1

There are more ways to do this in Python:

Pandas.Series.apply Lambda

df['romantic'] = df['romantic'].apply(lambda x: 0 if x == 'yes' else 1)
df['internet'] = df['internet'].apply(lambda x: 0 if x == 'yes' else 1)

Numpy.where

df['romantic'] = np.where(df['romantic'] == 'yes',0,1)
df['internet'] = np.where(df['internet'] == 'yes',0,1)

Pandas.Series.map

df['romantic'] = df['romantic'].map(dict(yes = 0, no = 1))
df['internet'] = df['internet'].map(dict(yes = 0, no = 1))

All yield the same result.