如何保存 .csv 文件，其中我使用 pandas 将字符串数据转换为 .csv 文件特定列的数字数据？

Question

I wrote this program where i converted the string data into numerical data of the given rows. The actual csv file is in here.

> df.Sex[df.Sex == 'M'] = 1 df.Sex[df.Sex == 'F'] = 0
> #changing ChestPainType of TA ,ATA,NAP and ASY into 1,2,3 and 4 df.ChestPainType[df.ChestPainType == 'TA'] = 1
> df.ChestPainType[df.ChestPainType == 'ATA'] = 2
> df.ChestPainType[df.ChestPainType == 'NAP'] = 3
> df.ChestPainType[df.ChestPainType == 'ASY'] = 4
> # changing ExerciseAngina of N = 0 and Y = 1 df.ExerciseAngina[df.ExerciseAngina == 'N'] = 0
> df.ExerciseAngina[df.ExerciseAngina == 'Y'] = 1
> # changing RestingECG of Normal,ST and LVH into 1,2 and 3 df.RestingECG[df.RestingECG == 'Normal'] = 1
> df.RestingECG[df.RestingECG == 'ST'] = 2 df.RestingECG[df.RestingECG
> == 'LVH'] = 3
> 
> df.ST_Slope[df.ST_Slope == 'Up'] = 1 df.ST_Slope[df.ST_Slope ==
> 'Flat'] = 2 df.ST_Slope[df.ST_Slope == 'Down'] = 3 df.head()

and it is showing the output of the first 5 rows of the file.

but afterwards when i try to print the correlations using this program:

pearsoncorr = df.corr(method = 'pearson')   #df = the .csv file i am working with. 

pearsoncorr

the output it is showing me is this.

Here, I want to see the correlations of the new csv file i made changes earlier which should be this file and the expected correlation output should be almost like this showing all the columns . But this correlative table is showing me the correlations of this csv file.

The question is, How can i save the modified .csv file?

PS I am new in this site so if there are any errors I made, i apologize and i will be glad if you let me know how can i change it.

Answer 1

You're most probably trying to set the values on a copy of a slice from the DataFrame

SettingWithCopyWarning: A value is trying to be set on a copy of a slice from a DataFrame

You can see on view-versus-a-copy and SettingWithCopyWarning why this may be a problem.

When using df.Sex[df.Sex == 'M'] = 1 the original DataFrame may not have been altered. You can check this using df.info() and inspect the Dtype column.

 #   Column          Non-Null Count  Dtype
---  ------          --------------  -----
 0   Age             918 non-null    int64
 1   Sex             918 non-null    object
 2   ChestPainType   918 non-null    object
 3   RestingBP       918 non-null    int64
 ...

In the previously mentioned link from Pandas documentation the recommended access method is using .loc for multiple items (using mask, df.Sex == 'M' in your case) and a single item using a fixed index:

df.loc[df.Sex == 'M', 'Sex'] = 1
df.loc[df.Sex == 'F', 'Sex'] = 0

Another option would be to use Pandas map , that, in my opinion better expresses the code intent.

df.Sex = df.Sex.map({'M':1, 'F':0})
df.ChestPainType = df.ChestPainType.map({'TA': 1, 'ATA': 2, 'NAP': 3, 'ASY': 4})
df.ExerciseAngina = df.ExerciseAngina.map({'N': 0, 'Y': 1})
...

Cheking again for the Dtype column we can ensure that the values are in fact of type integer, allowing you to use df.corr (or save the Dataframe with the new values to a csv with to_csv ).

Data columns (total 12 columns):
#   Column          Non-Null Count  Dtype
---  ------          --------------  -----
0   Age             918 non-null    int64
1   Sex             918 non-null    int64
2   ChestPainType   918 non-null    int64
3   RestingBP       918 non-null    int64
...