通过jupyter将a.csv文件转换为Python指定样式

Question

The file looks like:

The first column is user_id , and the second is the rating for joke 1 , and the rest can be done in the same manner. I want to convert the file shown above into the format likes:

user_id | joke_id | rating
--------------------------
   1    |   1     | -7.82
   1    |   2     | 8.79

In addition, after conversion, as the normal ratings are between -10 and +10, the number 99 means the user didn't rate for the corresponding jokes and should be removed.

Answer 1

Your question involves several steps, please avoid mixing all questions into one. Based on your question, following steps would be helpful:

• Reead csv file using pandas

import pandas as pd
raw = pd.read_csv('PATH-TO-FILE')

• Use melt to unpivot the DataFrame

Because only the image is provided, will use a sample DataFrame instead.

raw = pd.DataFrame([[1, -7, 8, 99], [2, 4, 0, 6]], columns = ['user_id', 'joke_1', 'joke_2', 'joke_3'])

   user_id  joke_1  joke_2  joke_3
0        1      -7       8      99
1        2       4       0       6

Unpivot DataFrame using melt :

df = pd.melt(raw, id_vars=['user_id'], value_vars=['joke_1', 'joke_2', 'joke_3'], var_name='joke', value_name='rating')

   user_id    joke  rating
0        1  joke_1      -7
1        2  joke_1       4
2        1  joke_2       8
3        2  joke_2       0
4        1  joke_3      99
5        2  joke_3       6

• Filter the Dataframe with condition

With pandas , you can easily filter DataFrame with condition:

df_processed = df[df.rating != 99].reset_index(drop=True)

Please notice the reset_index() is used to clean the index, not related to your question.

Hope above would help.