I am having a dataframe like this

| Text     | Label          |
| -------- | -------------- |
| ........ | ........       | 
|  65      | B-phone        |
| 5454     | I-phone        |
| 3485     | I-phone        |
| ........ | ........       | 

How could I merge the the above 3 rows into a single row like this? In specific, the B-phone label will be fixed (the 65 text can be any value and at any index), and any rows below it with the label "I-phone" will be merged into the row with label B-phone

CodePudding user response：

Adding an id attribute to assist in grouping the records:

df = pd.DataFrame([[' 65','B-phone'],['5454','I-phone'],['3485','I-phone']], columns=['Text','Label'])
df["id"] = 1
df_text = df.groupby('id')['Text'].aggregate(lambda x: ' '.join(tuple(x))).reset_index()
df_label = df.groupby('id')['Label'].aggregate(lambda x: x[0]).reset_index()
out = pd.merge(df_text,df_label)

Gives:

    id  Text            Label
0   1    65 5454 3485   B-phone

CodePudding user response：

Strictly addressing your stated requirements, this should work:

pd.DataFrame({'ID':[1], 'Text': [' '.join(df['Text'].to_list())], 'Label': [df.iat[0,1]]})

Result:

   ID           Text    Label
0   1   65 5454 3485  B-phone