CODEWITHSUNDEEP
pandasdataframe
Mk47
Mk47

Reputation: 155

Get pandas row names if value in pandas row is present

I am trying to convert a one hot key dataframe into a 2 d frame

Is there anyways I can iterate over rows and columns and fill the values having a 1 with the column name.

problem dataframe:

+------------------+-----+-----+
| sentence         | lor | sor |
+------------------+-----+-----+
| sam lived here   | 0   | 1   |
+------------------+-----+-----+
| drack lived here | 1   | 0   |
+------------------+-----+-----+

Solution dataframe:

+------------------+------+
| sentence         | tags |
+------------------+------+
| sam lived here   | sor  |
+------------------+------+
| drack lived here | lor  |
+------------------+------+

Upvotes: 1

Views: 102

Answers (1)

Abercrombie
Abercrombie

Reputation: 1086

You can segregate the rows having 1 for every column. For these columns, replace the value 1 with the name specified along with renaming the column names

lor_df = df.loc[df["lor"].eq(1), "lor"].rename(columns={"lor": "tags"}).replace(1, "lor")
sor_df = df.loc[df["sor"].eq(1), "sor"].rename(columns={"sor": "tags"}).replace(1, "sor")

After this, concatenate the individual results using pandas.concat, followed by dropping the columns which aren't required.

df["tags"] = pd.concat([lor_df, sor_df], sort=False)
df.drop(columns=["lor", "sor"], inplace=True)

To ensure unique values we can use pandas.DataFrame.drop_duplicates

df.drop_duplicates(inplace=True)
print(df)

Upvotes: 1

Related Questions