Got a dataframe df with a column "Id"
Id 0 -KkJz3CoJNM 1 08QMXEQbEWw 2 0ANuuVrIWJw 3 0pPU8CtwXTo 4 1-wYH2LEcmk I need to convert column "Id" into a set() but
set_id = set(df["Id"]) print(set_id) returns
{'Id'} instead of a set() of the strings from column "Id"?
For me working correctly if exist only one id column:
set_id = set(df["Id"]) print(set_id) {'1-wYH2LEcmk', '08QMXEQbEWw', '0pPU8CtwXTo', '0ANuuVrIWJw', '-KkJz3CoJNM'} But if there is more columns names id then df['id'] return DataFrame, so set(df["Id"]) return unique columns names:
#test for 2 columns with sample data df = pd.concat([df, df], axis=1) print (df["Id"]) Id Id 0 -KkJz3CoJNM -KkJz3CoJNM 1 08QMXEQbEWw 08QMXEQbEWw 2 0ANuuVrIWJw 0ANuuVrIWJw 3 0pPU8CtwXTo 0pPU8CtwXTo 4 1-wYH2LEcmk 1-wYH2LEcmk set_id = set(df["Id"]) print(set_id) {'Id'} Because:
L = list(df["Id"]) print(L) ['Id', 'Id'] working same like
L = list(df["Id"].columns) print(L) ['Id', 'Id'] and similar for sets:
set_id = set(df["Id"].columns) print(set_id) {'Id'} Possible solution for deduplicate columns:
c = df.columns.to_series() df.columns += c.groupby(c).cumcount().astype(str).radd('.').replace('.0','') print (df) Id Id.1 0 -KkJz3CoJNM -KkJz3CoJNM 1 08QMXEQbEWw 08QMXEQbEWw 2 0ANuuVrIWJw 0ANuuVrIWJw 3 0pPU8CtwXTo 0pPU8CtwXTo 4 1-wYH2LEcmk 1-wYH2LEcmk Or if always same values remove duplicated columns:
df = df.loc[:, ~df.columns.duplicated()] print (df) Id 0 -KkJz3CoJNM 1 08QMXEQbEWw 2 0ANuuVrIWJw 3 0pPU8CtwXTo 4 1-wYH2LEcmk 
df = df.loc[:, ~df.columns.duplicated()].df = df.T.drop_duplicates().T - Because be default pandas remove duplicates by rows, not by columns.df = df.drop_duplicates(axis=1)