Home > Mobile >  How to change column names of Pandas Series object?
How to change column names of Pandas Series object?

Time:09-01

I'm trying to prefix the names of the columns for each series in my Pandas Series, based on one of the other columns value. Currently my objective is to change a Pandas Dataframe that contains 3 columns into a Dataframe of only 1 column named 'Data' - or whatever. Below is an example of stacking a Dataframe to obtain a single dimension to work with.

df_single_level_cols = pd.DataFrame([[0, 1, 20], [2, 3, 40]],columns=['weight', 'height', 'girth'])
df = df_single_level_cols.stack()
print(df)
0  weight     0
   height     1
   girth     20
1  weight     2
   height     3
   girth     40
dtype: int64

For each series I need to prefix both column names weight and height with the value of girth. When that is done I will drop girth from the equation, leaving me with only the weight and height for my series. After prefixing and dropping the series object should look like the following:

0  20weight     0
   20height     1
1  40weight     2
   40height     3
dtype: int64

Then when converting this to a Dataframe I shall have the following:

             Data
20weight     0
20height     1
40weight     2
40height     3

I've tried messing around with .apply(...), .rename(...) and .add_prefix(...) but none of them seem to be doing the trick. If I do something like

df[0] = df[0].add_prefix("test")

I end up getting errors as I'm setting an array element with a sequence this does not actually use the value of girth - but was more a way of getting accustomed with the rename functionality..

CodePudding user response:

You can melt instead:

df = (df_single_level_cols
 .astype({'girth': str})
 .melt('girth', value_name='Data')
 .assign(**{'girth': lambda d: d['girth'] d.pop('variable')})
 .set_index('girth')
)

output:

          Data
girth         
20weight     0
40weight     2
20height     1
40height     3

CodePudding user response:

Like this?

df = pd.DataFrame([[0, 1, 20], [2, 3, 40]],columns=['weight', 'height', 'girth'])
df = df[['weight', 'height']].stack().reset_index(level=1).merge(df.girth, left_index=True, right_index=True, how='left')
df = df.set_index(df.girth.astype(str)   df.level_1).rename(columns={0: 'Data'})[['Data']]

> df

    Data
20weight    0
20height    1
40weight    2
40height    3
  • Related