Column¶

set columns in chained method¶

df = df.set_axis(columns, axis=1)

split col and convert to int¶

split a column in a DataFrame based on a delimiter (in this case, "-") and convert the first item of the split result to integers

df['new_col'] = df['col_with_dash'].str.split('-').str[0].astype(int)

split one col to two¶

df[['cola','colb']] = df['col'].str.rsplit('_', 1, expand=True)

newcol from two col levels (year and quarter)¶

qt_yr = [f'Q{quarter}{str(year)[2:]}' for year, quarter in df.columns]

merge df col levels¶

data = {
    ('A', 'X'): [1, 2, 3],
    ('A', 'Y'): [4, 5, 6],
    ('B', 'X'): [7, 8, 9],
    ('B', 'Y'): [10, 11, 12]
}
df = pd.DataFrame(data)
df.columns = df.columns.map('_'.join)
df.columns = [f'{i}_{j}' if j != '' else f'{i}' for i,j in df.columns]

# use pipe to chain the method
df = (
    df
    .pipe(
        lambda x: x.set_axis(x.columns.map('_'.join), axis=1)
    )
)

newcol based on conditions on another col¶

#use np.select
cons = [
    (df['cnt'] <= 2),
    (df['cnt'] > 2) & (df['cnt'] <= 9),
    (df['cnt'] > 9) & (df['cnt'] <= 15),
    (df['cnt'] > 15),
]
# create a list of the values we want to assign for each condition
lbls = ['band1', 'band2', 'band3', 'band4'] #label for each condition
# create a new column and use np.select to assign values to it
df['band'] = np.select(cons, lbls)

#use pd.cut
#bins = [-np.inf,2,9,15,np.inf] #or
bins = [float('-inf'),2,9,15,float('inf')]
lbls = ['band1', 'band2', 'band3', 'band4]
df['band'] = pd.cut(x=df['cnt'], bins=bins, right=True, labels=lbls).astype(str) #change category to str