You can filter by inverting (~) the logical mask for columns that do not need to be removed with loc
and str.endswith
also works str.contains
with $
for the end of the line end:
cols = ['SectorName', 'Name Sector', 'ItemName', 'Item', 'Counterpart SectorName']
df = pd.DataFrame([range(5)], columns = cols)
print (df)
SectorName Name Sector ItemName Item Counterpart SectorName
0 0 1 2 3 4
print (~df.columns.str.endswith('Name'))
[False True False True False]
df1 = df.loc[:, ~df.columns.str.endswith('Name')]
df1 = df.loc[:, ~df.columns.str.contains('Name$')]
Or first write down the column names:
print (df.columns[~df.columns.str.endswith('Name')])
Index(['Sector', 'Item'], dtype='object')
df1 = df[df.columns[~df.columns.str.endswith('Name')]]
print (df1)
Name Sector Item
0 1 3
source
share