I have a pandas dataframe with a row that contains data such as:
1 year
1 month
1 week
4 year
3 week
etc etc
I am trying to replace anything that contains "month" or "week" to 0
train_df.age["weeks" in train_df.age] = 0
and
for i in train_df['age']:
if "weeks" in i:
i = "0"
None of which seem to work.
Any advice on how to do this? Thanks.
Use str.contains:
train_df.loc[train_df['age'].str.contains(r'week|month'), 'age'] = 0
Here we pass a regex pattern that looks for whether the row contains either 'week' or 'month' and use the boolean mask to selectively update just the rows on interest:
In [4]:
df.loc[df['age'].str.contains(r'week|month'), 'age'] = 0
df
Out[4]:
age
1 year
1 0
1 0
4 year
3 0
stacked = train_df.stack() stacked.loc[stacked.str.contains(r'week|month'), 'age'] = 0 and then call unstack() unstack = stacked.unstack() or you can just call apply on the df to generate a boolean mask and then replace mask = df.apply(lambda x: x.str.contains(r'week|week')) and then do df[mask] = 0mask idea is actually working if i use lambda x: str('week') in str(x) as i have a df with mixed dtypes. otherwise with regex mask method as above i get error : nothing to repeat ... maybe i will post a dedicated question about this with full examples