How do I delete a specific year from df

Question

df.head()

          HEAT/HOT WATER
created_date    
2015-12-31  169457.0
2016-12-31  155578.0
2017-12-31  152009.0
2018-12-31  153277.0
2019-12-31  76712.0

I would like to delete 2019 from index

I have tried this with no success.

df = df[df['created_date'] != '2019']

@WeNYoBen I don't follow? Why would making a copy help in an assignment? — roganjosh, May 16 '19 at 16:20
@roganjosh check https://stackoverflow.com/questions/20625582/how-to-deal-with-settingwithcopywarning-in-pandas, prevent SettingWithCopyWarning — BENY, May 16 '19 at 16:25
Well aware of that answer. I'm just not sure how you came to the `.copy()` conclusion from the question — roganjosh, May 16 '19 at 16:26
@WeNYoBen I agree about the Warning and using .copy() but doesnt that only apply when youre setting the df to a new one? in this case the df is remaining the same. — Christian Torres, May 16 '19 at 16:30
@ChristianTorres that is just my coding behavior, everyone have their own coding style , like I always do df.loc[1,:] rather than df.loc[1] — BENY, May 16 '19 at 16:32

yatu · Accepted Answer · 2019-05-16T16:33:24.917

3

Convert the index to a series so it can support vectorized datetime operations and check which years are not 2019 using Series.ne

# df.index = pd.to_datetime(df.index)
df[df.index.to_series().dt.year.ne(2019)]

               HEAT/HOTWATER
2015-12-31       169457.0
2016-12-31       155578.0
2017-12-31       152009.0
2018-12-31       153277.0

Or simply df[df.index.year != 2019], as the index is already a DatetimeIndex as pointed out by @piRSquared in the comments

edited May 16 '19 at 16:33

answered May 16 '19 at 16:20

yatu

86,083
12
84
139

thanks! this worked for me. what would be the code if it were a column instead of index? – Christian Torres May 16 '19 at 16:25
1

This is sufficient `df[df.index.year != 2019]` – piRSquared May 16 '19 at 16:25
1

@ChristianTorres `df[df['created_date'].dt.year.ne(2019)]` – piRSquared May 16 '19 at 16:26
@piRSquared how would it look like if it were for two years, ie. 2019 and 2010, i tried `df[df['created_date'].dt.year.ne(2019,2010)]` but it didnt work – Christian Torres May 16 '19 at 16:26
2

@ChristianTorres `df[df['created_date'].year.isin([2019,2010])]` – BENY May 16 '19 at 16:29
1

thanks @piRSquared, of course its a `DatetimeIndex`! – yatu May 16 '19 at 16:32

How do I delete a specific year from df

1 Answers1