从 Pandas Dataframe 根据日期删除行

Question

I'm trying to create a dataframe with pandas and drop dates later than say 201702. The dataframe is structured as so

    Date         Account Number
1   2019-02-21   123841234
2   2017-01-01   193741927
3   2015-03-04   981237432
4   2018-05-29   134913473
5   2012-05-12   138749173
6   2009-01-04   174917239

I'm reading in the csv (data.csv) and attempting to remove any date after 2017-02-28 as so:

data_csv = pd.read_csv('data.csv')
data_csv[data_csv.DATE < '2017-02-28']

Is this supposed to work correctly with a date format of YYYY-MM-DD or is there something I'd have to do to the column format to ensure that these rows are dropped from the dataframe?

Thank you for your time.

Answer 1

I suggest you transform the string into a Timestamp , assuming data_csv.Date is also a Timestamp :

result = data_csv[data_csv.Date < pd.to_datetime('2017-02-28')]
print(result)

Output

        Date  Account Number
1 2017-01-01       193741927
2 2015-03-04       981237432
4 2012-05-12       138749173
5 2009-01-04       174917239

Answer 2

If your date strings are in YYYY-MM-DD format then lexicographical comparisons work out of the box (for python in general, not just pandas).

'2009-01-04' < '2017-02-28'  
# True

'2019-01-04' < '2017-02-28'
# False

So your comparison should work out without any changes. Although it'd be safer to convert to datetime first, so your code makes no assumptions while still working.

---

df.dtypes     

Date              object
Account Number     int64
dtype: object

df[df['Date'] < '2017-02-28']

         Date  Account Number
2  2017-01-01       193741927
3  2015-03-04       981237432
5  2012-05-12       138749173
6  2009-01-04       174917239

df['Date'] = pd.to_datetime(df['Date'], errors='coerce')
df.dtypes

Date              datetime64[ns]
Account Number             int64
dtype: object

df[df['Date'] < '2017-02-28']

        Date  Account Number
2 2017-01-01       193741927
3 2015-03-04       981237432
5 2012-05-12       138749173
6 2009-01-04       174917239