integer to TimeStamp in Python

Question

Consider a pd.Series which contains a date column with integers, e.g. 01052012. What is the best way to convert this?

Example:

        tDate
0    20040915
1    20041020
2    20041117
3    20041222
4    20050119
..        ...
203  20210818
204  20210915
205  20211020
206  20211117
207  20211222

Using pd.TimeStamp(x) outputs:

>> pd.Timestamp(x.data.tDate[0])
Timestamp('1970-01-01 00:00:00.020040915')

I can do loop over each element and do the following, but it's most likely not a good practice:

y = x.values[0][0])[:4]
m = x.values[0][0])[4:6]
d = x.values[0][0])[6:]
pd.Timestamp(int(y),int(m),int(d))

pd.datetime(x.data.tDate, format='%Y%m%d', errors='coerce'), considering tDate is a pandas series column — Anurag Reddy
– Anurag Reddy, Commented Oct 14, 2020 at 19:01

Fading Origami · Accepted Answer · 2020-10-14 19:07:43Z

3

this should work as well

 df['tDate'] = df['tDate'].apply(lambda x: pd.Timestamp(str(x)))

Another option will be to just cast the whole column like this

 df['tDate'] = df['tDate'].astype(str).astype('datetime64')

answered Oct 14, 2020 at 19:07

Fading Origami

2031 silver badge8 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

FObersteiner Over a year ago

no need for the apply/lambda... pd.to_datetime(s.astype(str)) allows for adjustments if needed and is slightly more readable.

deblue Over a year ago

This is slightly faster than the answer by wpercy

George Ogden Over a year ago

You might need .astype("datetime64[ms]") or .astype("datetime64[ns]") in newer versions

wpercy · Accepted Answer · 2020-10-19 22:31:06Z

0

I would first cast the integer as a string and use the strptime method of the datetime module.

Something like this:

In [3]: s = pd.Series([20040915,20041020])

In [4]: s = s.map(str).map(lambda x: datetime.datetime.strptime(x, '%Y%m%d'))

In [5]: s
Out[5]:
0   2004-09-15
1   2004-10-20
dtype: datetime64[ns]

As mentioned in the comments,

In [6]: s = pd.Series([20040915,20041020])

In [7]: pd.to_datetime(s.astype(str))
Out[7]:
0   2004-09-15
1   2004-10-20
dtype: datetime64[ns]

will give the same result and is likely faster with large series.

edited Oct 19, 2020 at 22:31

answered Oct 14, 2020 at 19:01

wpercy

10.2k4 gold badges35 silver badges50 bronze badges

2 Comments

FObersteiner Over a year ago

no need fo map, lambda or datetime here, simply pd.to_datetime(s.astype(str))

deblue Over a year ago

I took this as an accepted comment since it gives you more control

Collectives™ on Stack Overflow

integer to TimeStamp in Python

2 Answers 2

3 Comments

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

3 Comments

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related