Convert Integer into Datetime in Python

Question

I’m working on a df that gives me Date in a integer form, and i want to convert it into Datetime so i can continue to manipulate it.

Basically i have this DF where i have a Column like this below

    DATE
18010101
18010101
18010101
18010101
18010101
... ... ... ... ... ...
20123124
20123124
20123124
20123124
20123124

the respective order in this dateset is (Year, month, day, hour)

i’ve already tried to do something like this

df["Year"] = df.DATE[0:2]
df["Month"] = df.DATE[2:4]~

but it converts it into a float and for the first line example, it becomes similar to this

DATE    Year
18010101    18010101.0

where it was supposed to be

Year = 2018 or 18

I would appreciate a lot the help of you guys.

Asked By: Carlos Juvenal

||

Source

Answer 1

Use Series.str.extract and a regex

df = pd.DataFrame([[18010101], [18010101], [18010101], [18010101], [18010101],
                   [20123124], [20123124], [20123124], [20123124], [20123124]],
                  columns=['DATE'])
df[['year', 'month', 'day', 'hour']] = df['DATE'].astype(str).str.extract(r'(d{2})(d{2})(d{2})(d{2})')

DATE	year	month	day	hour
18010101	18	01	01	01
18010101	18	01	01	01
18010101	18	01	01	01
18010101	18	01	01	01
18010101	18	01	01	01
20123124	20	12	31	24
20123124	20	12	31	24
20123124	20	12	31	24
20123124	20	12	31	24
20123124	20	12	31	24

Answered By: azro

Answer 2

Earlier, you tried to split an int like a string.
But you can first convert it into a string (by .astype(str)) and then split it. Check the lines below.

date_str = df.DATE.astype(str)
df["Year"] = date_str.str[0:2]
df["Month"] = date_str.str[2:4]

Answered By: Jitesh

Answer 3

Are your hour stamps actually from 1 to 24? Normal convention for 24-hour time is 0-23 where you can use:

df = pd.DataFrame([[18010100], [18010100], [18010100], [18010100], [18010100],
                   [20123123], [20123123], [20123123], [20123123], [20123123]],
                  columns=['DATE'])

pd.to_datetime(df['DATE'], format='%y%m%d%H')

Or simply take 1 from each number beforehand like so:

df = pd.DataFrame([[18010101], [18010101], [18010101], [18010101], [18010101],
                   [20123124], [20123124], [20123124], [20123124], [20123124]],
                  columns=['DATE'])
pd.to_datetime(df['DATE']-1, format='%y%m%d%H')```

Answered By: Matt Rosinski

Convert Integer into Datetime in Python

Question:

Answers: