将日期和小时转换为日期时间 python pandas

Question

Using Pyhon 3.8, pandas 1.1.2

MRE:

df = pd.DataFrame({"date":["2020-07-20", "2020-07-20", "2020-07-20", "2020-07-20"],
                   "hour":["1", "2", "3", "4"]})

df["datetime"]  = df["date"] + " " + df["hour"]

This creates dataframe that looks like this:

    date       hour     datetime
0   2020-07-20  1   2020-07-20 1
1   2020-07-20  2   2020-07-20 2
2   2020-07-20  3   2020-07-20 3
3   2020-07-20  4   2020-07-20 4

I want to make datetime column an actual datetime object.

Desired output would be

    date       hour     datetime
0   2020-07-20  1   2020-07-20 01:00:00
1   2020-07-20  2   2020-07-20 02:00:00
2   2020-07-20  3   2020-07-20 03:00:00
3   2020-07-20  4   2020-07-20 04:00:00

I want to use this data for plotting a time series graph. datetime format doesn't need to be in specific format as long as it knows increasing order when plotting using plotly.

Thanks in advance!

Answer 1

Use pandas.to_datetime after zfill :

df["datetime"]  = df["date"] + " " + df["hour"].str.zfill(2)
pd.to_datetime(df["datetime"])

Output:

0   2020-07-20 01:00:00
1   2020-07-20 02:00:00
2   2020-07-20 03:00:00
3   2020-07-20 04:00:00
Name: datetime, dtype: datetime64[ns]

Answer 2

Another way

#pad df.hour with 0 on the left

s=df.hour.str.pad(width=2, side='left', fillchar='0')

#concat using df.str.cat

df['datetime']=pd.to_datetime(df['date'].str.cat(s+":00:00", sep=' '))

    date      hour        datetime
0  2020-07-20    1  2020-07-20 01:00:00
1  2020-07-20    2  2020-07-20 02:00:00
2  2020-07-20    3  2020-07-20 03:00:00
3  2020-07-20    4  2020-07-20 04:00:00

Answer 3

Another way using datetime module in python.

strftime converts datetime to str,
strptime converts str to datetime.

base_df["datetime"] = base_df["datetime"].apply(lambda row: datetime.strptime(row, "%Y-%m-%d %H"))