http://pandas.pydata.org/pandas-docs/stable/generated/pandas.to_datetime.html

Examples

Assembling a datetime from multiple columns of a DataFrame. The keys can be

common abbreviations like ['year', 'month', 'day', 'minute', 'second','ms', 'us', 'ns']) or plurals of the same

>>> df = pd.DataFrame({'year': [2015, 2016],  'month': [2, 3],  'day': [4, 5]})

>>> pd.to_datetime(df)

0  2015-02-04

1  2016-03-05

dtype: datetime64[ns]

If a date does not meet the `timestamp limitations

<http://pandas.pydata.org/pandas-docs/stable/timeseries.html

#timeseries-timestamp-limits>`_, passing errors='ignore'

will return the original input instead of raising any exception.

Passing errors='coerce' will force an out-of-bounds date to NaT,

in addition to forcing non-dates (or non-parseable dates) to NaT.

>>> pd.to_datetime('13000101', format='%Y%m%d', errors='ignore')

datetime.datetime(1300, 1, 1, 0, 0)

>>> pd.to_datetime('13000101', format='%Y%m%d', errors='coerce')

NaT

Passing infer_datetime_format=True can often-times speedup a parsing

if its not an ISO8601 format exactly, but in a regular format.

>>> s = pd.Series(['3/11/2000', '3/12/2000', '3/13/2000']*1000)

>>> s.head()

0    3/11/2000

1    3/12/2000

2    3/13/2000

3    3/11/2000

4    3/12/2000

dtype: object     

>>> %timeit pd.to_datetime(s,infer_datetime_format=True)

100 loops, best of 3: 10.4 ms per loop

>>> %timeit pd.to_datetime(s,infer_datetime_format=False)

1 loop, best of 3: 471 ms per loop

Using a unix epoch time

>>> pd.to_datetime(1490195805, unit='s')

Timestamp('2017-03-22 15:16:45')

>>> pd.to_datetime(1490195805433502912, unit='ns')

Timestamp('2017-03-22 15:16:45.433502912').. warning:: For float arg, precision rounding might happen. To prevent    unexpected behavior use a fixed-width exact type.Using a non-unix epoch origin

>>> pd.to_datetime([1, 2, 3], unit='D',  origin=pd.Timestamp('1960-01-01'))

0    1960-01-02

1    1960-01-03

2    1960-01-04

Parameters

----------

arg : integer, float, string, datetime, list, tuple, 1-d array, Series

    .. versionadded:: 0.18.1

      or DataFrame/dict-like

errors : {'ignore', 'raise', 'coerce'}, default 'raise'

    - If 'raise', then invalid parsing will raise an exception

    - If 'coerce', then invalid parsing will be set as NaT

    - If 'ignore', then invalid parsing will return the input

dayfirst : boolean, default False

    Specify a date parse order if `arg` is str or its list-likes.

    If True, parses dates with the day first, eg 10/11/12 is parsed as

    2012-11-10.

    Warning: dayfirst=True is not strict, but will prefer to parse

    with day first (this is a known bug, based on dateutil behavior).

yearfirst : boolean, default False

    Specify a date parse order if `arg` is str or its list-likes.

    - If True parses dates with the year first, eg 10/11/12 is parsed as

      2010-11-12.

    - If both dayfirst and yearfirst are True, yearfirst is preceded (same

      as dateutil).

    Warning: yearfirst=True is not strict, but will prefer to parse

    with year first (this is a known bug, based on dateutil beahavior).

    .. versionadded:: 0.16.1

utc : boolean, default None

    Return UTC DatetimeIndex if True (converting any tz-aware

    datetime.datetime objects as well).

box : boolean, default True

    - If True returns a DatetimeIndex

    - If False returns ndarray of values.

format : string, default None

    strftime to parse time, eg "%d/%m/%Y", note that "%f" will parse

    all the way up to nanoseconds.

exact : boolean, True by default

    - If True, require an exact format match.

    - If False, allow the format to match anywhere in the target string.

unit : string, default 'ns'

    unit of the arg (D,s,ms,us,ns) denote the unit, which is an

    integer or float number. This will be based off the origin.

    Example, with unit='ms' and origin='unix' (the default), this

    would calculate the number of milliseconds to the unix epoch start.

infer_datetime_format : boolean, default False

    If True and no `format` is given, attempt to infer the format of the

    datetime strings, and if it can be inferred, switch to a faster

    method of parsing them. In some cases this can increase the parsing

    speed by ~5-10x.

origin : scalar, default is 'unix'

    Define the reference date. The numeric values would be parsed as number

    of units (defined by `unit`) since this reference date.

    - If 'unix' (or POSIX) time; origin is set to 1970-01-01.

    - If 'julian', unit must be 'D', and origin is set to beginning of

      Julian Calendar. Julian day number 0 is assigned to the day starting

      at noon on January 1, 4713 BC.

    - If Timestamp convertible, origin is set to Timestamp identified by

      origin.

.. versionadded:: 0.20.0

cache : boolean, default False

    If True, use a cache of unique, converted dates to apply the datetime

    conversion. May produce sigificant speed-up when parsing duplicate date

    strings, especially ones with timezone offsets.

    .. versionadded:: 0.23.0

Returns

-------

ret : datetime if parsing succeeded.

    Return type depends on input:

    - list-like: DatetimeIndex

    - Series: Series of datetime64 dtype

    - scalar: Timestamp

    In case when it is not possible to return designated types (e.g. when

    any element of input is before Timestamp.min or after Timestamp.max)

    return will have datetime.datetime type (or corresponding

    array/Series).