Create a ~astropy.table.Table from a pandas.DataFrame instance

In addition to converting generic numeric or string columns, this supports conversion of pandas Date and Time delta columns to ~astropy.time.Time and ~astropy.time.TimeDelta columns, respectively.

Parameters¶

dataframepandas.DataFrame

A pandas pandas.DataFrame instance

indexbool

Include the index column in the returned table (default=False)

units: dict

A dict mapping column names to to a ~astropy.units.Unit. The columns will have the specified unit in the Table.

Returns¶

table~astropy.table.Table

A ~astropy.table.Table (or subclass) instance

Raises¶

ImportError

If pandas is not installed

Examples¶

Here we convert a pandas.DataFrame instance to a ~astropy.table.QTable.

>>> import numpy as np
>>> import pandas as pd
>>> from astropy.table import QTable

>>> time = pd.Series(['1998-01-01', '2002-01-01'], dtype='datetime64[ns]')
>>> dt = pd.Series(np.array([1, 300], dtype='timedelta64[s]'))
>>> df = pd.DataFrame({'time': time})
>>> df['dt'] = dt
>>> df['x'] = [3., 4.]
>>> with pd.option_context('display.max_columns', 20):
...     print(df)
        time              dt    x
0 1998-01-01 0 days 00:00:01  3.0
1 2002-01-01 0 days 00:05:00  4.0

>>> QTable.from_pandas(df)
<QTable length=2>
          time              dt       x
          Time          TimeDelta float64
----------------------- --------- -------
1998-01-01T00:00:00.000       1.0     3.0
2002-01-01T00:00:00.000     300.0     4.0