Setting index values to pd.DataFrame `index` argument removes data from columns

Question

When I create new DataFrame with custom index values it replaces data from columns with NaN/NaT values.

I've tried assigning pd.Index as an index argument as well, and result was the same.

pd.DataFrame(
    data={
        "date": pd.Series([datetime(2000, 1, 2)]),
        "duration": pd.Series([datetime(1970, 1, 1, 0, 5)]),
    }
    index = [1] 
    )

Instead of:

    date    duration
1   2000-01-02  1970-01-01 00:05:00

I receive:

    date    duration
1   NaT NaT

Is it a bug?

I use pandas 0.25.0

The indexes in your dataframe constructor and in your series have to match. pd.Series yields default indexes starting at 0, and you passed [1] as argument in your pd.DataFrame. — rafaelc
– rafaelc, Commented Sep 16, 2019 at 13:28

jezrael · Accepted Answer · 2019-09-16 13:34:06Z

Remove Series from DataFrame constructor, because their default index is 0, so different like index of DataFrame, so returned missing values (indices have to match):

df = pd.DataFrame(
    data={
        "date": [datetime(2000, 1, 2)],
        "duration": [datetime(1970, 1, 1, 0, 5)],
    },
    index = [1] 
    )

print (df)
        date            duration
1 2000-01-02 1970-01-01 00:05:00

Detail:

print (pd.Series([datetime(2000, 1, 2)]))
0   2000-01-02
dtype: datetime64[ns]

So if need Series, there is necessary set index to 1 too:

df = pd.DataFrame(
    data={
        "date": pd.Series([datetime(2000, 1, 2)], index = [1]),
        "duration": pd.Series([datetime(1970, 1, 1, 0, 5)],index = [1]),
    },
    index = [1] 
    )

Or remove in DataFrame for default 0 index:

df = pd.DataFrame(
    data={
        "date": pd.Series([datetime(2000, 1, 2)]),
        "duration": pd.Series([datetime(1970, 1, 1, 0, 5)]),
    },
    )

print (df)
        date            duration
0 2000-01-02 1970-01-01 00:05:00

Collectives™ on Stack Overflow

Setting index values to pd.DataFrame `index` argument removes data from columns

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related