问题
I have this DatetimeIndex:
dates = DatetimeIndex(['2017-06-09', '2017-06-10', '2017-06-11', '2017-06-12',
'2017-06-13', '2017-06-14'],
dtype='datetime64[ns]', freq='<DateOffset>')
I want to take dates and append them to my DataFrame df
:
for i in xrange(0,5):
df.append(dates[i],ignore_index=True)
I get this error TypeError: cannot concatenate a non-NDFrame object
.
UPDATE:
sample data of df
:
Out[85]:
2017-06-05 -0.944868
2017-06-06 0.073623
2017-06-07 -0.687232
Freq: <DateOffset>, dtype: float64
回答1:
If length of df
is same as DatetimeIndex
and need create index
:
df.index = dates
If not try filter by length of index
what is same as length of df
:
df.index = dates[:len(df.index)]
If need new column:
df['a'] = dates
If not:
df['a'] = dates[:len(df.index)]
If need use only first 5 values:
df['a'] = dates[:5]
EDIT:
I think you need union for concatenate index to dates
and then reindex:
df = df.reindex(df.index.union(dates), fill_value=-0.944868)
print (df)
2017-06-05 -0.944868
2017-06-06 0.073623
2017-06-07 -0.687232
2017-06-09 -0.944868
2017-06-10 -0.944868
2017-06-11 -0.944868
2017-06-12 -0.944868
2017-06-13 -0.944868
2017-06-14 -0.944868
Name: <DateOffset>, dtype: float64
来源:https://stackoverflow.com/questions/45167175/typeerror-cannot-concatenate-a-non-ndframe-object