Pandas: extracting multiple rows from DataFrame with list of DatetimeIndices

问题

I have a pandas Dataframe with the following index with seconds frequency:

DatetimeIndex(['2015-12-28 05:20:05', '2015-12-28 05:20:06',
               '2015-12-28 05:20:07', '2015-12-28 05:20:08',
               ...
               '2015-12-28 21:19:55', '2015-12-28 21:19:56',
               '2015-12-28 21:19:57', '2015-12-28 21:19:58']

I want to extract multiple rows at once, given a list of datetime strings. I tried:

df.loc[['2015-12-28 08:32:39', '2015-12-28 08:32:48']]

But I get the following error:

KeyError: "None of [['2015-12-28 08:32:39', '2015-12-28 08:32:48']] are in the [index]"

回答1:

You can convert list to datetime by DatetimeIndex or to_datetime:

d = ['2015-12-28 05:20:05', '2015-12-28 21:19:58']
print (df.loc[pd.DatetimeIndex(d)])

Or:

print (df.loc[pd.to_datetime(d)])

Sample:

idx = pd.DatetimeIndex(['2015-12-28 05:20:05', '2015-12-28 05:20:06',
               '2015-12-28 05:20:07', '2015-12-28 05:20:08',
               '2015-12-28 21:19:55', '2015-12-28 21:19:56',
               '2015-12-28 21:19:57', '2015-12-28 21:19:58'])
df = pd.DataFrame({'s': range(8)}, index=idx)  
print (df)
                     s
2015-12-28 05:20:05  0
2015-12-28 05:20:06  1
2015-12-28 05:20:07  2
2015-12-28 05:20:08  3
2015-12-28 21:19:55  4
2015-12-28 21:19:56  5
2015-12-28 21:19:57  6
2015-12-28 21:19:58  7

d = ['2015-12-28 05:20:05', '2015-12-28 21:19:58']
print (df.loc[pd.to_datetime(d)])
                     s
2015-12-28 05:20:05  0
2015-12-28 21:19:58  7

print (df.loc[pd.to_datetime(d)])
                     s
2015-12-28 05:20:05  0
2015-12-28 21:19:58  7

来源：https://stackoverflow.com/questions/45213834/pandas-extracting-multiple-rows-from-dataframe-with-list-of-datetimeindices

标签

python

pandas

dataframe

indexing