如何从数据框中切片唯一的日期数据？

Question

Existing slicing methods on dates have commonly extracted the data between two dates but not distinctive dates. My problem is slicing the data of distinctive dates. My dataframe is given by:

df =                           A         B
     2019-03-21 19:15:00   21.787958  16.728439  
     2019-03-25 19:16:00   20.983078  15.865983 
     2019-03-29 19:17:00   20.122042  15.073062  

I want to extract the data on days 21 and 29. My code is given below: Code1:

df.index == ['2019-03-21','2019-03-29']

Output:

ValueError: Lengths must match

Code2:

df['2019-03-21','2019-03-29']

Output:

KeyError: ('2019-03-21', '2019-03-25')

Could you help me to find the mistake here?

Answer 1

Few things going on here. First, when you compare one list to another with an "==", it doesn't necessarily return an elementwise comparison - you have to use pandas inbuilt 'isin' method.

Second, when you pass a mask to a dataframe to filter it, the mask needs to have the same number of elements as the rows in the dataframe.

Third, you have an index that is a datetime, that you want to compare with a date - so you have to extract the date component first to compare.

df=pd.DataFrame({'A':[21.787958,20.983078,20.122042], 'B':16.728439,15.865983,15.073062]})
df.index=pd.to_datetime(['2019-3-21 19:15:0','2019-3-25 19:16:0','2019-3-29 19:17:0'])

So here is the filtered dataframe:

df[pd.to_datetime(df.index.date).isin(pd.to_datetime(['2019-03-21','2019-03-29']))]

Answer 2

You can do it using

df.loc[df['date_column'].isin(['2019-03-21','2019-03-29'])]  

or

 df[(df['A'] == '2019-03-21') | (df['A'] ==  '2019-03-29')] 

the df.index that you wrote just gives the indexes of the rows. for instance if you run:

for i in df.index: 
    print(i)

it returns 0,1,2 which are the rows indexes. but what you are looking for is the content of a specific column which is df['date_column'] So you should compare the values in this column with what you are looking for.