Pandas - DataFrame - finding rows with the two latest date-CodePudding

I am not very familiar in python pandas programming.
I have a dataframe like this:

|  No   |   Date     |   hours  |    Flag   |
|:------|:-----------|:---------|:----------|
|  101  | 13.01.2021 |   2:45   |     0     |
|  101  | 13.01.2021 |   1:20   |     0     |
|  101  | 15.02.2021 |   3:15   |     0     |
|  101  | 16.02.2021 |   0:30   |     0     |
|  101  | 02.03.2021 |   1:20   |     0     |
|  101  | 02.03.2021 |   2:40   |     0     |
|  101  | 19.06.2021 |   2:05   |     0     |
|  101  | 19.06.2021 |   1:45   |     0     |
|  101  | 19.06.2021 |   1:25   |     0     |
|  101  | 19.06.2021 |   0:45   |     0     |

Then I want to extract all rows with the two latest dates. Alternativ with the latest date, the three latest dates etc.
A dataframe (with the two latest dates) like this should be the target:

|  No   |   Date     |   hours  |    Flag   |
|:------|:-----------|:---------|:----------|
|  101  | 02.03.2021 |   1:20   |     0     |
|  101  | 02.03.2021 |   2:40   |     0     |
|  101  | 19.06.2021 |   2:05   |     0     |
|  101  | 19.06.2021 |   1:45   |     0     |
|  101  | 19.06.2021 |   1:25   |     0     |
|  101  | 19.06.2021 |   0:45   |     0     |

I tried:

df.nlargest(2, 'Date', keep='all')

this produces sth. like this:

|  No   |   Date     |   hours  |    Flag   |
|:------|:-----------|:---------|:----------|
|  101  | 19.06.2021 |   2:05   |     0     |
|  101  | 19.06.2021 |   1:45   |     0     |
|  101  | 19.06.2021 |   1:25   |     0     |
|  101  | 19.06.2021 |   0:45   |     0     |

Can anybody help me please?
Thanks.

CodePudding user response：

Filter 2 largest Datetimes by Series.nlargest and then filter all Dates which match by Series.isin and boolean indexing:

df['Date'] = pd.to_datetime(df['Date'], dayfirst=True)

df = df[df['Date'].isin(df['Date'].nlargest(2))]

print (df)
    No       Date hours  Flag
6  101 2021-06-19  2:05     0
7  101 2021-06-19  1:45     0
8  101 2021-06-19  1:25     0
9  101 2021-06-19  0:45     0

CodePudding user response：

@PaulS yes this works, Thank you