data.sort_values(by=['Date'], ascending=False) doesn't works

Question

I need to plot some data series in x axis and I have a csv file which "Start time" colum is full of dates. As I work with DataFrame, I use pandas library to manipulate the data. My datetime data is:

Input:

print(paradas["Start time"])
    

Output:

0        31/12/2020 00:13:30
1        30/12/2020 19:30:00
2        30/12/2020 19:01:45
3        30/12/2020 19:00:10
4        30/12/2020 18:55:35
                ...         
10704    02/01/2020 08:37:33
10705    02/01/2020 08:32:33
10706    02/01/2020 08:28:03
10707    02/01/2020 08:19:03
10708    31/12/2019 02:41:01
Name: Start time, Length: 10709, dtype: object

As I am working with time data I transform to datetime64[ns] class all the timestamps from the column:

Input:

paradas["Start time"]=pd.to_datetime(paradas["Start time"])
print(paradas["Start time"])

Output:

0       2020-12-31 00:13:30
1       2020-12-30 19:30:00
2       2020-12-30 19:01:45
3       2020-12-30 19:00:10
4       2020-12-30 18:55:35
                ...        
10704   2020-02-01 08:37:33
10705   2020-02-01 08:32:33
10706   2020-02-01 08:28:03
10707   2020-02-01 08:19:03
10708   2019-12-31 02:41:01
Name: Start time, Length: 10709, dtype: datetime64[ns]

Now, since the dates are resversed, I tried to put them backwards by using:

Input:

paradas["Start time"]=paradas["Start time"].sort_values(by=['Date'], ascending=False)
print(paradas["Start time"])

However it doesn't recognise my code because of the 'by':

Output:

TypeError                                 Traceback (most recent call last)
<ipython-input-35-d4f349ab2092> in <module>()
    126 #print(paradas["Start time"])
    127 
--> 128 paradas["Start time"]=paradas["Start time"].sort_values(by=['Date'], ascending=False)
    129 print(paradas["Start time"])
    130

TypeError: sort_values() got an unexpected keyword argument 'by'

Also, I tried to evaluate it without the arguments, but it doesn't change anything eitherway.

So I don't know what I'm doing wrong, if it's the type of the elements or what. I read on another post about doing it with str, but since I need the datetime format, and I've already sawn other codes evaluating this project with datetime64[ns], I am almost sure that it is possible...

Answer 1

Ok, I solved it.

To whom it may concern, the problem was that paradas["Start time"]=paradas["Start time"].sort_values(by=['Date'], ascending=False) wasn't right because sort_values() it's not prepared to operate by calling just one column from the DataFrame. In particular, my data (writen as paradas) is in pandas.Dataframe format, and paradas["Start time"] (which is just one column from paradas) is in pandas.Seriers format.

We need to use sort_values() with the Dataframe format, so I needed to apply this comand to all my data, which means that we must use paradas :

Input:

paradas=paradas.sort_values(by=['Date'], ascending=False)

Output:

0        31/12/2020 00:13:30
10708    31/12/2019 02:41:01
2026     31/10/2020 05:04:06
2027     31/10/2020 04:59:06
2028     31/10/2020 04:57:46
                ...         
7642     01/04/2020 01:36:15
7643     01/04/2020 01:23:40
7644     01/04/2020 01:11:20
7645     01/04/2020 00:14:20
7646     01/04/2020 00:08:25
Name: Start time, Length: 10709, dtype: object

Nevertheless, it didn't sort like I wanted, so I just realised that if I needed to sort backwards the "Start time" column, so I finally did:

Input:

paradas=paradas.reindex(index=paradas["Start time"].index[::-1])

Output:

10708   2019-12-31 02:41:01
10707   2020-02-01 08:19:03
10706   2020-02-01 08:28:03
10705   2020-02-01 08:32:33
10704   2020-02-01 08:37:33
                ...        
4       2020-12-30 18:55:35
3       2020-12-30 19:00:10
2       2020-12-30 19:01:45
1       2020-12-30 19:30:00
0       2020-12-31 00:13:30
Name: Start time, Length: 10709, dtype: datetime64[ns]

(Now I've already changed into datetime64[ns] format) This is what it worked for me. I've just checked several times the documentation:

https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.DataFrame.sort_values.html

and compared it with:

https://pandas.pydata.org/pandas-docs/version/0.22/generated/pandas.Series.sort_values.html