How to specify the special time in one column using python - python-3.x

Question

Here I have a dataset with on input and date and time. Here I just want to convert time into 00:00:00 for specific value which is contain in input column, and other time will be display as it is.
Then I wrote the code for that.
Then what I want is specify that 00:00:00 only. So I wrote the code for it and got an error `'RangeIndex' object has no attribute 'strftime'"
Can anyone help me to solve this ?
My code :
df['time_diff']= pd.to_datetime(df['date'] + " " + df['time'],
format='%d/%m/%Y %H:%M:%S', dayfirst=True)
mask = df['x3'].eq(5)
df['Duration'] = np.where(df['x3'].eq(5), np.timedelta64(0), pd.to_timedelta(df['time']))
Then I got the output:
date time x3 Duration
0 10/3/2018 6:15:00 0 06:15:00
1 10/3/2018 6:45:00 5 00:00:00
2 10/3/2018 7:45:00 0 07:45:00
3 10/3/2018 9:00:00 0 09:00:00
4 10/3/2018 9:25:00 0 09:25:00
5 10/3/2018 9:30:00 0 09:30:00
6 10/3/2018 11:00:00 0 11:00:00
7 10/3/2018 11:30:00 0 11:30:00
8 10/3/2018 13:30:00 0 13:30:00
9 10/3/2018 13:50:00 5 00:00:00
10 10/3/2018 15:00:00 0 15:00:00
11 10/3/2018 15:25:00 0 15:25:00
12 10/3/2018 16:25:00 0 16:25:00
13 10/3/2018 18:00:00 0 18:00:00
14 10/3/2018 19:00:00 0 19:00:00
15 10/3/2018 19:30:00 0 19:30:00
16 10/3/2018 20:00:00 0 20:00:00
17 10/3/2018 22:05:00 0 22:05:00
18 10/3/2018 22:15:00 5 00:00:00
19 10/3/2018 23:40:00 0 23:40:00
20 10/4/2018 6:58:00 5 00:00:00
21 10/4/2018 13:00:00 0 13:00:00
22 10/4/2018 16:00:00 0 16:00:00
23 10/4/2018 17:00:00 0 17:00:00
Then I want to specify this 00:00:00 time only then :
match_time="00:00:00"
time = data['duration'].loc[data.index.strftime("%H:%M:%S") == match_time]
Got an error :
Expected output :
time
00:00:00
00:00:00
Just read only 00:00:00 time
My csv :
subset:
date time x3
10/3/2018 6:15:00 0
10/3/2018 6:45:00 5
10/3/2018 7:45:00 0
10/3/2018 9:00:00 0
10/3/2018 9:25:00 0
10/3/2018 9:30:00 0
10/3/2018 11:00:00 0
10/3/2018 11:30:00 0
10/3/2018 13:30:00 0
10/3/2018 13:50:00 5
10/3/2018 15:00:00 0
10/3/2018 15:25:00 0
10/3/2018 16:25:00 0
10/3/2018 18:00:00 0
10/3/2018 19:00:00 0
10/3/2018 19:30:00 0
10/3/2018 20:00:00 0
10/3/2018 22:05:00 0
10/3/2018 22:15:00 5
10/3/2018 23:40:00 0
10/4/2018 6:58:00 5
10/4/2018 13:00:00 0
10/4/2018 16:00:00 0
10/4/2018 17:00:00 0
My csv file :
My csv file

Answer 1

Because types of values in column Duration are timedeltas, compare by string converted to timedelta too:
print (data['Duration'].dtype)
#timedelta64[ns]
match_time="00:00:00"
time = data[data['Duration'] == pd.to_timedelta(match_time)]
print (time)
date time x3 Duration
1 10/3/2018 6:45:00 5 0 days
9 10/3/2018 13:50:00 5 0 days
18 10/3/2018 22:15:00 5 0 days
20 10/4/2018 6:58:00 5 0 days
EDIT: If always timedeltas are less like 1 day:
First convert timedeltas to strings - added 0 days:
print (df['Duration'].astype(str))
#0 0 days 06:15:00.000000000
#1 0 days 00:00:00.000000000
#2 0 days 07:45:00.000000000
#3 0 days 09:00:00.000000000
#4 0 days 09:25:00.000000000
#5 0 days 09:30:00.000000000
#6 0 days 11:00:00.000000000
#7 0 days 11:30:00.000000000
#8 0 days 13:30:00.000000000
#9 0 days 00:00:00.000000000
#10 0 days 15:00:00.000000000
#11 0 days 15:25:00.000000000
#12 0 days 16:25:00.000000000
#13 0 days 18:00:00.000000000
#14 0 days 19:00:00.000000000
#15 0 days 19:30:00.000000000
#16 0 days 20:00:00.000000000
#17 0 days 22:05:00.000000000
#18 0 days 00:00:00.000000000
#19 0 days 23:40:00.000000000
#20 0 days 00:00:00.000000000
#21 0 days 13:00:00.000000000
#22 0 days 16:00:00.000000000
#23 0 days 17:00:00.000000000
#Name: Duration, dtype: object
And then remove first and last part of strings by slicing:
print (df['Duration'].astype(str).str[-18:-10])
#0 06:15:00
#1 00:00:00
#2 07:45:00
#3 09:00:00
#4 09:25:00
#5 09:30:00
#6 11:00:00
#7 11:30:00
#8 13:30:00
#9 00:00:00
#10 15:00:00
#11 15:25:00
#12 16:25:00
#13 18:00:00
#14 19:00:00
#15 19:30:00
#16 20:00:00
#17 22:05:00
#18 00:00:00
#19 23:40:00
#20 00:00:00
#21 13:00:00
#22 16:00:00
#23 17:00:00
#Name: Duration, dtype: object
df['Duration'] = df['Duration'].astype(str).str[-18:-10]
match_time="00:00:00"
time = df[df['Duration'] == match_time]
print (time)
date time x3 Duration
1 10/3/2018 6:45:00 5 00:00:00
9 10/3/2018 13:50:00 5 00:00:00
18 10/3/2018 22:15:00 5 00:00:00
20 10/4/2018 6:58:00 5 00:00:00
Solution for all timedeltas:
def f(x):
ts = x.total_seconds()
hours, remainder = divmod(ts, 3600)
minutes, seconds = divmod(remainder, 60)
return ('{:02d}:{:02d}:{:02d}').format(int(hours), int(minutes), int(seconds))
df['Duration'] = df['Duration'].apply(f)
match_time="00:00:00"
time = df[df['Duration'] == match_time]

Answer 2

You are trying to convert the dataframe index (0,1,2,...23) to a string time format object, and not the items content in the 'Duration' column.
First convert each item in the column 'Duration' then compare it to 'match_time' and finally save the resulting sliced frame, all at once:
match_time="00:00:00"
df=data.loc[data['Duration'].apply(lambda x: x.strftime("%H:%M:%S"))==match_time]
Then you will get all indexes which match your desired 'match_time' as follows:
date time x3 Duration
1 2018-10-03 2018-10-03 00:00:00 5 00:00:00
9 2018-10-03 2018-10-03 00:00:00 5 00:00:00
18 2018-10-03 2018-10-03 00:00:00 5 00:00:00
20 2018-10-04 2018-10-04 00:00:00 5 00:00:00