Adding dates for unique EIDs row-wise to get Years of Service

I have the following dataset:

EID	Company	Start_Date	End_Date	T_F
A111	ABC	2015-07-20	NaT	True
B111	DEF	1983-06-01	NaT	False
B111	ABC	2017-01-01	NaT	True
C111	GHI	1980-10-01	1981-08-31	True
D111	JKL	1973-05-01	1977-11-30	True
E111	ABC	2006-04-24	NaT	True
F111	ABC	1991-06-10	1994-12-15	False
F111	MNO	1994-12-01	2002-08-31	False
F111	ABC	2002-08-01	NaT	True
G111	ABC	1979-01-01	NaT	True
H111	ABC	2002-02-01	NaT	True

The expected output is as follows:

EID	Company	Start_Date	End_Date	T_F	YoS
A111	ABC	2015-07-20	NaT	True	NaN
B111	DEF	1983-06-01	NaT	False	(2017-01-01) - (1983-06-01)
B111	ABC	2017-01-01	NaT	True	NaN
C111	GHI	1980-10-01	1981-08-31	True	(1981-08-31) - (1980-10-01)
D111	JKL	1973-05-01	1977-11-30	True	(1977-11-30) - (1973-05-01)
E111	ABC	2006-04-24	NaT	True	NaN
F111	ABC	1991-06-10	1994-12-15	False	(2002-08-01) - (1991-06-10)
F111	MNO	1994-12-01	2002-08-31	False	NaN
F111	ABC	2002-08-01	NaT	True	NaN
G111	ABC	1979-01-01	NaT	True	NaN
H111	ABC	2002-02-01	NaT	True	NaN

This is what I am trying to do:

1. Where an EID has only one record and company is ABC YoS column should be NULL. End_Date is always blank in these cases.
2. Where an EID has multiple records and his/her last record is company ABC then YoS column will be Start date of first company - Start date of ABC company.
3. Where an EID has only one record and company is not ABC then YoS will be calculated as End_Date - Start_Date
4. Only the first record will have YoS value other records will contain NaN value.
5. If an employee has multiple records 99% of the times employees last records will be ABC company.

I tried the following code but this is only half part (or incorrect) I believe:

 result.loc[~(result.CLEAN_NAME == 'HONEYWELL / HON') & (result.T_F == False),'Hon_StartDate'] = result['Start_Date']

Any leads would be appreciated. Thanks.!

from Recent Questions - Stack Overflow https://ift.tt/3fJZovU
https://ift.tt/eA8V8J