Query returns incorrect result - sql

Question

MY input table is patient_ID, and Appt_Resource_ID (Doctor) ( the second table is just getting the patient name )
Patient_ID Appt_Resource_ID
88299 47
88299 1
88299 40
88299 40
88299 40
88299 40
I am running an sql that is meant to write an output row for the patient_id and the Doctor_ID that occurs the most, in this case sb 40. But it's outputting Doctor_ID 1. Other cases I checked are doing correctly.
This is the query:
select distinct A.Patient_id, P.Patient_name, b.Appt_resource_id
from [PM].[vwGenPatApptInfo] A
inner join
(
select top 100 percent patient_id, Appt_resource_id, count(Appt_resource_id) as DR_count,
row_number() over (partition by patient_id order by count(*) desc) as seqnum
from [PM].[vwGenPatApptInfo]
where Patient_ID is NOT NULL
group by patient_id,Appt_resource_id
order by patient_id, seqnum
) B on B.Patient_ID = A.Patient_ID
and B.seqnum = 1
inner join [PM].[vwGenPatInfo] P on A.Patient_id = P.Patient_id
where A.Appt_DateTime >= DATEADD(yyyy, -2, GETDATE()) and A.Appt_Cancelled_Date is NULL
But the results are this:
Patient_ID Appt_Resource_ID
88299 1

Answer 1

Try this.
SELECT TOP 1 *
FROM (
SELECT Patient_ID, Appt_Resource_ID, COUNT(*) AS theCount
FROM vwGenPatApptInfo
GROUP BY Patient_ID, Appt_Resource_ID
WHERE ApptDateTime >= DATEADD(yyyy, -2, GETDATE()) and Appt_Cancelled_Date is NULL
) b
JOIN vwGenPatInfo p ON p.Patient_ID = b.Patient_ID
ORDER BY theCount DESC

Answer 2

This is not really an answer, sorry.
Your subquery finds the doctor that occurs most often for a patient. (If you want the most often occurring patient/doctor pair instead, then remove the partition clause from ROW_NUMBER.)
But then you join vwGenPatApptInfo again (as A). Why? And your join is even incomplete, because you only join by Patient_ID and not by Appt_resource_id, too. You are thus combining every most occurring doctor with other doctor records for the patient in a certain date range. These you dismiss again with DISTINCT.
I don't see however how your query gets the wrong doctor. If it had been A.Appt_resource_id you selected, that would have been a reason, but b.Appt_resource_id should be the most often occurring doctor.
I hope my comments above help you, though. Maybe you can simplify your query somehow and maybe get thus even rid of the error.
Good luck!

Answer 3

It might be not the most ideal solution, but try this:
;WITH CTE1 as (
SELECT Patient_ID, Appt_Resource_ID
, COUNT(*) OVER (PARTITION BY Patient_ID, Appt_Resource_ID) as CNT
FROM [PM].[vwGenPatApptInfo]
), CTE2 as (
SELECT Patient_ID, Appt_Resource_ID, CNT, RANK() OVER
(PARTITION BY Patient_ID ORDER BY CNT DESC) AS Rank
FROM CTE1
), CTE3 as (
SELECT Patient_ID, Appt_Resource_ID, CNT, Rank
FROM CTE2
WHERE Rank = 1
GROUP BY Patient_ID, Appt_Resource_ID, CNT, Rank
)
select distinct A.Patient_id, P.Patient_name, b.Appt_resource_id, B.CNT
from [PM].[vwGenPatApptInfo] A
INNER JOIN CTE3 as B on B.Patient_ID = A.Patient_ID
inner join [PM].[vwGenPatInfo] P on A.Patient_id = P.Patient_id
where A.Appt_DateTime >= DATEADD(yyyy, -2, GETDATE()) and A.Appt_Cancelled_Date is NULL