Combining queries into one final table with business date calculation

Question

I am new to sql and I have two measures that were done in tableau. The measure is as followed

{ FIXED [Client ID]: MAX(IF [Appt Status] = 'Complete' AND [Appointment Type Roll-Up] = '1' THEN [Appt Date]END)}

{ FIXED [Client ID]: MIN(IF[Appointment Status Roll-UP]='Active' and [Appt Date] > [Last Direct Trmt Appt] then [Appt Date]END)}

These measures, I converted into SQL as followed:

select 
a.client_id,
Max(a.appointment_date) as max_date
From #Last_Direct_treatment A
where  a.[appointment_status] = 'Complete' 
and  a.[Appointment_type_roll_up] = '1'
Group by client_id


select 
a.client_id,
min(a.appointment_date) as min_date
From #Last_Direct_treatment A
where  a.[Appointment_Status_Roll_up] = 'active' 
and a.appointment_date > a.last_direct_only_date
Group by client_id

In SQL, I already created the first query looking for my main fields:

SELECT 
s.[appointment_id]
,c.[full_name]
,s.[client_id]
,s.[case_number]
,s.[appointment_date]
,s.[appointment_type]
,s.[appointment_status]
,ca.[last_direct_only_date]
,t.[authorization_status]
,t.[service_type]
,case 
when s.appointment_status = '*_Cancelled' then 'Cancelled'
when s.appointment_status = '*_Complete' then 'Complete'
when s.appointment_status = 'Complete' then 'Complete'
else 'Active'
end as Appointment_Status_Roll_up
,case when
s.[appointment_type] like '%IND%' or s.[appointment_type] like '%indirect%' then '0' else '1'
end as Appointment_type_roll_up

Into #Last_Direct_treatment 
     
FROM [appointment] s
INNER JOIN [client] c
ON s.[client_id] = c.[client_id]
INNER JOIN [client_case] ca
ON c.[client_id] = ca.[client_id]
INNER JOIN [authorization] t
ON ca.[case_number] = t.[case_number]

ORDER BY appointment_date DESC

I was successful in making the following queries find the min and max date of the fields I needed.

My desired result would be to incorporate these two new fields into one final table where I have these as the last two columns and a business date calculation of the two columns as a new column with the fields from the first query.

select 
a.client_id,
Max(a.appointment_date) as max_date
From #Last_Direct_treatment A
where  a.[appointment_status] = 'Complete' 
and  a.[Appointment_type_roll_up] = '1'
Group by client_id


select 
a.client_id,
min(a.appointment_date) as min_date
From #Last_Direct_treatment A
where  a.[Appointment_Status_Roll_up] = 'active' 
and a.appointment_date > a.last_direct_only_date
Group by client_id

Answer 1

The original measures appear to resemble case expressions rather than a set of where clause predicates , eg

CASE WHEN MAX(IF [Appt Status] = 'Complete' 
         AND [Appointment Type Roll-Up] = '1' THEN [Appt Date] END

CASE WHEN MIN(IF[Appointment Status Roll-UP]='Active' 
         AND [Appt Date] > [Last Direct Trmt Appt] THEN [Appt Date] END

and in a combined query these might be as follows:

SELECT
      a.client_id
    , CASE 
        WHEN a.[appointment_status] = 'Complete'
            AND a.[Appointment_type_roll_up] = '1'
            THEN Max(a.appointment_date)
        END AS max_date
    , CASE 
        WHEN a.[Appointment_Status_Roll_up] = 'active'
            AND a.appointment_date > a.last_direct_only_date
            THEN min(a.appointment_date)
        END AS min_date
FROM #Last_Direct_treatment A

In the unlikely event that you need those dates in a single column, note that case expressions allow for multiple sets of conditions to be considered, eg:

SELECT
      a.client_id
    , CASE 
        WHEN a.[appointment_status] = 'Complete'
            AND a.[Appointment_type_roll_up] = '1'
            THEN Max(a.appointment_date)
        WHEN a.[Appointment_Status_Roll_up] = 'active'
            AND a.appointment_date > a.last_direct_only_date
            THEN min(a.appointment_date)
        -- ELSE ...
        END AS min_or_max_date
FROM #Last_Direct_treatment A

These condition sets are evaluated top-to bottom, if one applies that determines the result and the condition sets that follow are not used. If no condition sets apply else may be used.