Question

So i have a temporary table that I need to populate with the records that have to deal with enrollment history

I work with a dataset that looks like this:

sccphidsid    enroll_date disenroll_date    status
abc123x        2009-01-01  2010-31-12           0
abc123x        2011-01-01  null                 0
abc123x        2011-03-01  2012-01-01           0

So I need it to return all records with the same sccphidsid and return them including the null and not null values.

The conditions it has to meet are

• Disenroll dates are not null
• Disenroll dates are null and
• records have multiple disenroll_dates

Current Solution:

select *, count(disenroll_date is null)
              from enrollment _test
              where Status = 0
              group by Sub_Client_Cd, Policy_Holder_ID, Suffix_ID
              having count(disenroll_date is null) > 1;

However this is returning too many records, is there a way to isolate this query down, maybe simpler, that can prevent this from returning too many records by constraints or parameters?

Thanks all

Answer 1

SELECT Sub_Client_Cd, Policy_Holder_ID, Suffix_ID,
       SUM(CASE WHEN disenroll_date IS NULL THEN 1 ELSE 0 END) AS NullDateCount,
       SUM(CASE WHEN disenroll_date IS NOT NULL THEN 1 ELSE 0 END) AS NotNullDateCount
    FROM enrollment_test
    WHERE Status = 0
    GROUP BY Sub_Client_Cd, Policy_Holder_ID, Suffix_ID
    HAVING COUNT(*) > 1;