如何优化/重新考虑分区查询的排名

Question

I have a fairly straight forward query that takes forever to run - roughly a second per row that's returned. The issue is the rank over partition approach, but I'm not sure how to rewrite it (cross apply?)

Any pointers would be appreciated.

Query:

Select orders.orderid
    ,prescriptions.patientid as Rx_PtID
    ,prescriptions.prescriptiontypeid
    ,prescriptions.prescriptionid
    ,letter_history.employeename
    ,RANK() OVER (PARTITION BY prescriptions.prescriptionid ORDER BY letter_history.letter_date ASC) AS orank --Assigns ID based on first printing
from letter_history,orders,prescriptions
where letter_history.description LIKE CONCAT('%', orders.orderid, '%')
    AND prescriptions.patientid = orders.patientid
    AND orders.ordertypeid = 999                    --Exam Orders Only
    AND CONVERT(varchar(10),orders.orderdate, 120) = CONVERT(varchar(10),prescriptions.rxdate, 120)
    AND CAST(orders.ordereddate as date) >= @DateFrom
    AND CAST(orders.ordereddate as date) <= @DateTo

Answer 1

The only reason the rank() seems slow is that all the data needs to be processed in order for it to return results.

If you run the query without the rank() it will look fast -- because it is faster to return the first few results than the whole result set. You could check this out by make this a subquery and using count(*) to see the performance.

Also: Learn to use proper, explicit, standard JOIN syntax. Never use commas in the FROM clause.