PostgreSQL的。 改善查询。 WITH

Question

how can I improve my query?

WITH womenRating AS ( SELECT count(train_type) as trainCount, train_type FROM trainRoutine
INNER JOIN client ON client.id = trainRoutine.client_id
WHERE client.gender = 'Woman'
GROUP BY train_type
ORDER BY trainCount DESC ),
     menRating AS (
SELECT count(train_type) as trainCount, train_type FROM trainRoutine
INNER JOIN client ON client.id = trainRoutine.client_id
WHERE client.gender = 'Man'
GROUP BY train_type
ORDER BY trainCount DESC )
     select a.train_type, DENSE_RANK() OVER (ORDER BY a.traincount DESC) as women, DENSE_RANK() OVER (ORDER BY b.traincount DESC) as men
FROM womenRating a
FULL OUTER JOIN menRating b ON a.train_type = b.train_type;

There is a problem because of

FULL OUTER JOIN menRating b ON a.train_type = b.train_type

If the left table "womenRating" will not have any value of train_type that the second table has, there will be a empty field, and my DENSE_RANK() will count that row as №1

Answer 1

Use conditional aggregation:

SELECT train_type,
       sum( (c.gender = 'Woman')::int) as women_count,
       sum( (c.gender = 'Man')::int) as men_count,
       dense_rank() over (order by sum( (c.gender = 'Woman')::int) desc) as women_rank,
       dense_rank() over (order by sum( (c.gender = 'Man')::int) desc) as men_rank
FROM trainRoutine tr INNER JOIN
     client c
     ON c.id = tr.client_id
GROUP BY train_type
ORDER BY trainCount DESC