Postgres 查询需要很长时间才能获得结果

Question

I have a problem with my postgres query. The query taking too long, almost 27 seconds to get the results and i don't know what i need to do to solve

SELECT 
    sales.*, 
    customers.name AS customer_name, 
    customers.cpf_pf AS cpf, 
    customers.cnpj_pj AS cnpj, 
    customers.juridica AS tipo_pessoa, 
    status_sales.name AS status_name, 
    sales_status_histories.created_at AS data_modificacao, 
    sales_status_histories.obs as observacao, 
    users.name AS nome_vendedor

FROM sales 

RIGHT JOIN customers ON sales.customer_id = customers.id
RIGHT JOIN sales_user_owners ON sales_user_owners.sale_id = sales.id
RIGHT JOIN sales_status_histories ON sales_status_histories.sale_id = sales.id
RIGHT JOIN status_sales ON status_sales.id = sales_status_histories.status_sale_id
RIGHT JOIN users ON users.id = sales_user_owners.user_id 

WHERE sales_status_histories.id IN (SELECT id FROM ( SELECT *, row_number() over (PARTITION BY sale_id ORDER BY created_at DESC) AS rn FROM sales_status_histories) AS ssh WHERE rn=1 AND Date( (sales_status_histories.created_at at time zone 'UTC') at time zone 'America/Sao_Paulo') BETWEEN '2018-01-01 00:00:00' AND '2018-01-04 23:59:59')
    AND sales_user_owners.id IN ( SELECT id FROM ( SELECT *, row_number() over (PARTITION BY sale_id ORDER BY created_at DESC) AS rn FROM sales_user_owners) AS so WHERE rn=1)

The big problem is here:

WHERE sales_status_histories.id IN (SELECT id FROM ( SELECT *, row_number() over (PARTITION BY sale_id ORDER BY created_at DESC) AS rn FROM sales_status_histories) AS ssh WHERE rn=1 AND Date( (sales_status_histories.created_at at time zone 'UTC') at time zone 'America/Sao_Paulo') BETWEEN '2018-01-01 00:00:00' AND '2018-01-04 23:59:59')

This part searches the last status of the sale in the sales_status_histories table with specific range of time

The sale can has multiples statuses in different phases, but only the last is the current status and just the last is important to me in this query, because i need get all sales that were updated in specific range of time

Someone can help me to build a optimized query that take less time to get the same results, please?

See the entire structure of my database in this image:

My application is a ruby on rails with postgres

INDEXES:

--- EDIT 1 ----

Sometimes i need to find the sales with specific current status. Using the @bma suggestion i tried something like this:

(SELECT id, row_number() OVER (PARTITION BY sale_id ORDER BY created_at DESC) AS rn FROM sales_status_histories WHERE status_sale_id IN (2) AND DATE((sales_status_histories.created_at AT TIME ZONE 'UTC') AT TIME ZONE 'America/Sao_Paulo') BETWEEN '2018-01-01' AND '2018-01-04') AS ssh WHERE rn=1

As you can see, I added the status_sale_id IN (2), but in this case only sales that at some point (not always the current status) had that specific status are returned

Answer 1

Not a full answer, but a thing to try might be making the query more readable with a WITH clause to pull out the sales_status_histories rows you are interested in first, and then join the main SELECT on that.

Note: you can maybe use SELECT DISTINCT sale_id FROM sales_status_histories ORDER BY created_at DESC to get only the row with the latest change instead of a subquery with row_number(). I don't know if that's more efficient, but it's simpler to read.