You are right, I dropped BRIN and created btree and the performance on 0
rows matching criteria table is good, below is the plan with BTREE. I will
test by inserting lot of data.
Hash Join (cost=50186.91..3765911.10 rows=5397411 width=291) (actual
time=1.501..1.504 rows=0 loops=1)
Hash Cond: (
Hi
pá 4. 6. 2021 v 10:32 odesílatel Ayub Khan napsal:
> BRIN index is only on the date_time column, I even tried with btree index
> with no performance gains.
>
-> Bitmap Heap Scan on restaurant_order ro
(cost=5427.94..3353966.60 rows=19275986 width=108) (actual
time=1036.793..1
BRIN index is only on the date_time column, I even tried with btree index
with no performance gains.
On Fri, Jun 4, 2021 at 11:23 AM Pavel Stehule
wrote:
>
>
> pá 4. 6. 2021 v 10:07 odesílatel Ayub Khan napsal:
>
>>
>> below query is slow even with no data
>>
>>
>> explain ANALYZE
>>
>> WITH bu
pá 4. 6. 2021 v 10:07 odesílatel Ayub Khan napsal:
>
> below query is slow even with no data
>
>
> explain ANALYZE
>
> WITH business AS( SELECT * FROM get_businessday_utc_f() start_date)
> SELECT ro.order_id,
> ro.date_time,
> round(ro.order_amount, 2) AS order_amount,
> b.branch_id
below query is slow even with no data
explain ANALYZE
WITH business AS( SELECT * FROM get_businessday_utc_f() start_date)
SELECT ro.order_id,
ro.date_time,
round(ro.order_amount, 2) AS order_amount,
b.branch_id,
b.branch_name,
st_x(b.location) AS from_x,
st_y(b.location
