Need to find Three most expensive path on average having code 'A' with delivery in May 2021 - mysql

Question

Statement:
Need to find Three most expensive path(from_to_ in table) on average that code (with_ in table) 'A' with delivery in May 2021? If two ties then include both.
Schema:
'observed_on', date,
'from_', varchar(3),
'to_', varchar(3),
'from_to_', varchar(8),
'with_', varchar(3),
'cart_no', varchar(8),
'deliver_on', date,
'd_charge', double,
Sample data:
click to view
Solution I tried:
SELECT
from_to_
,avg_price
FROM
(
SELECT
from_to_
,ROUND(AVG(d_charge),2) AS avg_price
,DENSE_RANK() OVER(ORDER BY ROUND(AVG(d_charge),2) DESC) rank_by_avgp
FROM
(
SELECT
*
FROM DELIVERY
WHERE deliver_on BETWEEN '2021-05-01' AND '2021-05-30'
AND with_ = 'A'
) AS A
GROUP BY from_to_
) AS bb
WHERE bb.rank_by_avgp <=3;
I know it's a workaround so I am looking for a better solution

Answer 1

#nishant, you have a quite poorly asked question here.
The issues with your question
The sample data, you reference, is a picture? Why not give text values or even better the DML statements to set the example up.
In the referenced dataset(picture) all the DELIVERY.deliver_on date values are before 2010. So the conditions you have here deliver_on BETWEEN '2021-05-01' AND '2021-05-30' will just not return anything for the example data.
If you say
I know it's a workaround so I am looking for a better solution
then based on what? It is a workaround for what? It looks to be producing correct results, so what is the problem with it? Do you want it to perform better or what?
You are not specifing the DB version. The different MySQL versions can have different solutions.
One possible solution
The example dataset setup:
DDL
CREATE TABLE DELIVERY (
observed_on DATE
, from_ VARCHAR(3)
, to_ VARCHAR(3)
, from_to_ VARCHAR(8)
, with_ VARCHAR(3)
, cart_no VARCHAR(8)
, deliver_on DATE
, d_charge DOUBLE
);
DML
INSERT INTO DELIVERY VALUES ('2012-01-19','Aus','Nzl','AusNzl','A','2118','2021-04-19',82.3);
INSERT INTO DELIVERY VALUES ('2012-01-19','Aus','Nzl','AusNzl','A','2118','2021-05-19',82.3);
INSERT INTO DELIVERY VALUES ('2012-01-19','Aus','Nzl','AusNzl','A','2118','2021-05-19',82.3);
INSERT INTO DELIVERY VALUES ('2013-01-19','Ind','Sla','IndSla','B','2233','2021-05-19',70.32);
INSERT INTO DELIVERY VALUES ('2013-01-19','Ind','Sla','IndSla','A','2233','2021-05-19',70.32);
INSERT INTO DELIVERY VALUES ('2013-01-19','Eur','Usa','EurUsa','C','2434','2021-05-19',67.53);
INSERT INTO DELIVERY VALUES ('2013-01-19','Eur','Usa','EurUsa','A','2434','2021-05-19',67.53);
INSERT INTO DELIVERY VALUES ('2013-01-19','Xyz','Usa','XyzUsa','A','2434','2021-05-19',67.53);
INSERT INTO DELIVERY VALUES ('2013-01-19','Xyz','Sla','XyzSla','A','2434','2021-05-19',67.51);
INSERT INTO DELIVERY VALUES ('2012-01-19','Aus','Nzl','AusNzl','A','2323','2021-05-19',82.3);
INSERT INTO DELIVERY VALUES ('2012-01-19','Aus','Nzl','AusNzl','A','2118','2021-06-19',82.3);
QUERY
SELECT from_to_
, avg_d_charge
, denserank_avg_d_charge
FROM /*SUB_to_calculate_the_denserank*/
(
SELECT from_to_
, ROUND(avg_d_charge, 2) AS avg_d_charge
, DENSE_RANK() OVER (
ORDER BY ROUND(avg_d_charge, 2) DESC
) denserank_avg_d_charge /*dense ranking*/
FROM /*SUB_to_calculate_the_averages*/
(
SELECT from_to_
, ROW_NUMBER() OVER (PARTITION BY from_to_) AS rownumber /*To filter for only one row per from_to_.*/
, AVG(d_charge) OVER (PARTITION BY from_to_) AS avg_d_charge /*Average caldulation*/
FROM /*DELIVERY*/
DELIVERY
WHERE 1 = 1
AND with_ = 'A' /* The "code" filter*/
AND DATE_SUB(deliver_on, INTERVAL DAYOFMONTH(deliver_on) - 1 DAY) = '2021-05-01' /* The 2021-05 filter*/
) SUB_to_calculate_the_averages
WHERE 1 = 1
AND rownumber = 1
) SUB_to_calculate_the_denserank
WHERE 1 = 1
AND denserank_avg_d_charge < 4;
The only main difference here from your solution is then that I do not use the aggegate GROUP BY here, only analytical functions. I prefer this quite often as it allows to carry later other attributes through the query without the need to apply the aggregate functions on them etc. But in the end then this comes down to performance and the requirements what/how should be done.