Try to optimized a PL/SQL query - oracle

I am trying to improve the speed of the query below, but the required field cannot be changed. So I got stuck here. Please, help me out of this trap. A little hint or inspiration is also helpful!!
select cg.province_id,
(select count(distinct(c.guidance_user_id))
from case_guidance c
where c.guidance_status = '2'
and c.province_id = cg.province_id) as guidance_cnt,
(select count(distinct(c.guidance_user_id))
from case_guidance c
where c.guidance_status = '2'
and c.guidance_user_type = 'role.type.teacher'
and c.province_id = cg.province_id) as guidance_teacher_cnt,
(select count(distinct(c.guidance_user_id))
from case_guidance c
where c.guidance_status = '2'
and c.guidance_user_type = 'role.type.jyy'
and c.province_id = cg.province_id) as guidance_jyy_cnt,
(select count(distinct(c.guidance_user_id))
from case_guidance c
where c.guidance_status = '2'
and c.guidance_user_type = 'role.type.expert'
and c.province_id = cg.province_id) as guidance_expert_cnt,
(select count(distinct(c.case_id))
from case_guidance c
where c.guidance_status = '2'
and c.province_id = cg.province_id) as guidance_case_cnt
from case_guidance cg
where cg.province_id is not null
group by cg.province_id
order by guidance_cnt desc

Replace the correlated subqueries with CASE to eliminate all joins:
select
province_id,
count(distinct(case when guidance_status = '2' then guidance_user_id else null end)) guidance_cnt,
count(distinct(case when guidance_status = '2' and guidance_user_type = 'role.type.teacher' then guidance_user_id else null end)) guidance_teacher_cnt,
count(distinct(case when guidance_status = '2' and guidance_user_type = 'role.type.jyy' then guidance_user_id else null end)) guidance_jyy_cnt,
count(distinct(case when guidance_status = '2' and guidance_user_type = 'role.type.expert' then guidance_user_id else null end)) guidance_expert_cnt,
count(distinct(case when guidance_status = '2' then case_id else null end)) guidance_case_cnt
from case_guidance
group by province_id;
order by guidance_cnt desc
(I intentionally left the code lines extra long so that the conditions will line up. This helps make it clear what the difference between the columns are, and that they're all doing almost the exact same thing.)

Related

Oracle:if left join table return no rows then change to another left join table

I have two query as below and both of them work well. Now I need to combine them and if the first select find rows then no need to run the second select
1.
select f.*, o.org_sn, ttt.name linkman_name, ttt.phone phone_num
from td_archive_feedback f
left join TD_SM_ORG o
on f.recv_org_id = o.org_id
left join (--if this select returns no row , then change.if find matching rows then don't change
select wm_concat(linkman_name) name,
wm_concat(phone_num) phone,
org_id
from (select linkman_name, phone_num, LINK_ORG_ID, org_id
from TD_SM_LINKMAN
where STATE = '2'
and (LINK_ORG_ID = #link_org_id)) t
group by org_id) ttt
on ttt.org_id = o.org_id
--
left join td_sm_dict_item di
on o.org_level = di.item_id
where f.businessid = #businiessid
select f.*, o.org_sn, ttt.name linkman_name, ttt.phone phone_num
from td_archive_feedback f
left join TD_SM_ORG o
on f.recv_org_id = o.org_id
left join (--if the first left join return no rows ,then change to this
select wm_concat(linkman_name) name,
wm_concat(phone_num) phone,
t.org_id
from (
select linkman_name, phone_num, LINK_ORG_ID, org_id
from TD_SM_LINKMAN
where STATE = '2'
and (LINK_ORG_ID is null or LINK_ORG_ID = '')) t
group by t.org_id) ttt
on ttt.org_id = o.org_id
--
left join td_sm_dict_item di
on o.org_level = di.item_id
where f.businessid = #businessid;
Just like I remarked, if
select wm_concat(linkman_name) name,.... where STATE = '2'
and (LINK_ORG_ID = #link_org_id)) t
... returns no rows,
I want to change this to:
select wm_concat(linkman_name) name,.... where STATE = '2' and (LINK_ORG_ID is null or LINK_ORG_ID = '')) t.
Please help me, thanks.
Use Or condition:
select f.*, o.org_sn, ttt.name linkman_name, ttt.phone phone_num
from td_archive_feedback f
left join TD_SM_ORG o
on f.recv_org_id = o.org_id
left join (--if this select returns no row , then change
select wm_concat(linkman_name) name,
wm_concat(phone_num) phone,
org_id
from (select linkman_name, phone_num, LINK_ORG_ID, org_id
from TD_SM_LINKMAN
where STATE = '2'
and (
(LINK_ORG_ID = #link_org_id) or
(LINK_ORG_ID is null or LINK_ORG_ID = '')
) t
group by org_id) ttt
on ttt.org_id = o.org_id
--
left join td_sm_dict_item di
on o.org_level = di.item_id
where f.businessid = #businiessid

Subquery with Select statement works in 12C but not on 11g

Select
c.id AS case_id,
c.id AS case_number,
tt.description as Target_Type,
coalesce(
v_vendors.vendor_number,
vma.org_code,
TO_CHAR(sm.staff_member_id),
la.org_code,
oe.other_fns_number,
TO_CHAR(oe.other_id),
TO_CHAR(c_clients.client_id)
) as Target_Identifier,
coalesce(
v_vendors.vendor_name,
vma.name,
s_wic_users.username,
la.name,
oe.other_name,
UPPER(
c_clients.last_name || ', ' || c_clients.first_name
)
) as Target_Name,
c.date_opened AS Open_Date,
c.date_closed AS Close_Date,
t.target_type_id,
t.target_id,
s.description AS Status,
case when aIosm.last_name is null then '' else UPPER(
aIosm.last_name || ', ' || aIosm.first_name
) end Investigator,
case when aSosm.last_name is null then '' else UPPER(
aSosm.last_name || ', ' || aSosm.first_name
) end Supervisor,
c.Notes AS Supervisor_Comments,
aI.Date_Assigned as Assign_Date,
----subquery starts
(
select
lat.description
from
investigation_actions_taken iat
join investig_lu_action_type lat on iat.action_type_id = lat.id
where
iat.investigation_case_id = c.id
and iat.is_deleted <> 'T'
order by
iat.date_assessed desc,
iat.date_created desc OFFSET 0 ROWS FETCH NEXT 1 ROWS ONLY
) AS Last_Action
from
v_investigation_cases c
join investigation_target t on c.target_id = t.target_id
join v_lu_inv_status s on (
case when date_closed <= TRUNC(sysdate) then 2 else case when date_opened > TRUNC(sysdate) then 2 else 1 end end
) = s.status_id
join investigation_target_type tt on t.target_type_id = tt.target_type_id
---left join starts
left join investigation_case_assignment aI on c.id = aI.investigation_case_id
and aI.date_roll_off is null
and aI.assignment_type = 'I'
left join o_staff_members aIosm on aI.staff_member_id = aIosm.staff_member_id
left join investigation_case_assignment Super on c.id = Super.investigation_case_id
and Super.date_roll_off is null
and Super.assignment_type = 'S'
left join o_staff_members aSosm on Super.staff_member_id = aSosm.staff_member_id
-----left join starts
left join v_vendors on t.vendor_id = v_vendors.id
left join o_organizational_units vma on t.vendor_management_area_id = vma.seq_id
left join c_clients on t.client_id = c_clients.client_id
left join o_staff_members sm on t.user_id = sm.staff_member_id
left join s_wic_users on sm.staff_member_id = s_wic_users.sm_staff_member_id
left join investigation_other_entity oe on t.non_wic_vendor_id = oe.other_id
left join o_organizational_units la on t.local_agency_id = la.seq_id
---end of the query
Here's an article about what you can achieve in Oracle 12c with the OFFEST FETCH based on Kaushik Nayak.
http://www.dba-oracle.com/t_offset_fet_first_rows_only.htm
I hope it can help you to convert your original query to 11g. Good luck.

Oracle/SQL Query issue

I am sort of a newbie to oracle/sql. I am trying to pull from the same column different values and add them with some other information. The other information is not the issue it is trying to count and add here the problem comes in.
I am connecting to an oracle database. Here is what i have
SELECT
EV.PUBLIC_DESCRIPTION,
EV.EVENT_DATE,
ES.PRICE,
BT.BUYER_TYPE_CODE,
PCA.ADDR1,
PCA.ADDR2,
PCA.CITY,
PCA.POSTAL_CODE,
PCE.EMAIL,
PC.FORMATTED_NAME,
PCP.PHONE_NUMBER,
PCP.SECONDARY,
SUM(COUNT(CASE WHEN BT.BUYER_TYPE_CODE = 'GRADLT' THEN 1 ELSE 0 END) + COUNT(CASE WHEN BT.BUYER_TYPE_CODE = 'GRADTE' THEN 1 ELSE 0 END) + COUNT(CASE WHEN BT.BUYER_TYPE_CODE = 'GRSTND' THEN 1 ELSE 0 END) + COUNT(CASE WHEN BT.BUYER_TYPE_CODE = 'GSTDTE' THEN 1 ELSE 0 END) + COUNT(CASE WHEN BT.BUYER_TYPE_CODE = 'GROUDI' THEN 1 ELSE 0 END)) AS "Adults",
SUM(COUNT(CASE WHEN BT.BUYER_TYPE_CODE = 'GRCHILD' THEN 1 ELSE 0 END) + COUNT(CASE WHEN BT.BUYER_TYPE_CODE = 'GRCHTE' THEN 1 ELSE 0 END)) AS 'Paid Child',
SUM(COUNT(CASE WHEN BT.BUYER_TYPE_CODE = 'GRPCH' THEN 1 ELSE 0 END)) AS 'Free Child',
SUM(COUNT(CASE WHEN BT.BUYER_TYPE_CODE = 'GRCOMP' THEN 1 ELSE 0 END)) AS 'Comps'
FROM EVENT EV
INNER JOIN EVENT_SEAT ES ON EV.EVENT_ID = ES.EVENT_ID
INNER JOIN BUYER_TYPE BT ON ES.BUYER_TYPE_ID = BT.BUYER_TYPE_ID
INNER JOIN PATRON_ORDER PO ON ES.ORDER_ID = PO.ORDER_ID
INNER JOIN PATRON_ACCOUNT PA ON ES.ATTENDING_PATRON_ACCOUNT_ID = PA.PATRON_ACCOUNT_ID
INNER JOIN PATRON_CONTACT PC ON PA.PATRON_ACCOUNT_ID = PC.PATRON_ACCOUNT_ID
INNER JOIN PATRON_CONTACT_ADDRESS PCA ON PC.PATRON_ACCOUNT_ID = PCA.PATRON_ACCOUNT_ID
INNER JOIN PATRON_CONTACT_EMAIL PCE ON PCA.PATRON_ACCOUNT_ID = PCE.PATRON_ACCOUNT_ID
INNER JOIN PATRON_CONTACT_PHONE PCP ON PCE.PATRON_ACCOUNT_ID = PCP.PATRON_ACCOUNT_ID
GROUP BY EV.PUBLIC_DESCRIPTION, EV.EVENT_DATE
ORDER BY ES.TRANSACTION_ID DESC, PCP.SECONDARY DESC, PCP.PHONE_NUMBER DESC, PC.FORMATTED_NAME DESC, PCE.EMAIL DESC, PCA.POSTAL_CODE DESC, PCA.CITY DESC, PCA.ADDR2 DESC, PCA.ADDR1 DESC, BT.BUYER_TYPE_CODE DESC, ES.PRICE DESC;
any help would be greatly appreciated
You need to decide which technique you want to use, currently you are using 2 techniques and they are colliding.
For this you must know: COUNT() will increment by one for every NON-NULL value
So, to use COUNT() with a case expression do this
COUNT(CASE WHEN BT.BUYER_TYPE_CODE = 'GRSTND' THEN 1 END)
or
COUNT(CASE WHEN BT.BUYER_TYPE_CODE = 'GRSTND' THEN 1 ELSE NULL END)
OR, don't use COUNT(), use SUM() instead
SUM(CASE WHEN BT.BUYER_TYPE_CODE = 'GRSTND' THEN 1 ELSE 0 END)
To add conditions together, I suggest you use the case expression better
Instead of something like this:
, COUNT(CASE WHEN BT.BUYER_TYPE_CODE = 'GRADLT' THEN 1 END)
+ COUNT(CASE WHEN BT.BUYER_TYPE_CODE = 'GRADTE' THEN 1 END)
+ COUNT(CASE WHEN BT.BUYER_TYPE_CODE = 'GRSTND' THEN 1 END)
+ COUNT(CASE WHEN BT.BUYER_TYPE_CODE = 'GSTDTE' THEN 1 END)
+ COUNT(CASE WHEN BT.BUYER_TYPE_CODE = 'GROUDI' THEN 1 END) AS "Adults"
Use this:
COUNT(CASE WHEN BT.BUYER_TYPE_CODE IN ('GRADLT','GRADTE','GRSTND','GSTDTE','GROUDI') THEN 1 ELSE NULL END)
There is also an issue with your GROUP BY, which MUST contain ALL non-aggregating columns. I think your query should look more like this:
SELECT
EV.PUBLIC_DESCRIPTION
, EV.EVENT_DATE
, ES.PRICE
/* , BT.BUYER_TYPE_CODE */
, PCA.ADDR1
, PCA.ADDR2
, PCA.CITY
, PCA.POSTAL_CODE
, PCE.EMAIL
, PC.FORMATTED_NAME
, PCP.PHONE_NUMBER
, PCP.SECONDARY
, COUNT(CASE WHEN BT.BUYER_TYPE_CODE IN ('GRADLT', 'GRADTE', 'GRSTND', 'GSTDTE', 'GROUDI') THEN 1 ELSE NULL END) AS "Adults"
, COUNT(CASE WHEN BT.BUYER_TYPE_CODE IN ('GRCHILD', 'GRCHTE', 'GRPCH', 'GRCOMP') THEN 1 ELSE NULL END) AS "Free Child"
, COUNT(CASE WHEN BT.BUYER_TYPE_CODE = 'GRCOMP' THEN 1 ELSE NULL END) AS "Comps"
FROM EVENT EV
INNER JOIN EVENT_SEAT ES
ON EV.EVENT_ID = ES.EVENT_ID
INNER JOIN BUYER_TYPE BT
ON ES.BUYER_TYPE_ID = BT.BUYER_TYPE_ID
INNER JOIN PATRON_ORDER PO
ON ES.ORDER_ID = PO.ORDER_ID
INNER JOIN PATRON_ACCOUNT PA
ON ES.ATTENDING_PATRON_ACCOUNT_ID = PA.PATRON_ACCOUNT_ID
INNER JOIN PATRON_CONTACT PC
ON PA.PATRON_ACCOUNT_ID = PC.PATRON_ACCOUNT_ID
INNER JOIN PATRON_CONTACT_ADDRESS PCA
ON PC.PATRON_ACCOUNT_ID = PCA.PATRON_ACCOUNT_ID
INNER JOIN PATRON_CONTACT_EMAIL PCE
ON PCA.PATRON_ACCOUNT_ID = PCE.PATRON_ACCOUNT_ID
INNER JOIN PATRON_CONTACT_PHONE PCP
ON PCE.PATRON_ACCOUNT_ID = PCP.PATRON_ACCOUNT_ID
GROUP BY
EV.PUBLIC_DESCRIPTION
, EV.EVENT_DATE
, ES.PRICE
/* , BT.BUYER_TYPE_CODE */
, PCA.ADDR1
, PCA.ADDR2
, PCA.CITY
, PCA.POSTAL_CODE
, PCE.EMAIL
, PC.FORMATTED_NAME
, PCP.PHONE_NUMBER
, PCP.SECONDARY
/* check all these columns exist in the select clause
ORDER BY
ES.TRANSACTION_ID DESC
, PCP.SECONDARY DESC
, PCP.PHONE_NUMBER DESC
, PC.FORMATTED_NAME DESC
, PCE.EMAIL DESC
, PCA.POSTAL_CODE DESC
, PCA.CITY DESC
, PCA.ADDR2 DESC
, PCA.ADDR1 DESC
, BT.BUYER_TYPE_CODE DESC
, ES.PRICE DESC
*/
When you come the the final clause: ORDER BY you can ONLY reference columns that exist in the select clause. This example would FAIL
select column1 from table1 group by column1 order by fred
but this would work:
select column1 from table1 group by column1 order by column1
Your column aliases 'Paid Child', 'Free Child', 'Comps' should not be wrapped in single quotes. You should be using double quotes like you are already for "Adults".
So they should instead be:
"Paid Child"
"Free Child"
"Comps"
Or better yet, consider naming your aliases without any spaces, so you don't have to worry about wrapping the aliases in anything, like this:
paid_child
free_child
comps
Documentation on Database Object Names and Qualifiers:
Database Object Naming Rules
Every database object has a name. In a SQL statement, you represent the name of an object with a quoted identifier or a nonquoted identifier.
A quoted identifier begins and ends with double quotation marks ("). If you name a schema object using a quoted identifier, then you must use the double quotation marks whenever you refer to that object.
A nonquoted identifier is not surrounded by any punctuation.
...
Although column aliases, table aliases, usernames, and passwords are not objects or parts of objects, they must also follow these naming rules unless otherwise specified in the rules themselves.

How do I "Order by" more than on column?

This is my query:
SELECT count(oi.id) imgCnt, o.*,
IF(pricet=2,c.currency_value*o.attributes_36*o.price,
c.currency_value*o.price) AS pprice, od.title, oi.image,
MIN(oi.id), ( c.currency_value * o.price ) AS fprice,
ag.agent_name, DATE_FORMAT( o.date_added, '%d-%m-%Y') as dadded
FROM i_offers_12 o
LEFT JOIN i_agents ag ON o.agents_id = ag.id
LEFT JOIN i_currencies c ON o.currencies_id = c.id
LEFT JOIN i_offers_details od ON ( o.id = od.offers_id AND od.languages_id = 1 )
LEFT JOIN i_offers_images oi ON ( oi.offers_id = o.id AND oi.o_id = '12' )
WHERE ( o.offer_status='active' OR o.offer_status='sold')
AND actions_id = '1'
AND c.id = o.currencies_id
AND o.counties_id = '2'
AND o.cities_id = '3'
GROUP BY o.id
ORDER BY dadded
DESC
I want to sort after dadded(which is of type date) and offer_status(which is of type enum).
I want to display first, all of the elements which have offer_status = 'active' and sort by dadded and after that all of the elements which have offer_status = 'sold' and sort also by dadded. How can I do that? thx
The fields you want to sort on MUST be part of the select statement:
SELECT count(oi.id) imgCnt, o.*,
IF(pricet=2,c.currency_value*o.attributes_36*o.price,
c.currency_value*o.price) AS pprice, od.title, oi.image,
MIN(oi.id), ( c.currency_value * o.price ) AS fprice,
ag.agent_name, DATE_FORMAT( o.date_added, '%d-%m-%Y') as dadded ,
offer_status
FROM i_offers_12 o
LEFT JOIN i_agents ag ON o.agents_id = ag.id
LEFT JOIN i_currencies c ON o.currencies_id = c.id
LEFT JOIN i_offers_details od ON ( o.id = od.offers_id AND od.languages_id = 1 )
LEFT JOIN i_offers_images oi ON ( oi.offers_id = o.id AND oi.o_id = '12' )
WHERE ( o.offer_status='active' OR o.offer_status='sold')
AND actions_id = '1'
AND c.id = o.currencies_id
AND o.counties_id = '2'
AND o.cities_id = '3'
AND o.offer_status='active'
GROUP BY o.id
ORDER BY offer_status, dadded
DESC
Note that you normally should group by ALL non summarized fields (o.*,
IF(pricet=2,c.currency_value*o.attributes_36*o.price,
c.currency_value*o.price) AS pprice, od.title, oi.image, ( c.currency_value * o.price ) AS fprice,
ag.agent_name, DATE_FORMAT( o.date_added, '%d-%m-%Y') as dadded ,
offer_status)
MySQL is not enforcing it, but other DB like Oracle does.
You can use a comma, like
ORDER BY supplier_city DESC, supplier_state ASC;
I believe you can write the SQL with
ORDER BY offer_status, dadded

Dynamic query to dynamic data

I am new to the oracle database, I am trying to execute the following query
select o.id as ovaid ,
(case when(select count(m.cid) from ovamapper m where m.id = o.id and m.solutionid = 1)>0 then 1 else 0 end) as sol1,
(case when(select count(m.cid) from ovamapper m where m.id = o.id and m.solutionid = 2)>0 then 1 else 0 end) as sol1,
(case when(select count(m.cid) from ovamapper m where m.id = o.id and m.solutionid = 3)>0 then 1 else 0 end) as sol1 from ovatemplate o order by o.id
Instead of static values for solutionid , I would like to select it from other table.
Any help on this is really appreciated
you could use
join
to table that contain the solutionid. ex
Select * from ovatemplate JOIN solutiontable ON (solutiontable.ovaid=ovatempate.ovaid)
after that, change the static values to solutionid
Try this query
select o.id as ovaid ,
count(case when solutionid = 1 then m.cid else null end) as sol1 ,
count(case when solutionid = 2 then m.cid else null end) as sol2 ,
count(case when solutionid = 3 then m.cid else null end) as sol3
from ovamapper m , ovatemplate o
where m.id = o.id
group by o.id
order by o.id
If you dont need the aggregations as columns you should probably do that instead
select o.id as ovaid , solutionid , count(*) as sol
from ovamapper m , ovatemplate o
where m.id = o.id
and m.solutionid in (1,2,3)
group by o.id , solutionid
order by o.id

Resources