Oracle Update one to many

Hello all,

I have a table that holds all of my investigations for each individual. This table has a column called account_id. The account id is associated to one company. I noticed my previous worker update my case table with the wrong id. So I need to correct this. My investigation table as an invest_id and the account_id. My case table has multiple records for the account_id due to different users. I need to update the many records in my case table with the correct id from the investigation table. Please see sample data below for better detail.

invest table:
id       company     account_id
1        abc              12455

case table:
rpt_mth     user     case_id
2012-02     jane     12455
2012-02     john     12455
2012-02     mike    12455

I need the case table to read like this:
rpt_mth     user     case_id
2012-02     jane     1
2012-02     john     1
2012-02     mike    1

The case table has over 200k records over the last two years. So I need an update that will cross that period. I have tried the query below but am getting an error on updating many records from one.

update case_tbl
set case_id = invest_tbl.id
where case_tbl.case_id = invest_tbl.account_id;


Any help would be great.

Thanks,
Jeff
woodjeAsked:
Who is Participating?

[Webinar] Streamline your web hosting managementRegister Today

x
 
HainKurtConnect With a Mentor Sr. System AnalystCommented:
maybe this

UPDATE case_tbl c
   SET case_id =
          (SELECT id
             FROM invest i
            WHERE i.account_id = c.case_id)

Open in new window

0
 
HomerTNachoCheeseCommented:
update case_tbl
set case_id = 1
where case_tbl.case_id = 12455;

I think all you might have been doing was finding records with 12455 and re-entering 12455 again.  Or, you were ending up with no records to change.

All you need to do is look for 12455 and set it to 1 - no need to cross reference other tables' data I think.
0
 
woodjeAuthor Commented:
HomerTNachoCheesem

Thank you for the quick response. However the data i provided is only a small sampling. I am dealing with updating over 200,000 records for server rpt_mth's and several different accounts I can't do them one at a time. Can you suggest a way to modify my query to pull the current account id from the case_tbl and pull the id from the investigation table and update my case_id in my case_tbl with the id from the invest_tbl.
0
The new generation of project management tools

With monday.com’s project management tool, you can see what everyone on your team is working in a single glance. Its intuitive dashboards are customizable, so you can create systems that work for you.

 
slightwv (䄆 Netminder) Commented:
Can you use a simple PL/SQL block?

drop table myinvest purge;
create table myinvest(id number, account_id number);

insert into myinvest values(1,12455);
insert into myinvest values(2,11111);
commit;

drop table mycase purge;
create table mycase(myuser varchar2(10), case_id number);

insert into mycase values('jane',12455);
insert into mycase values('john',12455);
insert into mycase values('mike',12455);
insert into mycase values('bill',11111);
commit;

begin
	for i in (select id,account_id from myinvest) loop
		update mycase set case_id=i.id where case_id=i.account_id;
	end loop;
	commit;
end;
/

Open in new window

0
 
HomerTNachoCheeseCommented:
I don't have Oracle installed anywhere to figure it out and get the syntax correct.

What you could try is create a select query that pulls all your data together for starters, where all fields you mentioned in your original post are present.

I am only running Access right now, so if I wrote it for Access SQL, it would look like something this:

UPDATE case_tbl
INNER JOIN invest_tbl ON case_tbl.case_id = invest_tbl.account_id
SET case_tbl.case_id = invest_tbl.id

Potential problem is that you cannot update the field that is joined.

Probably not the best way of doing things, but if you created a copy of the case_id field and all of its data into a field called case_id_old, then you could join on that field instead if you ran into that problem.
0
 
woodjeAuthor Commented:
HainKurt,

Your idea looked great but I get the error More than one value was returned by a subquery.
0
 
slightwv (䄆 Netminder) Commented:
Looks like you might have duplicate account ids in the invest table.

select account_id from invest group by account_id  having count(account_id) > 1;
0
 
woodjeAuthor Commented:
slightwv,

Yes looks like I have about 16k of the 200k that do have dups for the account_id. If it is not one thing it is another.
0
 
slightwv (䄆 Netminder) Commented:
My pl/sql loop will run but it will grap the last account_id it comes to which probably isn't what you want...

Once you get everything cleaned up:  Constraints, constraints, constraints...  Then you'll not get back into this mess.
0
 
HainKurtSr. System AnalystCommented:
if you have multiple ids for case_id what is your solution?

invest table:
id       company     account_id
1        abc              12455
2        abc              12455

case table:
rpt_mth     user     case_id
2012-02     jane     12455
2012-02     john     12455
2012-02     mike    12455

???
0
 
HainKurtSr. System AnalystCommented:
you can try one of these:

random pickup:

UPDATE case_tbl c
   SET case_id =
          (SELECT id
             FROM ((  SELECT id, ROWNUM rn
                        FROM invest i
                       WHERE i.account_id = c.case_id))
            WHERE rn = 1)

min id pickup

UPDATE case_tbl c
   SET case_id =
          (SELECT MIN (id)
             FROM invest i
            WHERE i.account_id = c.case_id)


max id pickup

UPDATE case_tbl c
   SET case_id =
          (SELECT MAX(id)
             FROM invest i
            WHERE i.account_id = c.case_id)
0
 
woodjeAuthor Commented:
This worked with some modification to use max invest_id and filtering on dates. But your sample pointed me in the right direction to get the job done. Thanks.
0
All Courses

From novice to tech pro — start learning today.