I have a customer table.
I have created stored procedure I can use to insert new data into the table. But what if I wanted to use the same procedure to update OR delete data from that table. Could I do this easily or do I have to use a separate function/procedure for each function?
create or replace procedure add_customer(custid in table.id%type,
name table.name%type)
is
begin
insert into table(id, name)
values(id, name);
commit;
end;
/
You can add parameter like action in example below and use it in code:
create or replace procedure modify_customer(
action in varchar2, custid in table.id%type, custname table.name%type)
is
begin
if action = 'insert' then
insert into table(id, name) values(custid, name);
commit;
elsif action = 'delete' then
delete from table where id = custid and name = custname;
commit;
end if;
end;
You can add a discriminator parameter to your add_customer procedure which says whether the action is INSERT, UPDATE or DELETE. Based on this parameter you can create the required insert, update or delete statement. This way you will be able to use a common procedure for all the actions.
As far as using one procedure or multiple procedures, if the table is a simple one with limited number of columns, one procedure should do fine. But once the number of columns increases in the table, one procedure might become more complicated than required.
Related
I'm trying to set an "after insert" trigger that executes a procedure. The procedure would take all inserted rows in table A, group them by a column and insert the result in a table B. I know about "new" variable but it gets inserted rows one by one. Is it possible to get all of them?
I think I can't use a for each row statement as I need to group rows depending on the "trackCode" variable, shared by different rows in tableA.
CREATE OR REPLACE PROCEDURE Public.my_procedure(**inserted rows in tableA?**)
LANGUAGE 'plpgsql'
AS $$
BEGIN
INSERT INTO Public."tableB" ("TrackCode", "count")
SELECT "TrackCode", count(*) as "count" FROM Public."tableA" --new inserted rows in this table
GROUP BY "vmsint"."TrackCode" ;
COMMIT;
END;
$$;
create trigger Public.my_trigger
after insert ON Public.tableA
execute procedure Public.my_procedure(**inserted rows in tableA?**)
Thank you!
You create a statement lever trigger, but do not attempt to pass parameters. Instead use the clause referencing new table as reference_table_name. In the trigger function you use the reference_table_name in place of the actual table name. Something like: (see demo)
create or replace function group_a_ais()
returns trigger
language 'plpgsql'
as $$
begin
insert into table_b(track_code, items)
select track_code, count(*)
from rows_inserted_to_a
group by track_code ;
return null;
end;
$$;
create trigger table_a_ais
after insert on table_a
referencing new table as rows_inserted_to_a
for each statement
execute function group_a_ais();
Do not attempt to commit in a trigger, it is a very bad id even if allowed. Suppose the insert to the main table is part of a larger transaction, which fails later in its process.
Be sure to refer to links provided by Adrian.
I trying to write a trigger that will update a column when user insert or updates a row, within the same table.
Example:
insert into user(ID, F_NM, L_NM, EMAIL) values ('1', 'John','Doe','john.doe#market.org.com');
after the insert, i want to call: update user set ORG = 'market' where ID = '1'.
create or replace trigger user_change
after insert or update of EMAIL on USER
for each row
declare
NEW_ORG VARCHAR(10);
BEGIN
CASE
when :NEW.EMAIL like '$#market.org.com' then
NEW_ORG := 'market';
........
END CASE;
UPDATE USER set ORG = NEW_ORG where ID = :NEW.ID
END;
Calculating the new ORG work, but I can't get the update statement to work.
I get 'ORA-04091 table USER is mutating, trigger/funtion may not see it', figure its due to me inserting/updating the same record at same time. Tried adding 'pragma autonomous_transaction' and 'commit' to the trigger, the insert/update of fields works but the ORG does not get updated.
Also tried changing to INSTEAD OF INSERT OR UPDATE OF EMAIL but I keep getting 'ORA-04073 column list not valid for this trigger type'
create or replace trigger user_change
instead of insert or update of EMAIL on USER
while i get 'ORA-25002 cannot create instead of triggers on tables'
create or replace trigger user_change
instead of insert on USER
Why not simply turn the trigger to a before trigger, when you can set the value before it is written? This way, you don't need to run a new DML statement on the table, which avoid the "mutating" error.
create or replace trigger user_change
after insert or update of email on user
for each row
begin
if :new.email like '%#market.org.com' then
:new.org := 'market';
end if;
end;
Looks like your org column can be calculated virtual column. In this case it would be better to create user-defined deterministic pl/sql function that returns correct calculated value and add it to your table, for example:
Alter table t add org varchar2(30) generated always as (f_get_org(email))
Here I have a question that I need to write a PL/SQL. The structure of the database is also linked. The question requires to use a sequence inside a procedure. I'm new to this and don't know if this works properly and my exec command doesn't seem working please help me out. Also is this how to look up the max shareholder_id that the sequence should start with, or can I a select inside create sequence?
Write a PL/SQL procedure called INSERT_DIRECT_HOLDER which will be used to insert new direct holders. Create a sequence to automatically generate shareholder_ids. Use this sequence in your procedure.
-Input parameters: first_name, last_name
DROP SEQUENCE shareholder_id_seq;
SELECT
MAX(shareholder_id)
FROM shareholder;
CREATE SEQUENCE shareholder_id_seq
INCREMENT BY 1
START WITH 25;
CREATE OR REPLACE PROCEDURE insert_direct_holder(
p_first_name in direct_holder.first_name%type,
p_last_name in direct_holder.last_name%type)
IS
v_shareholder_id NUMBER(6);
BEGIN
INSERT INTO DIRECT_HOLDER(direct_holder_id,first_name,last_name) values(shareholder_id_seq.nextval, p_first_name, p_last_name);
INSERT INTO shareholder (shareholder_id, type) VALUES (shareholder_id_seq.nextval,'Direct_Holder');
COMMIT;
END;
/
/* test command*/
exec insert_direct_holder( p_first_name, p_last_name );
You need to insert the record with same ID in both the tables.
also, you need to insert record into your parent table (shareholder) first and then child table(direct_holder).
CREATE OR REPLACE PROCEDURE insert_direct_holder(
p_first_name in direct_holder.first_name%type,
p_last_name in direct_holder.last_name%type)
IS
v_shareholder_id NUMBER(6);
BEGIN
v_shareholder_id := shareholder_id_seq.nextval;
INSERT INTO shareholder (shareholder_id, type) VALUES (v_shareholder_id,'Direct_Holder');
INSERT INTO DIRECT_HOLDER(direct_holder_id,first_name,last_name) values(v_shareholder_id, p_first_name, p_last_name);
COMMIT;
END;
/
/* test command*/
exec insert_direct_holder( p_first_name, p_last_name );
Sequence is created once to implement auto incrementing feature for any numeric column.
For the current use case it has to be created just once and left forever hopefully.Sequence can be modified in future if required.
If shareholder table has records in it already, then create the sequence with start value as SELECT MAX(shareholder_id) + 1 FROM shareholder; to avoid primary key constraint violation.
A slight modification is required for the stored procedure to use the same SHAREHOLDER.SHAREHOLDER_ID as the column has foreign key relationship with DIRECT_HOLDER.
Use INSERT ALL to insert to both the tables for the same sequence.nextval.
CREATE OR REPLACE PROCEDURE insert_direct_holder(
p_first_name in direct_holder.first_name%type,
p_last_name in direct_holder.last_name%type)
IS
BEGIN
INSERT ALL
INTO SHAREHOLDER
(shareholder_id, type) values(shareholder_id_seq.nextval,'Direct_Holder')
INTO DIRECT_HOLDER
(direct_holder_id,first_name,last_name) values
(shareholder_id_seq.nextval,p_first_name,p_last_name)
SELECT 'DUMMY' FROM dual;
COMMIT;
END;
/
dbfiddle demo with working code : https://dbfiddle.uk/?rdbms=oracle_18&fiddle=5d80488fb69d78d4b5087f06a5becf96
I have a following question, for example I have a following table:
CREATE TABLE "regions" (gid serial PRIMARY KEY,
"__gid" int8,
"name" varchar(20),
"language" varchar(7),
"population" int8);
And I want to insert some records, say one of the values for "name" is - 'B', what sort of code would I have to write to change 'B' to 'English-Speaking'? Is that done with some sort of trigger? So would I have to write a trigger to change the values automatically on insert? Any help greatly appriciated!!!
It's an UPDATE statement which will do what you wish, in this case:
UPDATE regions set name = 'English-Speaking' where name = 'B';
To put this in a function use something like:
CREATE OR REPLACE FUNCTION insert_into_wgs()
RETURNS void AS
$$
BEGIN
UPDATE regions SET name = 'English-Speaking' WHERE name = 'B';
END
$$
LANGUAGE 'pgpsql';
Then you create a trigger to run this function:
CREATE TRIGGER log_update
AFTER UPDATE ON accounts
FOR EACH ROW
WHEN (OLD.* IS DISTINCT FROM NEW.*)
EXECUTE PROCEDURE
insert_into_wgs();
Assuming I've guessed what you mean correctly from your description:
You will need a simple BEFORE INSERT OR UPDATE ... FOR EACH ROW EXECUTE trigger to invoke a PL/PgSQL trigger procedure that changes the value of the NEW record and then does a RETURN NEW.
The documentation contains abundant details, and since this is homework I'm not going to provide a complete example. Start with CREATE TRIGGER and PL/pgSQL trigger procedures.
I want to write a trigger for a table "TRANSACTION".When a new line is inserted, I want to trigger to update the field "TRANSACTIONID" to the maximum + 1 of all the previous records.
I on't know much about SQL. Can someone help me?
many thanks
This is a really bad idea for a multi-user environment, as it will serialise inserts into the table. The usual approach is to use an Oracle sequence:
create sequence transaction_seq;
create trigger transaction_bir before insert on transaction
for each row
begin
:new.id := transaction_seq.nextval;
end;
To write a trigger based solution that actually got the max current value plus 1, you would need to write a complex 3-trigger solution to avoid the "mutating table" issue. Or you could create a simpler solution using another table to hold the current maximum value like this:
create table transaction_max (current_max_id number);
insert into transaction_max values (0);
create trigger transaction_bir before insert on transaction
for each row
declare
l_current_max_id number;
begin
update transaction_max set current_max_id = current_max_id + 1
returning current_max_id into l_current_max_id;
:new.id := l_current_max_id;
end;
This will avoid the mutating table issue and will serialize (slow down) inserts, so I don't see any advantage of this over using a sequence.
CREATE TRIGGER trigger1 on TransactionTable
INSTEAD OF INSERT
AS
BEGIN
DECLARE #MaxTranId INT
SELECT
#MaxTranId = MAX(TransactionId)
FROM
TransactionTable
INSERT INTO TransactionTable
SELECT
#MaxTranId + 1 ,
RestOfYourInsertedColumnsHere ,
FROM
inserted
END
GO