How do I return the value of an identity column (id) in Oracle 12c after insertion? Seems like most of the approaches out there uses sequence to get back the id of the inserted item.
Simply use the RETURNING clause.
For example -
RETURNING identity_id INTO variable_id;
Test case -
SQL> set serveroutput on
SQL> CREATE TABLE t
2 (ID NUMBER GENERATED ALWAYS AS IDENTITY, text VARCHAR2(50)
3 );
Table created.
SQL>
SQL> DECLARE
2 var_id NUMBER;
3 BEGIN
4 INSERT INTO t
5 (text
6 ) VALUES
7 ('test'
8 ) RETURNING ID INTO var_id;
9 DBMS_OUTPUT.PUT_LINE('ID returned is = '||var_id);
10 END;
11 /
ID returned is = 1
PL/SQL procedure successfully completed.
SQL>
SQL> select * from t;
ID TEXT
---------- --------------------------------------------
1 test
SQL>
Related
I have a function that returns a table of custom objects. I wish to select a certain column by name from the returned result.
create or replace type sd_Serial_Number as object (
serial_number VARCHAR2(32)
);
The table of objects
create or replace type sd_Serial_Number_Table as table of sd_Serial_Number;
The function
create function get_result
return sd_Serial_Number_Table as
v_ret sd_Serial_Number_Table;
begin
select sd_Serial_Number(selected.SERIAL_NUMBER)
bulk collect into v_ret
from (
selection here
) selected;
return v_ret;
end get_result;
When I call the function this way, I get a result with a single column called SERIAL_NUMBER
select * from table(get_result());
However, I can't do something like this
select SERIAL_NUMBER from table(get_result());
Is there a way to select the column SERIAL_NUMBER ?
"I can't" is difficult to debug. I'll show you that I can (on the same database version you use).
SQL> SELECT * FROM v$version WHERE rownum = 1;
BANNER
--------------------------------------------------------------------------------
Oracle Database 11g Enterprise Edition Release 11.2.0.4.0 - 64bit Production
SQL> CREATE OR REPLACE TYPE sd_serial_number AS OBJECT
2 (
3 serial_number VARCHAR2 (32)
4 );
5 /
Type created.
SQL> CREATE OR REPLACE TYPE sd_serial_number_table AS TABLE OF sd_serial_number;
2 /
Type created.
SQL> CREATE OR REPLACE FUNCTION get_result
2 RETURN sd_serial_number_table
3 AS
4 v_ret sd_serial_number_table;
5 BEGIN
6 SELECT sd_serial_number (deptno)
7 BULK COLLECT INTO v_ret
8 FROM dept;
9
10 RETURN v_ret;
11 END get_result;
12 /
Function created.
Testing:
SQL> SELECT * FROM TABLE (get_result ());
SERIAL_NUMBER
--------------------------------
10
20
30
40
SQL> SELECT serial_number FROM TABLE (get_result ());
SERIAL_NUMBER
--------------------------------
10
20
30
40
SQL>
I have a table:
id
name
amount
If the same name was inserted, then amount should be incremented.
Else, insert with 0 amount.
How to create a trigger with a given condition?
You shouldn't really be doing that. AMOUNT column effectively counts number of NAME appearances in that table. You can always count it, can't you? So, what's the purpose of doing that?
If you want to know which name was inserted prior (or after) some other name (equal to the previous one), sort them by ID (if it is incremental, such as an identity column or if it gets its value from a sequence).
If ID isn't incremental, add DATE_INSERTED column (and sort by it; or apply ROW_NUMBER analytic function which orders values by DATE_INSERTED).
Also, what happens if you delete one of those duplicate names? Will you retroactively decrement AMOUNT column for all previous instances of that NAME?
But, if you insist, here's one option. As you can't just select from a table you're inserting into (because of the mutating table error), I'm using a compound trigger. The part of "insert amount 0" is done by setting the column's default value (you don't need any code for that).
SQL> create table test
2 (id number primary key,
3 name varchar2(20),
4 amount number default 0);
Table created.
SQL> create or replace trigger trg_ai_test
2 for insert on test
3 compound trigger
4
5 type test_rt is record (id test.id%type,
6 name test.name%type);
7 type rli_t is table of test_rt index by pls_integer;
8 g_rli rli_t;
9
10 after each row is
11 begin
12 g_rli(g_rli.count + 1).id := :new.id;
13 g_rli(g_rli.count).name := :new.name;
14 end after each row;
15
16 after statement is
17 l_cnt number;
18 begin
19 for i in 1 .. g_rli.count loop
20 dbms_output.put_Line('x');
21 dbms_output.put_line(i ||' '|| g_rli(i).id ||' '||g_rli(i).name);
22 select count(*) into l_cnt
23 from test
24 where name = g_rli(i).name;
25
26 update test set
27 amount = l_cnt
28 where id = g_rli(i).id;
29 end loop;
30
31 end after statement;
32 end trg_ai_test;
33 /
Trigger created.
SQL>
Testing:
SQL> insert into test (id, name) values (1, 'little');
1 row created.
SQL> insert into test (id, name) values (2, 'foot');
1 row created.
SQL> insert into test (id, name) values (3, 'little');
1 row created.
SQL> insert into test (id, name) values (9, 'little');
1 row created.
SQL> select * from test;
ID NAME AMOUNT
---------- -------------------- ----------
1 little 1
2 foot 1
3 little 2
9 little 3
SQL>
Or, a simpler solution (described above), without using that much code:
SQL> select id, name,
2 row_number() over (partition by name order by id) as amount
3 from test;
ID NAME AMOUNT
---------- -------------------- ----------
2 foot 1
1 little 1
3 little 2
9 little 3
SQL>
I'm trying to start a sequence that automatically populates a column when an insertion has been made. It should start from 500 and increment by 1. Any idea how would I go about doing this? So far I have something like this, but it seems to crash
CREATE TRIGGER ADD_TRIGGER ON tablename
AFTER INSERT
AS
BEGIN
ADD colname NUMBER IDENTITY(500,1);
END
GO
You can create a sequence
CREATE SEQUENCE mySeq
START WITH 500
INCREMENT BY 1
CACHE 100;
and then use it in your trigger
CREATE OR REPLACE TRIGGER myTrigger
AFTER INSERT ON table_name
FOR EACH ROW
BEGIN
SELECT mySeq.nextval
INTO :new.colname
FROM dual;
END;
Oracle 12c introduces IDENTITY COLUMNS.
SQL> CREATE TABLE new_identity_table
2 (
3 ID NUMBER GENERATED ALWAYS AS IDENTITY,
4 text VARCHAR2(50)
5 );
Table created.
SQL>
SQL> INSERT
2 INTO new_identity_table
3 (
4 text
5 )
6 VALUES
7 (
8 'This table has an identity column'
9 );
1 row created.
SQL> column text format A40;
SQL>
SQL> select * from new_identity_table;
ID TEXT
---------- ----------------------------------------
1 This table has an identity column
Oracle creates a sequence to populate the identity column. You can find it named as ISEQ$$
SQL> select sequence_name, min_value, max_value, increment_by from user_sequences;
SEQUENCE_NAME MIN_VALUE MAX_VALUE INCREMENT_BY
-------------------- ---------- ---------------------------- ------------
ISEQ$$_93199 1 9999999999999999999999999999 1
More more information about the identity columns, use the ALL_TAB_IDENTITY_COLS view.
SQL> SELECT table_name,
2 column_name,
3 generation_type,
4 identity_options
5 FROM all_tab_identity_cols
6 WHERE owner = 'LALIT'
7 ORDER BY 1, 2;
TABLE_NAME COLUMN_NAME GENERATION IDENTITY_OPTIONS
-------------------- --------------- ---------- --------------------------------------------------
NEW_IDENTITY_TABLE ID ALWAYS START WITH: 1, INCREMENT BY: 1, MAX_VALUE: 9999999
999999999999999999999, MIN_VALUE: 1, CYCLE_FLAG: N
, CACHE_SIZE: 20, ORDER_FLAG: N
For pre 12c releases, see Auto-increment primary key in Pre 12c releases (Identity functionality)
I am working on a small school project using oracle database. I have created some tables and two of them are Mobile (Mobile_Number,Status_Flag) Status_Flag shows if a number is active or not and there is another table Owner_Mobile(Owner_Id FK,Mobile_ID FK). Now I should write a Constraint that prohibits the insert operation if the corresponding Status_Flag is N for the specified number. I tried to make it using sub query but this is not possible.
the constrain should be applied to OWNER_MOBILE table of course. For example if I say: INSERT INTO OWNER_MOBILE(25541,042536) the constrain should check the Mobile table and see if the Mobile 042536 is active or not . If the number is not active the insert statement should generate a error
You can use the trigger or another PL/SQL API for this but you should take into account ACID transaction principles. Let's consider the case when flag value = 0 should prevent insertion:
SQL> create table mobile (mobile_id int primary key, flag int)
2 /
SQL> create table owner_mobile(owner_id int,
2 mobile_id int references mobile(mobile_id))
3 /
SQL> insert into mobile values (1,1)
2 /
SQL> commit
2 /
SQL> create or replace trigger
2 tr_owner_mobile
3 before insert on owner_mobile
4 for each row
5 declare
6 l_flag mobile.flag%type;
7 begin
8 select flag into l_flag
9 from mobile where mobile_id = :new.mobile_id;
10
11 if l_flag = 0 then
12 raise_application_error(-20000, 'Unavalable mobile');
13 end if;
14 end;
15 /
In the code above I simply select flag and rely on the retrieved value - I don't care of ACID.
In the first transaction I update flag value but don't commit:
SQL> update mobile set flag = 0 where mobile_id = 1;
In the second transaction I insert into owner_mobile and get the success:
SQL> insert into owner_mobile values(1,1);
1 row inserted.
Next, I commit the first transaction and later - the second one. What I get then:
SQL> select * from mobile;
MOBILE_ID FLAG
---------- ----------
1 0
SQL> select * from owner_mobile;
OWNER_ID MOBILE_ID
---------- ----------
1 1
Seems this is not what I expect.
I can use select for update to prevent inconsistent behavoiur:
SQL> update mobile set flag = 1;
1 row updated.
SQL> delete from owner_mobile;
1 row deleted.
SQL> commit;
SQL> create or replace trigger
2 tr_owner_mobile
3 before insert on owner_mobile
4 for each row
5 declare
6 l_flag mobile.flag%type;
7 begin
8 select flag into l_flag
9 from mobile where mobile_id = :new.mobile_id
10 for update;
11
12 if l_flag = 0 then
13 raise_application_error(-20000, 'Unavalable mobile');
14 end if;
15 end;
16 /
Now do the same:
SQL> update mobile set flag = 0 where mobile_id = 1;
1 row updated.
Second transaction is waiting because parent row is locked:
SQL> insert into owner_mobile values(1,1);
After commit in first transaction I get in the second one:
SQL> insert into owner_mobile values(1,1);
insert into owner_mobile values(1,1)
*
error in line 1:
ORA-20000: Unavalable mobile
ORA-06512: at "SCOTT.TR_OWNER_MOBILE", line 9
ORA-04088: error in trigger 'SCOTT.TR_OWNER_MOBILE'
So whatever you do to achieve requirements you will have to consider transaction isolation.
I have an oralcle SP forced on me that will not accept an empty parameter in an update. So if I wanted to set a value back to the default of ('') it will not let me pass in the empty string. Is there a keyword you can use such as default, null, etc that oracle would interpret back to the default specified for a particular column?
Sometimes things are just as simple as you hope they might be.
First, a table with a default value ...
SQL> create table t23 (
2 id number not null primary key
3 , col_d date default sysdate not null )
4 /
Table created.
SQL> insert into t23 values (1, trunc(sysdate, 'yyyy'))
2 /
1 row created.
SQL> select * from t23
2 /
ID COL_D
---------- ---------
1 01-JAN-10
SQL>
Next a procedure which updates the default column ...
SQL> create or replace procedure set_t23_date
2 ( p_id in t23.id%type
3 , p_date in t23.col_d%type )
4 is
5 begin
6 update t23
7 set col_d = p_date
8 where id = p_id;
9 end;
10 /
Procedure created.
SQL>
... but which doesn't work as we would like:
SQL> exec set_t23_date ( 1, null )
BEGIN set_t23_date ( 1, null ); END;
*
ERROR at line 1:
ORA-01407: cannot update ("APC"."T23"."COL_D") to NULL
ORA-06512: at "APC.SET_T23_DATE", line 6
ORA-06512: at line 1
SQL>
So, let's try adding a DEFAULT option ...
SQL> create or replace procedure set_t23_date
2 ( p_id in t23.id%type
3 , p_date in t23.col_d%type )
4 is
5 begin
6 if p_date is not null then
7 update t23
8 set col_d = p_date
9 where id = p_id;
10 else
11 update t23
12 set col_d = default
13 where id = p_id;
14 end if;
15 end;
16 /
Procedure created.
SQL>
... and lo!
SQL> exec set_t23_date ( 1, null )
PL/SQL procedure successfully completed.
SQL>
SQL> select * from t23
2 /
ID COL_D
---------- ---------
1 28-FEB-10
SQL>
I ran this example on an 11g database. I can't remember when Oracle introduced this exact support for DEFAULT, but it has been quite a while (9i???)
edit
The comments are really depressing. The entire point of building PL/SQL APIs is to make it easier for application developers to interact with the database. That includes being sensible enough to rewrite stored procedures when necessary. The big difference between building something out of software and, say, welding cast-iron girders together is that software is malleable and easy to change. Especially when the change doesn't alter the signature or behaviour of an existing procedure, which is the case here.
The procedure that's been forced on you:
create or replace procedure notEditable(varchar2 bar) as
begin
--update statement
null;
end;
How to use:
begin
notEditable(bar=>null);
end;
I didn't actually compile, but I believe this is the correct syntax.