I have an error I'm failing to understand. This is happening on some production code, but I created dummy tables to simulate the issue.
create table dl_test_nullable_src(col1 varchar2(20) NOT NULL, col2 varchar2(20) NOT NULL);
create table dl_test_nullable_tgt(col1 varchar2(20) NOT NULL, col2 varchar2(20) NOT NULL);
begin
for i in 1..10 loop
insert into dl_test_nullable_src(col1,col2) values(i*100,i*300);
insert into dl_test_nullable_tgt(col1,col2) values(i*100,i*300);
end loop;
commit;
end;
The below works
merge into dl_test_nullable_tgt tgt
using
(select * from dl_test_nullable_src) src
on (tgt.col1 = src.col1)
when matched then update
set tgt.col2 = src.col2
when not matched then insert
(tgt.col2)
values
(src.col2);
10 rows merged.
Now I add some parallel and append to direct path load any NEW records.
alter session enable parallel dml;
merge /*+ append parallel(tgt,4) */ into dl_test_nullable_tgt tgt
using
(select * from dl_test_nullable_src) src
on (tgt.col1 = src.col1)
when matched then update
set tgt.col2 = src.col2
when not matched then insert
(tgt.col2)
values
(src.col2);
SQL Error: ORA-01400: cannot insert NULL into ("PBTUSER5"."DL_TEST_NULLABLE_TGT"."COL1")
01400. 00000 - "cannot insert NULL into (%s)"
*Cause: An attempt was made to insert NULL into previously listed objects.
*Action: These objects cannot accept NULL values.
Even though this doesn't make sense on its own, it should not try to be inserting any records since everything that's in the SRC table is in the TGT table. The only way to bypass this is to also add col1 in the insert and values clause.
Table Contents
as far as I can see this is described in MOS document 1547251.1 - and classified as "not a bug, but a feature". The difference is based on the evaluation of contraints in parallel and non-parallel execution.
Related
I would receive an error:
ORA-02437: cannot validate (%s.%s) - primary key violated
Cause: attempted to validate a primary key with duplicate values or null values
I found it was because I have a stored procedure that increments the ID, but it had failed to do so when it re-ran and had an error related to one of my datatypes. I found I now had a duplicate ID in my database table. All this made sense and I was able to easily rectify it with a DELETE FROM MyTable WHERE ID = x, where x was the offending duplicate ID. The problem I have is the only way I was able to even find the IDs that were duplicated is in the first place is because I did a SELECT * FROM MyTable WHERE ID = x -- where x was one greater than the last ID I could actually see. I found it just by an educated guess. So:
Why can't I see these duplicate IDs when I open the table in Oracle SQL Developer? It only shows the last row as the ID before the duplicates. I don't think it is because of my primary key constraint, since the first line in my stored procedure is to remove that (and put it back, at the end - probably when I got my error), and it was not present when I looked at my table.
Is there some way to make these last IDs that got inserted into the table visible, so I wouldn't have to guess or assume that the duplicate IDs are "hiding" as one greater than the last ID I have in my table, in the future? There is a commit; in my stored procedure, so they should have appeared -- unless, of course, the procedure got hung up before it could run that line of code (highly probable).
Stored procedure that runs:
create or replace
PROCEDURE PRC_MYTABLE_INTAKE(
, EMPLOYEE_ID IN NVARCHAR2
, TITLE_POSITION IN NVARCHAR2
, CREATED_DATE IN DATE
, LAST_MODIFIED IN DATE
) AS
myid integer := 0;
appid integer := 0;
BEGIN
-- disable PK constraint so it can be updated
EXECUTE IMMEDIATE 'ALTER TABLE MYTABLE DROP CONSTRAINT MYTABLE_PK';
COMMIT;
-- assign ID to myid
SELECT ID INTO myid FROM MYTABLE WHERE ROWID IN (SELECT MAX(ROWID) FROM MYTABLE);
-- increment
myid := myid + 1;
-- assign APPLICATION_ID to appid
SELECT APPLICATION_ID INTO appid FROM MYTABLE WHERE ROWID IN (SELECT MAX(ROWID) FROM MYTABLE);
-- increment
appid := appid + 1;
-- use these ids to insert with
INSERT INTO MYTABLE (ID, APPLICATION_ID,
, EMPLOYEE_ID
, TITLE_POSITION
, CREATED_DATE
, LAST_MODIFIED
) VALUES(myid, appid,
, EMPLOYEE_ID
, TITLE_POSITION
, CREATED_DATE
, LAST_MODIFIED
);
COMMIT;
-- re-enable the PK constraint
EXECUTE IMMEDIATE 'ALTER TABLE PASS ADD CONSTRAINT MYTABLE_PK PRIMARY KEY (ID)';
COMMIT;
END;
Here's one problem:
SELECT ID
INTO myid
FROM MYTABLE
WHERE ROWID IN (SELECT MAX(ROWID) FROM MYTABLE)
There is no correlation between ID and ROWID, so you're not getting the maximum current ID, you're just getting the one that happens to be on the row that is furthest from the start of a datafile with a high number.
The code you need is:
SELECT COALESCE(MAX(ID),0)
FROM MYTABLE;
Or better yet, just use a sequence.
No idea why you're dropping the PK either.
Furthermore, when you issue the query:
SELECT APPLICATION_ID INTO appid ...
... that could be for a different row than the one you already got the id for, because a change could have been committed to the table.
Of course another issue is that you can't run two instances of this procedure at the same time either.
For David Aldridge, since he wants to look at code instead of the real reason I posted my question, run this ---
CREATE TABLE YOURSCHEMA.TESTING
(
TEST_ID NVARCHAR2(100) NOT NULL
, TEST_TYPE NVARCHAR2(100) NOT NULL
, CONSTRAINT TEST_PK PRIMARY KEY
(
TEST_ID
)
ENABLE
);
create or replace
PROCEDURE PRC_TESTING_INSERT(
TEST_TYPE IN NVARCHAR2
) AS
testid integer := 0;
BEGIN
-- disable PK constraint so it can be updated
EXECUTE IMMEDIATE 'ALTER TABLE TESTING DROP CONSTRAINT TEST_PK';
COMMIT;
-- assign TEST_ID to testid
SELECT TEST_ID INTO testid FROM TESTING WHERE ROWID IN (SELECT MAX(ROWID) FROM TESTING);
-- increment
testid := testid + 1;
-- use this id to insert with
INSERT INTO TESTING (TEST_ID, TEST_TYPE) VALUES(testid, TEST_TYPE);
COMMIT;
-- re-enable the PK constraint
EXECUTE IMMEDIATE 'ALTER TABLE TESTING ADD CONSTRAINT TEST_PK PRIMARY KEY (TEST_ID)';
COMMIT;
END;
SET serveroutput on;
DECLARE
test_type varchar(100);
BEGIN
test_type := 'dude';
YOURSCHEMA.PRC_TESTING_INSERT(test_type);
-- to verify the variable got set and procedure ran, could do:
--dbms_output.enable;
--dbms_output.put_line(test_type);
END;
Now, because there is no data in the table, the stored procedure will fail with ORA-06512: no data found. If you then try and run it again, you will get ORA-02443: cannot drop constraint - nonexistent constraint, because the EXECUTE IMMEDIATE 'ALTER TABLE TESTING DROP CONSTRAINT TEST_PK'; successfully dropped it, and the procedure never ran the command at the end to re-add it. This is what made me think I needed the commits, but even without them, it still will not complete the whole procedure.
To prove that the procedure DOES run, if given proper data, run this after creating the table, but before creating/running the stored procedure:
INSERT INTO TESTING (TEST_ID, TEST_TYPE)
VALUES ('1', 'hi');
And if you run the proc from a new table (not one with its constraint dropped), it will run fine.
Since mathguy didn't post this as the answer, though I'll credit him for the information...
Answer to why I can't see the duplicates is because the COMMIT does not occur in the procedure when it failed due to a datatype mismatch (which we found was actually in the application's code that sent the variable's values into this procedure, not in the stored procedure, itself). (It's also why I'll mark down anyone that says you don't have to add so many COMMIT lines in this procedure.) The commands were run in the session of the user that starts it - in my case, another session of the same DB user I was logged in with, but started from my application, instead of my SQL Developer session. It also explains why I could do a COMMIT, myself, but it did not affect the application's session - I could not commit any actions ran from another session. Had I ran a COMMIT as an OracleCommand and did an .ExecuteNonQuery on my OracleConnection right after the failure within the catch of my application, I would have seen the rows in SQL Developer without having to do a special query.
So, in short, the only way to see the items was with a direct query using WHERE ID =, find the last ID and increment it, and put it in the query.
Procedure is to check the the eid ,and do merge. While updating the existing row, it needs to update with the eid.seq.nextval.
I have created a the sequence and calling in Procedure.
CREATE OR REPLACE PROCEDURE Temp(
Eid Number,
dpt varchar2
) As
BEGIN
MERGE INTO Src1 e
USING (select v_eid as eid
, v_dept as dept
FROM dual) d
ON (e.eid = d.eid)
WHEN NOT MATCHED THEN
INSERT (e.eid,e.dept)
VALUES(d.eid, d.dept)
WHEN MATCHED THEN
UPDATE SET e.eid = eid_SEQ.nextval, e.dept = d.dept;
END;
/
Error:
1.ORA--38104:ORA-38104: Columns referenced in the ON Clause cannot be updated.
IF I remove the ON clause condition then PK cannot be null error .
Also, the best procedure to call the seq.nextval in the procedure.
Any help is greatly appreciated!
What you want to do is impossible with a merge statement. Think about it: you're trying to update the value of a field that is used to determine if the field should be updated. This is not safe, especially if the merge statement updates/inserts more than one row.
Apart from that it is close to impossible to suggest an alternative since eid seems to be a primary key and updating primary keys is normally a bad thing.
This question already has answers here:
How to create id with AUTO_INCREMENT on Oracle?
(18 answers)
Closed 8 years ago.
I am facing issue while inserting multiple row in one go into table because column id has primary key and its created based on sequence.
for ex:
create table test (
iD number primary key,
name varchar2(10)
);
insert into test values (123, 'xxx');
insert into test values (124, 'yyy');
insert into test values (125, 'xxx');
insert into test values (126, 'xxx');
The following statement creates a constraint violoation error:
insert into test
(
select (SELECT MAX (id) + 1 FROM test) as id,
name from test
where name='xxx'
);
This query should insert 3 rows in table test (having name=xxx).
You're saying that your query inserts rows with primary key ID based on a sequence. Yet, in your insert/select there is select (SELECT MAX (id) + 1 FROM test) as id, which clearly is not based on sequence. It may be the case that you are not using the term "sequence" in the usual, Oracle way.
Anyway, there are two options for you ...
Create a sequence, e.g. seq_test_id with the starting value of select max(id) from test and use it (i.e. seq_test_id.nextval) in your query instead of the select max(id)+1 from test.
Fix the actual subselect to nvl((select max(id) from test),0)+rownum instead of (select max(id)+1 from test).
Please note, however, that the option 2 (as well as your original solution) will cause you huge troubles whenever your code runs in multiple concurrent database sessions. So, option 1 is strongly recommended.
Use
insert into test (
select (SELECT MAX (id) FROM test) + rownum as id,
name from test
where name='xxx'
);
as a workaround.
Of course, you should be using sequences for integer-primary keys.
If you want to insert an ID/Primary Key value generated by a sequence you should use the sequence instead of selecting the max(ID)+1.
Usually this is done using a trigger on your table wich is executed for each row. See sample below:
CREATE TABLE "MY_TABLE"
(
"MY_ID" NUMBER(10,0) CONSTRAINT PK_MY_TABLE PRIMARY KEY ,
"MY_COLUMN" VARCHAR2(100)
);
/
CREATE SEQUENCE "S_MY_TABLE"
MINVALUE 1 MAXVALUE 999999999999999999999999999
INCREMENT BY 1 START WITH 10 NOCACHE ORDER NOCYCLE NOPARTITION ;
/
CREATE OR REPLACE TRIGGER "T_MY_TABLE"
BEFORE INSERT
ON
MY_TABLE
REFERENCING OLD AS OLDEST NEW AS NEWEST
FOR EACH ROW
WHEN (NEWEST.MY_ID IS NULL)
DECLARE
IDNOW NUMBER;
BEGIN
SELECT S_MY_TABLE.NEXTVAL INTO IDNOW FROM DUAL;
:NEWEST.MY_ID := IDNOW;
END;
/
ALTER TRIGGER "T_MY_TABLE" ENABLE;
/
insert into MY_TABLE (MY_COLUMN) values ('DATA1');
insert into MY_TABLE (MY_COLUMN) values ('DATA2');
insert into MY_TABLE (MY_ID, MY_COLUMN) values (S_MY_TABLE.NEXTVAL, 'DATA3');
insert into MY_TABLE (MY_ID, MY_COLUMN) values (S_MY_TABLE.NEXTVAL, 'DATA4');
insert into MY_TABLE (MY_COLUMN) values ('DATA5');
/
select * from MY_TABLE;
In Oracle, I have a requirement where in I need to insert records from Source to Target and then update the PROCESSED_DATE field of source once the target has been updated.
1 way is to use cursors and loop row by row to achieve the same.
Is there any other way to do the same in an efficient way?
No need for a cursor. Assuming you want to transfer those rows that have not yet been transfered (identified by a NULL value in processed_date).
insert into target_table (col1, col2, col3)
select col1, col2, col3
from source_table
where processed_date is null;
update source_table
set processed_date = current_timestamp
where processed_date is null;
commit;
To avoid updating rows that were inserted during the runtime of the INSERT or between the INSERT and the update, start the transaction in serializable mode.
Before you run the INSERT, start the transaction using the following statement:
set transaction isolation level SERIALIZABLE;
For more details see the manual:
http://docs.oracle.com/cd/E11882_01/server.112/e26088/statements_10005.htm#i2067247
http://docs.oracle.com/cd/E11882_01/server.112/e25789/consist.htm#BABCJIDI
A trigger should work. The target table can have a trigger that on update, updates the source table's column with the processed date.
My preferred solution in this sort of instance is to use a PL/SQL array along with batch DML, e.g.:
DECLARE
CURSOR c IS SELECT * FROM tSource;
TYPE tarrt IS TABLE OF c%ROWTYPE INDEX BY BINARY_INTEGER;
tarr tarrt;
BEGIN
OPEN c;
FETCH c BULK COLLECT INTO tarr;
CLOSE c;
FORALL i IN 1..tarr.COUNT
INSERT INTO tTarget VALUES tarr(i);
FORALL i IN 1..tarr.COUNT
UPDATE tSource SET processed_date = SYSDATE
WHERE tSource.id = tarr(i).id;
END;
The above code is an example only and makes some assumptions about the structure of your tables.
It first queries the source table, and will only insert and update those records - which means you don't need to worry about other sessions concurrently inserting more records into the source table while this is running.
It can also be easily changed to process the rows in batches (using the fetch LIMIT clause and a loop) rather than all-at-once like I have here.
Got another answer from some one else. Thought that solution seems much more reasonable than enabling isolation level as all my new records will have the PROCESSED_DATE as null (30 rows which inserted with in the time the records got inserted in Target table)
Also the PROCESSED_DATE = NULL rows can be updated only by using my job. No other user can update these records at any point of time.
declare
date_stamp date;
begin
select sysdate
into date_stamp
from dual;
update source set processed_date = date_stamp
where procedded_date is null;
Insert into target
select * from source
where processed_date = date_stamp;
commit;
end;
/
Let me know any further thoughts on this. Thanks a lot for all your help on this.
Do want to create Stored procc which updates or inserts into table based on the condition if current line does not exist in table?
This is what I have come up with so far:
PROCEDURE SP_UPDATE_EMPLOYEE
(
SSN VARCHAR2,
NAME VARCHAR2
)
AS
BEGIN
IF EXISTS(SELECT * FROM tblEMPLOYEE a where a.ssn = SSN)
--what ? just carry on to else
ELSE
INSERT INTO pb_mifid (ssn, NAME)
VALUES (SSN, NAME);
END;
Is this the way to achieve this?
This is quite a common pattern. Depending on what version of Oracle you are running, you could use the merge statement (I am not sure what version it appeared in).
create table test_merge (id integer, c2 varchar2(255));
create unique index test_merge_idx1 on test_merge(id);
merge into test_merge t
using (select 1 id, 'foobar' c2 from dual) s
on (t.id = s.id)
when matched then update set c2 = s.c2
when not matched then insert (id, c2)
values (s.id, s.c2);
Merge is intended to merge data from a source table, but you can fake it for individual rows by selecting the data from dual.
If you cannot use merge, then optimize for the most common case. Will the proc usually not find a record and need to insert it, or will it usually need to update an existing record?
If inserting will be most common, code such as the following is probably best:
begin
insert into t (columns)
values ()
exception
when dup_val_on_index then
update t set cols = values
end;
If update is the most common, then turn the procedure around:
begin
update t set cols = values;
if sql%rowcount = 0 then
-- nothing was updated, so the record doesn't exist, insert it.
insert into t (columns)
values ();
end if;
end;
You should not issue a select to check for the row and make the decision based on the result - that means you will always need to run two SQL statements, when you can get away with one most of the time (or always if you use merge). The less SQL statements you use, the better your code will perform.
BEGIN
INSERT INTO pb_mifid (ssn, NAME)
select SSN, NAME from dual
where not exists(SELECT * FROM tblEMPLOYEE a where a.ssn = SSN);
END;
UPDATE:
Attention, you should name your parameter p_ssn(distinguish to the column SSN ), and the query become:
INSERT INTO pb_mifid (ssn, NAME)
select P_SSN, NAME from dual
where not exists(SELECT * FROM tblEMPLOYEE a where a.ssn = P_SSN);
because this allways exists:
SELECT * FROM tblEMPLOYEE a where a.ssn = SSN