SQL & PL/SQL :: Using Subquery In Insert Statement
Jun 5, 2010
I read in a book that you can't use subquery in an insert statement . E.g:
1)insert into dates (date_col) values (select sysdate fom dual) but when i tried using subquery like this:
2)insert into regions values ((select max(region_id)+1 from regions), 'Oce');
This query worked but 1st query didnt.From my assumptions if we try inserting values in table with the subqueries for a particular column as in 1st query , it will throw error but not while inserting values in all columns as in 2nd query.
Step# 2 manipulated the tables: Inserted 3 records in Table_C.
Then I created an Insert Trigger to Table_A with an insert statement into Table_B and a subquery to Table_C. Please see below:
CREATE OR REPLACE TRIGGER TABLE_A_TR after INSERT OR UPDATE OR DELETE ON TABLE_A FOR EACH ROW DECLARE [code]......
Step# 3 compiled the created trigger and I've successfully compiled it. Step# 4 Tested the trigger (TABLE_A_TR) using an insert statement to TABLE_A. Insert into TABLE_A values (1,'testa','testb')
I've successfully insert the values into TABLE_A however I've observed that the trigger didn't execute the insert statement because TABLE_B has an empty rows. I tried to manually execute the insert statement just to see if there's an issue in my insert statement but I've successfully populated the values into TABLE_B. So I'm wondering why the trigger didn't execute the insert statement.
select * from (select pa_request_id ,max(status) status ,max(approved_amount) approved_amount ,min(level_id) level_id ,max(req_amount) req_amount from target_aggregation_attendee group by pa_request_id) where status = 'Approved')
My Main Query:
select TARGET_EMPLOYEE.FIRST_NAME||' ' ||TARGET_EMPLOYEE.LAST_NAME as Requestername, TARGET_EMPLOYEE.GE_ID as requesterGEID, TARGET_ATTENDEE.FIRST_NAME||' ' ||TARGET_ATTENDEE.LAST_NAME as AttendeeName, TARGET_ATTENDEE.ATTENDEE_TYPE_FLAG as Attendeetyflg, TARGET_ATTENDEE.US_GO_ATTENDEE_FLAG as usgoflg, TARGET_ATTENDEE.COMP_GOVT_AGENCY_DEPT as Atcomp,
I have 3 tables, user_login_event, person and resource_viewed_event. What I want to do have a report for each month, users logged in our application and then show for each month, how many records were created in table person and how many resource views events were logged in resource_viewed_event.
Lets only worry about the timestamp fields in these tables now as I want to use them to join the tables together or at least build correlated subqueries along the months. I have tried several options, all not leading to a desired result:
Left outer join. Works but its incredibly slow:
SELECT distinct to_char(ule.TIMESTAMP,'YYYY-MM') as "YYYY-MM", count(distinct ule.id) as "User Logins", count(distinct ule.user_id) as "Users logged on", count(distinct p2.id) as "Existing Users", count(distinct p1.id) as "New Users", count(distinct r1.id) as "Resources created"
[code]....
Tried the same with left outer joins of temporary tables created through select statements:
select distinct ule.month as "Month", count(distinct p1.user_id) as "Users created", count (ule.id) as "Logins", count (distinct ule.user_id) as "Users logged in", count(rv.id) as "Resource Views", count(distinct rv.resource_id) as "Resources Viewed"
[code]....
Tried the same with left outer joins of temporary tables created through select statements:
select distinct ule.month as "Month", count(distinct p1.user_id) as "Users created", count (ule.id) as "Logins", count (distinct ule.user_id) as "Users logged in", count(rv.id) as "Resource Views", count(distinct rv.resource_id) as "Resources Viewed"
[code]....
another approach is to create my own temporary tables using select statements and create fixed Month values which I can use to directly link the sets together.
select distinct ule.loginday as "Month", count(distinct ule.id) as "Logins", count(distinct ule.user_id) as "Users logged in", count(distinct p1.user_id) as "Users created", count(distinct p2.user_id) as "Existing users1"
[code]....
performance is OK with 2 tables but the example above takes forever to execute.
Tried an approach with union but this creates new rows for each table
SELECT DISTINCT p1.MONTH AS "Month", COUNT(DISTINCT p1.user_id) AS "Users created", NULL AS "Logins", NULL AS "Users Logged in", NULL AS "Resource views", NULL AS "Resources viewed" FROM (SELECT To_char(person.created_on_date, 'YYYY-MM') AS MONTH,
how to insert data in oracle table without writing insert statement in oracle 9i or above. i am not going to write insert all, merge, sqlloder and import data.
I have one query i have create one table T1 in this table i have this three colm (EMPNO,ENAME,HIREDATE) and i have to insert raw from this two table T2 and T3 in T2 table colms are (EMPNO,SAL) and T3 table colm are (EMPNO,MGRID) so which query i have to run
I wanted to print 'null' when the column value is null. Actually, i am doing something like this
select empno||','||''''||ename||'''''||','||comm||','||sal from emp
It gives the following output for example 7369,'pointers',,200 If I use the above values to form a insert statement it throws an error. As 'comm' value is not there.
I wish to get something like 7369,'pointers','',200 or 7369,'pointers',null,200 from the above select query
note I dint copy paste the query exactly from my sql*plus session as I am away from my oracle machine
i'm executing a load test using the below statement:
INSERT INTO BPMBI.PPROCINSTANCE SELECT * FROM BPMBI.PPROCINSTANCE_BKP WHERE ROWNUM < 501; COMMIT; EXIT;
I've been using the same for a few days without any error while now wehn i try to use the same i get below error:
SQL> insert into pprocinstance select * from pprocinstance_bkp where rownum <= 500; insert into pprocinstance select * from pprocinstance_bkp where rownum <= 500 * ERROR at line 1: ORA-01858: a non-numeric character was found where a numeric was expected.
I've been trying to check what the issue is and when i execute the rownum < 140 works fine while rownum < 150 is showing the same error.
I have a table with unique index, and the application is doing inserts/updates into this table.Suddenly and for about half a minute I faced a high concurrency waits on all the processes running these inserts for one node. I saw this high concurrency wait in the top activity screen of the OEM only on one of the nodes. knowing that the processes doing these inserts are running on both nodes.
All what I have that in this half minute I see high concurrency wait in OEM top activity screen related to this insert statement and when I clicked on the insert I found high "enq: TX - index contention". Again this was only on one node.After this half minute everything went back to normal.What could be the reason and how can I investigate it ?
Given below a block of code, this code compiled successfully on one DB but returns error on other.DB version is same. I know sqlerrm can not be used directly but how it compiled successfully on one DB.
I am creating the following two tables...no issues here: CODECREATE TABLE COURSE_SECTION ( Csecid NUMBER(8) CONSTRAINT COURSE_SELECTION_NUMBER_pk PRIMARY Key, Cid NUMBER(6) NOT NULL CONSTRAINTS COURSE_SELECTION_Cid_fk REFERENCES COURSE, Termid NUMBER(5) NOT NULL CONSTRAINTS COURSE_SELECTION_Termid_fk REFERENCES TERM,
[code]...
The issue I am having is actually inserting data into the table: CODEINSERT INTO ENROLLMENT VALUES (100, 1000, 'A' ); INSERT INTO ENROLLMENT VALUES (100, 1003, 'A' );
[code]...
But I get an ORACLE error of
ORA-02291- integrity constraint (User1.ENROLLMENT_CSECID_FK) violated - parent key not foundHow can the parent key not be found when I have it declared/created in the above statement?
I would like to insert a value if that value is not existing in the table (example for a column which contains date only new dates should be inserted and if the date already exists in the column then it needs to get updated )
example of scenario...
date s1 s2 s3
in the above if the date is new..it should get inserted with the appropriate slot no.(s1,s2,s3) if the date already exists it needs to update the no.in slot no.
I want to grant a privilege through an insert statement into a sys table.Why do not grant the privilege through the classic way : grant select on t to l_user; ?
Because I want to do it remotely.I am connected to db1.I want to grant select on t2 to u2_b from u2_a.I assume that all DDL are DML. So a grant is equivalent "somewhere" to an insert.I tried to do my requirement locally, and here is the output.
SQL> conn scott/aa Connecté. SQL> -- step 1 : try to grant "normally" a select on dept to hr from scott SQL> grant select on dept to hr;
Autorisation de privilèges (GRANT) acceptée.
SQL> SQL> conn sys/a as sysdba Connecté. SQL> -- step 2 : Then, we connect to sys to see the row inserted in dba_tab_privs SQL> SQL> col GRANTEE format A10 SQL> col OWNER format A10
[code]...
Then if I can do it locally, I can do it remotely through a db link.
If we insert a row in a database table then the new row stays at database buffer cache in SGA (until commit), right?. The target table is not affected (before commit). The new row is saved after commit.
I saw a concepts at Sybex oracle 10g oca book (Page 406) as follows:
" INSERT statements use little space in an undo segment; only the pointer to the new row is stored in the undo tablespace. To undo an INSERT statement, the pointer locates the new row and deletes it from the table if the transaction is rolled back. "
My question is If the row is not saved at table before commit, if we issue rollback then how oracle delete from table? I think the new row is deleted from database buffer cache in SGA.
DECLARE v_seq_num NUMBER; BEGIN SELECT SEQ_ID.NEXTVAL INTO v_seq_num FROM DUAL; INSERT INTO TABLEA (COL1, COL2) VALUES (v_seq_num, 'test'); INSERT INTO TABLEB (COL3) VALUES (v_seq_num); END;
-- Option2 - Using sequence.NEXTVAL in INSERT USING RETURNING INTO clause
DECLARE v_seq_num NUMBER; BEGIN INSERT INTO TABLEA (COL1, COL2) VALUES (SEQ_ID.NEXTVAL, 'test') RETURNING COL1 INTO v_seq_num; INSERT INTO TABLEB (COL3) VALUES (v_seq_num); END;
i'm working in an Oracle 10g database on an IBM AIX server.
I have 3 tables (tables A, B and C).
Table A has columns -- product, rate and expiration date.
Table B has columns -- product, rate and deductible.
Table C has columns -- product, rider, gender, age and rate.
I also have a Master table which is used to store the data from Tables A, B and C via the insert statement.
I'm trying to create a dynamic SQL insert statement using a shell script to insert data from the columns in Tables A, B and C into my Master table. Master table does contains all columns from Tables A, B and C, although a column name could be spelled differently. For example, Master table contains a column named "deduct", while Table B has the same column spelled as "deductible".
I build the dynamic query using a for loop in my shell script (see below).
The problem is that i can't get the correct columns in the Master table in the dynamic SQL for the insert because depending on the table i'm selection from, the columns are different. So how do i get the correct columns in the SQL for the Master table?
Example Shell Script
--Archive_Rates.txt contains: Table A, Table B, Table C (but the next time my process runs, Archive_Rates might contain Table D, Table E and Table F -- each which have different column...but all columns are still in the Master table)
for tbl in `more Archive_Rates.txt` do echo 'BEGIN WORK; ' > rc1.sql echo ' ' >> rc1.sql echo 'insert into Master' >> rc1.sql echo '(prod, rate, rate_exp) ' >> rc1.sql
In a trigger(on update of a table t1) I am trying to write, I am doing an insert on t2 accessing ':new' values of the update on t1.
But in my Insert statement, I am having get one of the column values from another table. How can I write my insert statement in such a way as to insert values contained in ':new' pseudo columns and a select from another table. Below is my insert statement in the trigger : -------
IF (:old.GROUP_YELLOW <> :new.GROUP_YELLOW) THEN INSERT INTO TEST.W_THRESHOLD_LOG (THRESHOLD_LOG_WID, CHANGE_DATE, MEASURE_TYPE_WID, MEASURE_NAME, CUSTOMER_WID, CUSTOMER_NAME, USER_ID, CHANGED_ITEM, PREV_VALUE, NEW_VALUE) VALUES(TEST.W_THRESHOLD_LOG_SEQ.NEXTVAL, SYSDATE, :new.MEASURE_TYPE_WID, 'Rolling Stabilty' , :new.CUSTOMER_WID, 'Customer1', 'User1', 'GROUP_YELLOW', :old.GROUP_YELLOW , :new.GROUP_YELLOW); END IF; -------
In the above code if the hardcoded value 'Customer1' need to be picked from another table, i.e .
SELECT NAME FROM W_CUSTOMER_DIM WHERE CUSTOMER_WID = THRESHOLD.CUSTOMER_WID
how can I rewrite my query to the above value from the select into my insert statement..?
