Dynamic SQL is a valid option?
HelloI have question on insertion I have a large amount of lines that I want to insert into a table. Insert rows are like;
001 09 00 FF
001 09 FF 11
FF 00 11 001
001 00 01 00
There are four types of conditions. Only the last condition can be inserted, while the rest cannot be inserted (in other words, on four columns to all who have the value 'FF' must not be inserted).
I wanted to use a simple constraint checking, but our DBA is not agree on this. The reason is that overall with many conditions will create a problem with the database-level locking.
Now, I think to use dynamic SQL and the place of the condition in the WHERE clause. I think right and make any sense?
I think I have to be aware of performance because there will be a mass insertion.
Published by: uk113 on July 28, 2009 06:03
In fact, I agree with kind of your DBA, but not his reasons. A check against the content of a column constraint will not cause a locking more in the database column without constraint. If you update a row in the table, Oracle will have a lock on the rows updated if they are limited or not. A newly inserted row never "locked" because it is no longer visible in the session that he introduced to the validation of this session.
You say "I have a lot of lines that I want to insert into a table" which implies to me that you have the source data in an accessible form electronically, which means you can do single insertion as a SQL statement. As many others have suggested, I would use a predicate on each of the columns that cannot contain FF to remove these rows from the result set that you will use to make the insertion.
Given the constraints on the table for a loading block, particulalry if you know you will have a fairly large number of rows in your data source, which will infringe the constraint actually will force you in a row by row (or bulk collect bulk collect) scenario so that you can catch constraint violations and throw these lines. Although you can still use a predicate to filter the lines that would not allow the validation.
If you need constraints after loading mass will depend on whether or not you expect to get new data in the table, and how it will be inserted or updated. If new data come mainly from single row inserts and updates (for example from the application screen), then I would certainly limit the columns to allow not FF. If you do only loads in bulk, I would add probably not these constraints. If you are likely to have a mix of bulk and individual tips, I add the constraints, but still have a predicate in loads in bulk to filter incorrect records.
John
Tags: Database
Similar Questions
-
The generation dynamic SQL when records using custom
Hi all
We have a few custom folders (we did not create the views) where we joined the table of facts with several dimensions of code so that users don't need to go through several folders. Currently every time a user selects a field in the record, it will automatically joins all the dims in accordance with the definition of record. Is it possible, when the user select only the fact of the table fields in the folder then the sql access only the table of facts (as in dynamic sql) leaving aside the joints.
I'm not positive all able to do so. But even post the question to check if there is a solution found by any one.
Here's the hypothetical scenario:
SQL for my custom folder:
Select
EMP.emp_id,
EMP.emp_name,
job_id EMP.,.
jobs. job_desc,
EMP.location_id,
Location.location_desc,
EMP. Salary
of the emp, jobs, location
where emp.job_id = jobs.job_id
and emp.location_id = location.location_id;
In the example above, when the user select only the emp_id and emp_name, discoverer only accesses the emp table, leaving aside joins jobs and location tables?
Appreciate your answers.Hello
There are 3 approaches you can take, which is best for you depends on your database and the version of discoverer and how much change you want to make:
1. optimization of data - if you use database 11 g and the query sent to the database does not refer to an attachment internal table in the first select statement in the query, for example, the query
Select
EMP.emp_id,
EMP.emp_name,
job_id EMP.,.
jobs. job_desc,
EMP.location_id,
EMP. Salary
of the emp, jobs, location
where emp.job_id = jobs.job_id
and emp.location_id = location.location_id;does not reference the table of localization, so as long as location_id is the primary key for location and emp.location_id has a validated and enabled foreign key constraint then the database query plan will not otherwise the table of localization.
2 synthesis discoverer files - you can manually set the summary files in admin discoverer that contain queries custom with less joins of tables. Thus, for example, you could put
Select
EMP.emp_id,
EMP.emp_name,
job_id EMP.,.
jobs. job_desc,
EMP.location_id,
EMP. Salary
of the emp, jobs
where emp.job_id = jobs.job_id;in a personalized summary folder and discoverer would redirect the request to the summary file if the location has not been used. You can also do a similar thing to the database level using the views and materialized views.
3. complex records discoverer - you can change your custom folder in a folder complex bringing together two or more straightforward cases. If the columns mapped to a single file are not reference, then all joins and discoverer preferences are set correctly the generated query does not reference the table in the simple folder.
Rod West
-
Dynamic SQL with dynamic identifiers
Oracle 10gXE
I have a table that is used to track student attendance. If the student is present on a given day, a record is inserted with the student ID and the date they frequented.
presence of DESC
Name of Type Null
--------------- -------- ------
ID NOT NULL NUMBER
NUMBER OF STUDENT_ID
SITE_ID NUMBER
DATE OF ATTENDANCE_DATE
I want to view a report (in TOP) which shows all students for a given site, the days when school was in session for this month (not Saturday, Sunday) and a X for the student who participated in that day or a null value if they do not have.
ID # FNAME LNAME 3 4 5 6 7 10
38754636 POE JANE X X X
81248754 DOE KAYLA X X X X X
43127409 RAO JOHN X X X X X
In this example, days 3 and 7 of the month are from Monday to Friday and 8-9 have been excluded because they are on Saturday and Sunday.
I am trying to find a way to dynamically generate a query for any month/year and returning to a report. The obstacle is that given that I don't know in advance which days in a month will be excluded, the identifiers in my select statement are not fixed.
I found a query that returns a list of day numbers and dates for a given month. It looks like this:
SELECT LEVEL lv, TO_DATE (TO_CHAR (LEVEL, '09')
|| TO_CHAR (EXTRACT (TO_DATE('01-SEP-11') MONTHS), '09')
|| To_char (EXTRACT (TO_DATE('01-SEP-11') YEAR), ' 9999'), 'dd.mm.yyyy') mon_day
OF double WHERE ROWNUM < = EXTRACT (DAY OF LAST_DAY (TO_DATE('01-SEP-11')))
CONNECT BY LEVEL = ROWNUM
I combined this request with another brings my student attendance data and results that pivots in what I need. Then, I created a function that accepts the month and year as parameters and creates the combined request.
SELECT c.ID student_id, c.last_name, c.first_name,.
MAX (decode (mon_day, ' 01 - SEPT.-11', 'X', NULL)) '1', MAX (decode (mon_day, ' 02 - SEVEN.-11', 'X', NULL)) '2 ',.
MAX (decode (mon_day, ' 05 - SEPT.-11', 'X', NULL)) '5', MAX (decode (mon_day, ' 06 - SEPT.-11', 'X', NULL)) "6."
MAX (decode (mon_day, ' 07-SEP-11', 'X', NULL)) '7', MAX (decode (mon_day, ' 08 - SEPT.-11', 'X', NULL)) "8."
MAX (decode (mon_day, ' 09 - SEPT.-11', 'X', NULL)) '9', MAX (decode (mon_day, 12-SEP-11', 'X', NULL)) "12."
MAX (decode (mon_day, 13-SEP-11', 'X', NULL)) '13', MAX (decode (mon_day, 14-SEPT-11', 'X', NULL)) "14."
MAX (decode (mon_day, 15-SEP-11', 'X', NULL)) '15', MAX (decode (mon_day, 16-SEP-11', 'X', NULL)) "16."
MAX (decode (mon_day, 19-SEP-11', 'X', NULL)) '19', MAX (decode (mon_day, 20-7.-11', 'X', NULL)) "20."
MAX (decode (mon_day, 21-SEP-11', 'X', NULL)) "21", MAX (decode (mon_day, 22-SEP-11', 'X', NULL)) '22 ',.
MAX (decode (mon_day, 23-SEP-11', 'X', NULL)) '23', MAX (decode (mon_day, 26-SEP-11', 'X', NULL)) "26."
MAX (decode (mon_day, 27-SEP-11', 'X', NULL)) '27', MAX (decode (mon_day, 28-SEP-11', 'X', NULL)) "28."
MAX (decode (mon_day, 29-SEP-11', 'X', NULL)) '29', MAX (decode (mon_day, 30-SEP-11', 'X', NULL)) '30 '.
(SELECT level lv, TO_DATE (TO_CHAR (LEVEL, '09')
|| TO_CHAR (EXTRACT (TO_DATE('01-SEP-11') MONTHS), '09')
|| To_char (EXTRACT (TO_DATE('01-SEP-11') YEAR), ' 9999'), 'dd.mm.yyyy') mon_day
OF double WHERE ROWNUM < = EXTRACT (DAY OF LAST_DAY (TO_DATE('01-SEP-11')))
CONNECTION LEVEL = ROWNUM) a, b of attendance, student c
WHERE to_char(MON_DAY,'D') NOT IN (1.7)
AND A.mon_day = b.attendance_date
AND b.student_id = c.ID
Group of c.ID, student_id, c.last_name, c.first_name
Now I'm stuck on what I can pass this request in to get my result. What I've read, dynamic SQL method 4 would work if I incorporate Pro/C. I hope this isn't my only option. I have considered to try to do this in a function table in pipeline, but since the columns returned are not known in advance, I can't create a corresponding data type.
I'm starting to wonder if I forgot a simpler method to achieve this.
Published by: David Sumner on April 12, 2012 18:52Ignoring the actual query and looking at the dynamic SQL part of the question.
There are 3 ways to run a dynamic select SQL in PL/SQL code.
immediate execution requires a fixed number of bind variables, linking him in position and translates a single implicit output cursor extraction - requiring the projection of cursor to be known at the time of coding.
REF CURSOR are essentially the same - the only exception being that it requires an extraction explicit coding process output cursor.
DBMS_SQL differs in all of these respects. The connection is by name and not position. The connection is dynamic. Projection of the cursor must not be known at the time of coding. It is dynamically determined at runtime and fetch offers of release of the cursor with the projection of unknown SQL - coding time.
Apex uses DBMS_SQL internally (in fact he uses DBMS_SYS_SQL properly run dynamic and specific patterns of Oracle SQL, as if a session connected as a scheme runs the SQL code). This allows the Apex to take a dynamic SQL that you enter a region (which can contain a variable number of bind variable), run it and make the projection of the cursor as an HTML report.
Apex supports in turn also dynamic SQL - what you (Apex developer) to provide a dynamic SQL code for a region, instead of you having to provide SQL fixed (with bind variable) for the region in question.
As Apex creates DBMS_SQL sliders, it doesn't have your SQL as a ref cursor or cursor DBMS_SQL. All they need is the source of the SQL statement. And that's what you create dynamically.
So in the Apex, dynamic SQL means using a function from PL/SQL to return the source for Apex SQL analyze like a slider DBMS_SQL - instead of coding the source SQL statement in a fixed statement.
So, using the option of function for an area considered in the Apex, allows you to provide the following in the code that must run to get the SQL statement for the report Apex (Apex running this as a dynamic function that returns a string):
--// return the dynamic SQL for the report region return( case when :P1_OPTION = 1 then 'select * from emp' when :P1_OPTION = 2 then 'select * from dept where dept_id = :P1_DEPT_ID' end );
You can also write a PL/SQL function that is stored in the database and call this function so that it can determine what is the dynamic SQL source code statement. For example
--// calling database function to return the dynamic SQL for the report region return( GetDynamicReportSQL( page => 1, option => :P1_OPTION ) );
Dynamic SQL is fully supported by Apex - but does not require that you, the developer, to create the dynamic cursor and treat the dynamic linking and dynamic recovery of this slider.
All you have to do is to provide the source code of this dynamic SQL (including the bind variable) to the Apex and it will do the whole thing from cursor for you.
-
GETTING AN ERROR - slider 'P_REFCUR' cannot be used in the OPEN dynamic SQL stat
DECLARE
create or replace procedure partial_single (p_fileid in NUMBER, p_filename IN VARCHAR2 (2000), p_temptablename IN VARCHAR2 (2000), p_temppartialtablename IN VARCHAR2 (2000), p_retval ON the NUMBER)
--)
p_fileid NUMBER;
p_filename VARCHAR2 (2000);
p_temptablename VARCHAR2 (2000);
p_temppartialtablename VARCHAR2 (2000);
p_retval NUMBER;
p_refcur types_pkg.return_cur;
v_strquery varchar2 (4000);
BEGIN
p_fileid: = 5080;
p_filename: = "TAGSUR1HM2011013111160838654.000019265";
-p_temptablename: = "TEMP_RECORDS_MED_0004";
p_temptablename: = 'TEMP_RECORDS_MED_DATE ';
p_temppartialtablename: = 'TEMP_MED_PARTIAL_RECORDS_0002 ';
p_retval: = 0;
v_strquery: = 'SELECT imsi, connectedcallingnumber, callstart MIN (calleventstarttimestamp), SUM (calleventduration), MAX (sequence_number) sequencenumber, msisdn, max (callreleasetime) callreleasetime period';
v_strquery: = v_strquery | ' A ' | p_temppartialtablename | ' GROUP BY connectedcallingnumber, imsi, msisdn';
OPEN p_refcur - GETTING AN ERROR - slider 'P_REFCUR' cannot be used in a dynamic OPEN SQL statement
FOR v_strquery;
FOR CC IN (SELECT imsi,
connectedcallingnumber,
Callstart MIN (calleventstarttimestamp),
Duration of the SUM (calleventduration),
SequenceNumber MAX (sequence_number),
MSISDN,
Max (callreleasetime) callreleasetime
OF TEMP_MED_PARTIAL_RECORDS_0002
GROUP BY connectedcallingnumber, imsi, msisdn)
LOOP
UPDATE TEMP_RECORDS_MED_DATE
SET calleventstarttimestamp = cc.callstart,
calleventduration = calleventduration + cc.duration
WHERE connectedcallingnumber = cc.connectedcallingnumber
AND imsi = cc.imsi
AND sequencenumber = cc.sequencenumber + 1
AND msisdn = cc.msisdn
AND calleventstarttimestamp = cc.callreleasetime;
IF SQL % ROWCOUNT > 0
THEN
UPDATE TEMP_MED_PARTIAL_RECORDS_0002
SET isprocessed = 1
WHERE connectedcallingnumber = cc.connectedcallingnumber
AND imsi = cc.imsi
AND msisdn = cc.msisdn;
- AND callreleasetime = cc.callreleasetime;
ON THE OTHER
UPDATE TEMP_MED_PARTIAL_RECORDS_0002
SET calleventduration = calleventduration + cc.duration
IMSI WHERE = cc.imsi
AND msisdn = cc.msisdn
AND callreleasetime = cc.callreleasetime
AND calleventduration! = cc.duration;
END IF;
END LOOP;
REMOVE FROM TEMP_MED_PARTIAL_RECORDS_0002
WHERE isprocessed = 1;
p_retval: = 0;
-VALIDATION;
/ * EXCEPTION
WHILE OTHERS
THEN
-RESTORATION;
p_retval: = 1;
p3_errorlog ("partial" p_fileid, SQLERRM, |) ':' || p_filename);
COMMIT;
*/
END;Is - what your refcursor has a return type? In this case you can not open it with dynamic SQL. Change the Refcursor as a weakly typed cursor and give it a try.
-
ORA-00922: missing or not valid option?
Can someone tell me what is the problem with this statement:
SQL > CREATE TABLE CUSTOMER2
2 2 (employe_id NUMBER (10),)
3 3 FIRST NAME VARCHAR2 (40),
4 4 LASTNAME VARCHAR2 (40),
5 5 VARCHAR2 (40) OF COMPANYNAME,.
6 6 PHONE VARCHAR2 (40),
7 7 E-MAIL VARCHAR2 (40),
8 8 CITY VARCHAR2 (40),
9 9 VARCHAR2 (40) STATE.
10 10 ZIP VARCHAR2 (40)
11 (11);
2 (employe_id NUMBER (10),)
*
ERROR on line 2:
ORA-00922: missing or not valid optionInstead of your code with below
CREATE TABLE CUSTOMER2 ( employee_id NUMBER(10), FIRSTNAME VARCHAR2(40), LASTNAME VARCHAR2(40), COMPANYNAME VARCHAR2(40), PHONE VARCHAR2(40), EMAIL VARCHAR2(40), CITY VARCHAR2(40), STATE VARCHAR2(40), ZIP VARCHAR2(40) );
-
Dynamic SQL instead of SQLLOADER
Hello
I use dynamic SQL to load and transform data files.
It works well, except that for reasons of speed I load the entire file into a temporary table before loading to its final destination along the lines of; -
It works very fast, BUT my problems come when; my example above; a field contains a non-numeric is loaded into a numeric field.insert into ap_invoices_interface (invoice_num, invoice_amount, etc....) (select field1, field2, etc from temp_table);
Is anyway, once again using dynamic SQL I can QUICKLY validate before for this kind of error, without doing it on a row by row basis?
NB: I have a table xx_interface_mapping - which maps field1 = > invoice_num, Field2 = > invoice_amount etc. - and I put this together to make the insert statement, is it possible to combine with all_tables to know the type of data in the fields and ensure that they are valid - ALL IN the SQL DYNAMICS zippy instructions?
Thanks for the suggestions - I is not looking for you to write the code, suggest only the concepts to use to get there...
Robert.You can also the values digital alpha eliminate by this way.
with TABLE1 as)
Select "999,99' column1 of union double all the»
Select ' + 999.112' of all the double union. "
Select "12321.54"of any double union. "
Select '-12ABC0.99' of all the double union
Select '122' in double union
Select '9878' across double Union
Select "xyzd" from double
)
SELECT * FROM TABLE1
where LENGTH (TRIM (TRANSLATE (column1, ' + -. 0123456789', ' '))) a null value -
the number of columns in dynamic sql
Hello
I was wondering if you could count the number of columns had a dynamic sql code.
Very simple eg.
Say I had a string 'select 1, 2,3,4,5,6 double' and I only immediate excute. Is it possible to count the number of column IE 6 columns.
I know s/n | THE USER | ALL_TAB_COLUMNS, but it could be a few different columns columns, and it is not a view.
Thanks for you advice/help.with the "immediate execution" option you cannot get metadata such as the number of columns for the executed query. Use rather dbms_sql...
Research of DESCRIBE_COLUMNS string in the link below... It will give you an idea of how to make to get the same.
http://www.morganslibrary.org/reference/DBMS_SQL.htmlRavi Kumar
Published by: ravikumar.sv on July 30, 2010 15:51
-
Difference between static &; dynamic sql
Hello
It is my first question in OTN.
I have a pl/sql procedure with no insert statements.
Insert into table values...
If I turn all these insert statements in dynamic sql, my procedure will give more effective? (execution)
Please help me.
Thanks in adv
Friend
Hello
8b00c42d-7716-467e-BCEF-c756fd9599e2 wrote:
Hello
It is my first question in OTN.
Welcome to the Forum!
To get the most out of this Forum, see the FAQ in the Forum: Re: 2. How can I ask a question on the forums?
For all performance issues, see also this Forum FAQ page: Re: 3. how to improve the performance of my query? / my query is slow.
I have a pl/sql procedure with no insert statements.
Insert into table values...
If I turn all these insert statements in dynamic sql, my procedure will give more effective? (execution)
Probably not. Dynamic SQL is generally less effective than static SQL, as well as being more difficult to code, debug and maintain.
Is there a reason why you think that the dynamic SQL would be preferable in this case? If so, what is it?
-
A loop dynamic sql in the procedure
Hi guys,.
I'm having a procedure that will have two input start_date and end_date parameters
I need a loop in alter procedure statement something like this using dynamic sql
CHANGE TABLE M1
SWAP PARTITION FOR (TO_DATE('01-JAN-2015','dd-MON-yyyy'))
WITH TABLE T2
INCLUDING THE INDEX;
CHANGE TABLE M1
BY EXCHANGE COMPETITION FOR (TO_DATE('02-JAN-2015','dd-MON-yyyy'))
WITH TABLE T3
INCLUDING THE INDEX;
.
.
.
For example I have dates in setting something like January 1, 2015-4 January 2015
I need something like this, where T is a constant variable, I need to add T1, T2. T3 according to the dates of the range.
Hello
your procedure can not simply be rolled up, but the call session can be rolled up the output to a file.
Try something like this
col act_date noprint new_value act_date SET TERMOUT OFF select to_char(sysdate, 'yyyy_mm_dd_hh24miss') act_date from dual; SET TERMOUT ON set serveroutput on set feedback off set linesize 3000 col statement for a200 heading "-- alter statements" spool d:\temp\&act_date._&_CONNECT_IDENTIFIER._alter_partition.sql -- exec your_procedure -- or simple plain sql with data (start_date, end_date) as ( select to_date('01/01/2016', 'dd/mm/yyyy'), to_date('05/01/2016', 'dd/mm/yyyy') from dual ) select --level, to_char(start_date + (level - 1), 'dd-MON-yyyy' ) cur_date 'ALTER TABLE M1 EXCHANGE PARTITION FOR (TO_DATE(''' || to_char(start_date + (level - 1), 'dd-MON-yyyy' ) || ''',''dd-MON-yyyy'')) WITH TABLE T2 INCLUDING INDEXES;' || chr(10) || 'ALTER TABLE M1 EXCHANGE PARTITION FOR (TO_DATE(''' || to_char(start_date + (level - 1), 'dd-MON-yyyy' ) || ''',''dd-MON-yyyy'')) WITH TABLE T3 INCLUDING INDEXES;' as statement from data connect by level <= end_date - start_date + 1; spool off set feedback on prompt @d:\temp\mk_alter.sql
concerning
Kay -
Error when using native dynamic SQL
Please can someone help me understand what's wrong with my code.
Code:
create or replace procedure p1 (name varchar2)
as
Start
run immediately "select * from ' |" name;
end;
/
Goal: Just to test the native dynamic SQL code using a placeholder.
COmpliation:
SQL > create or replace procedure p1 (name varchar2)
2 as
3
4 start
5 run immediately "select * from ' |" name;
6 end;
7.
Created procedure.
Error:
SQL > exec p1 ('employees');
BEGIN p1 ('employees'); END;
*
ERROR on line 1:
ORA-00923: THE KEYWORD not found where expected
ORA-06512: at "HR. "P1", line 5
ORA-06512: at line 1
SQL >
You need a space after the clause 'from', like this:
run immediately "select * from ' |" name;
-
How to pass the Record of entry to the dynamic SQL Code
Hi all
I am trying to run under Oracle applications API on the remote database using DB Link. I try to run using dynamic SQL. Part of the code is given below.
DECLARE
l_batch_info_rec wsh_picking_batches_pub.batch_info_rec;
l_rule_id NUMBER;
l_rule_name VARCHAR2 (240);
l_batch_prefix VARCHAR2 (240);
BEGIN
l_batch_info_rec.document_set_id: = v_document_set_id;
l_batch_info_rec.order_type_id: = v_order_type_id;
l_batch_info_rec. Default_Stage_Subinventory: = v_default_stage_subinventory;
l_batch_info_rec.pick_grouping_rule_id: = v_pick_grouping_rule_id;
l_batch_info_rec.pick_sequence_rule_id: = v_pick_sequence_rule_id;
l_batch_info_rec.autopack_level: = v_autopack_level;
l_batch_info_rec.autopack_flag: = v_autopack_flag;
l_batch_info_rec.ac_delivery_criteria: = v_ac_delivery_criteria;
l_batch_info_rec.backorders_only_flag: = v_backorders_only_flag;
l_batch_info_rec.existing_rsvs_only_flag: = v_existing_rsvs_only_flag;
l_batch_info_rec.customer_id: = v_customer_id;
l_batch_info_rec.order_header_id: = v_header_id;
l_batch_info_rec.from_scheduled_ship_date: = NULL;
l_batch_info_rec.organization_id: = v_organization_id;
l_batch_info_rec.include_planned_lines: = v_include_planned_lines;
l_batch_info_rec.autocreate_delivery_flag: = v_autocreate_deliveries_flag;
l_batch_info_rec.autodetail_pr_flag: = v_autodetail_pr_flag;
l_batch_info_rec.allocation_method: = 'I ';
l_batch_info_rec.pick_from_locator_id: = NULL;
l_batch_info_rec.auto_pick_confirm_flag: = 'n'; -The value manually to run Transact order
l_batch_info_rec.autopack_flag: = 'n'; -Manually configure to perform the manual packaging
l_rule_id: = NULL;
l_rule_name: = NULL;
l_batch_prefix: = NULL;
v_sqlstmt1: = ' START
wsh_picking_batches_pub.create_batch' | g_db_link | "(1.0,
fnd_api.g_true,
fnd_api.g_true,
: l_return_status,.
: l_msg_count,.
: l_msg_data,.
: l_rule_id,.
: l_rule_name,.
: l_batch_info_rec,.
: l_batch_prefix,.
: p_new_batch_id
);
END;';
--
EXECUTE IMMEDIATE V_sqlstmt1
With the HELP OF THE l_return_status,.
ON l_msg_count,.
ON l_msg_data,.
BY l_rule_id,
BY l_rule_name,
BY l_batch_info_rec,
BY l_batch_prefix,
OUT p_new_batch_id;
--
END;
After you run this script, I get below error "PLS-00457: expressions must be SQL types ' for variable l_batch_info_rec which is of type record. can someone please guide me how can I pass variable type dynamic SQL record.
Thank you
Priyanka
Food for thought:
On remote db:
SQL > create or replace
package 2 pkg
3 is
4 type r_type is (record
Number 5,
6 name varchar2 (10)
7 );
(p) 8 procedure
9 p_rec in r_type,
10 p_out out varchar2
11 );
12 end;
13.Package created.
SQL > create or replace
package 2 body pkg
3 is
4 procedure p)
5 p_rec in r_type,
6 p_out out varchar2
7 )
8 is
9 start
10 p_out: = "ID = ' |" p_rec.ID | 'Name =' | p_rec. Name;
11 end;
12 end;
13.Package body created.
SQL >
On local db:
SQL > set serveroutput on
SQL > declare
v_rec 2 pkg.r_type@pdb1sol12;
3 v_out varchar2 (50);
4 start
v_rec.ID 5: = 1;
6 v_rec.name: = 'XXX ';
7 immediate execution"
8 begin
9 pkg.p@pdb1sol12 (: 1,: 2);
10 end;'
11 using v_rec,
12 v_out;
13 dbms_output.put_line (v_out);
14 end;
15.
using v_rec;
*
ERROR on line 11:
ORA-06550: line 11, column 20:
PLS-00457: expressions must be SQL types
ORA-06550: line 7, column 5:
PL/SQL: Statement ignoredSQL > declare
2 number of v_id: = 1;
3 v_name varchar2 (10): = 'XXX ';
4 v_out varchar2 (50);
5. start
6 immediate execution"
7 report
8 v_rec pkg.r_type@pdb1sol12.
9 begin
10 v_rec.id := :1;
11 v_rec.name: =: 2;
12 pkg.p@pdb1sol12 (v_rec,: 3);
13 end;'
14 using v_id,
15 in v_name,
16 v_out;
17 dbms_output.put_line (v_out);
18 end;
19.
ID = 1 name = XXXPL/SQL procedure successfully completed.
SQL >
SY.
-
Is Hi possible to create dynamic sql statements to create a select statement
for example I have table with 50 lines, I need to contract a statemnt for all columns selection how can I do this Dynamics
Hi Alex,
Start with something like this
select 'select ' || LISTAGG(column_name, ', ') within GROUP (ORDER BY COLUMN_ID ) || ' from ' || table_name || ';' as "-- select statement" from user_tab_columns where table_name like upper ( '&1' ) group by table_name;
concerning
Kay -
How the date can be filled with timestamp in dynamic sql
Hi all
Below we have sql
v_stmt: = ' MERGE IN MTH_EQUIPMENTS_EXT_B ED
WITH THE HELP OF)
SELECT TAG_DATA,
EQUIPMENT_FK_KEY,
WORKORDER_FK_KEY,
SEGMENT_FK_KEY,
SHIFT_WORKDAY_FK_KEY,
HOUR_FK_KEY,
ITEM_FK_KEY,
READ_TIME,
ATTR_GROUP_ID,
RECIPE_NUM,
RECIPE_VERSION,
NVL (FND_GLOBAL. User_Id,-1) l_updated_by,.
NVL (FND_GLOBAL. Login_Id,-1) l_last_update_login
OF MTH_TAG_READINGS_T_STG
WHERE DB_COL = ' | " ' || v_colname | " ' ||') TS
WE (';)
v_stmt: = v_stmt | "ED. EQUIPMENT_PK_KEY = TS. EQUIPMENT_FK_KEY AND
ED. READ_TIME = TS. READ_TIME AND
ED. ATTR_GROUP_ID = TS. ATTR_GROUP_ID)
WHEN MATCHED THEN
UPDATE
GAME ED.'. v_colname |' = TS. TAG_DATA,
ED. LAST_UPDATED_BY = 1, ";
v_stmt: = v_stmt | "ED. LAST_UPDATE_LOGIN = TS.l_last_update_login
WHEN NOT MATCHED THEN
INSERT ('| v_colname |', EXTENSION_ID, EQUIPMENT_PK_KEY, WORKORDER_FK_KEY, SEGMENT_FK_KEY, SHIFT_WORKDAY_FK_KEY, HOUR_FK_KEY, ITEM_FK_KEY, READ_TIME, ATTR_GROUP_ID, LAST_UPDATE_DATE, LAST_UPDATED_BY,';)
v_stmt: =.
v_stmt | ("LAST_UPDATE_LOGIN, CREATED_BY, CREATION_DATE, RECIPE_NUM, RECIPE_VERSION)
VALUES (TS. TAG_DATA, EGO_EXTFWK_S.NEXTVAL, TS. EQUIPMENT_FK_KEY, TS. WORKORDER_FK_KEY, TS. SEGMENT_FK_KEY, TS. SHIFT_WORKDAY_FK_KEY, TS. HOUR_FK_KEY, TS. ITEM_FK_KEY, TS. READ_TIME,';
v_stmt: = v_stmt | "TS. ATTR_GROUP_ID,' | '''|| SYSDATE | "'| ', 1, TS.l_last_update_login, TS.l_updated_by,'|"' | SYSDATE | " ' ||', TS. RECIPE_NUM, TS. RECIPE_VERSION)';
EXECUTE IMMEDIATE v_stmt;
After you run the above SQL last date of update in MTH_EQUIPMENST_EXT_B is filled with sysdate but timestamp is not its always 00:00:00. Can we have timestamp also populated through dynamic sql in the last update date
Kind regards
Amrit
Values to hardcode in the dynamic SQL - never use bind variables:
v_stmt: = ' MERGE IN MTH_EQUIPMENTS_EXT_B ED
WITH THE HELP OF)
SELECT TAG_DATA,
EQUIPMENT_FK_KEY,
WORKORDER_FK_KEY,
SEGMENT_FK_KEY,
SHIFT_WORKDAY_FK_KEY,
HOUR_FK_KEY,
ITEM_FK_KEY,
READ_TIME,
ATTR_GROUP_ID,
RECIPE_NUM,
RECIPE_VERSION,
NVL (FND_GLOBAL. User_Id,-1) l_updated_by,.
NVL (FND_GLOBAL. Login_Id,-1) l_last_update_login
OF MTH_TAG_READINGS_T_STG
WHERE DB_COL =: 1) TS
WE (';)
v_stmt: = v_stmt | "ED. EQUIPMENT_PK_KEY = TS. EQUIPMENT_FK_KEY AND
ED. READ_TIME = TS. READ_TIME AND
ED. ATTR_GROUP_ID = TS. ATTR_GROUP_ID)
WHEN MATCHED THEN
UPDATE
GAME ED.'. v_colname |' = TS. TAG_DATA,
ED. LAST_UPDATED_BY = 1, ";
v_stmt: = v_stmt | "ED. LAST_UPDATE_LOGIN = TS.l_last_update_login
WHEN NOT MATCHED THEN
INSERT ('| v_colname |', EXTENSION_ID, EQUIPMENT_PK_KEY, WORKORDER_FK_KEY, SEGMENT_FK_KEY, SHIFT_WORKDAY_FK_KEY, HOUR_FK_KEY, ITEM_FK_KEY, READ_TIME, ATTR_GROUP_ID, LAST_UPDATE_DATE, LAST_UPDATED_BY,';)
v_stmt: =.
v_stmt | ("LAST_UPDATE_LOGIN, CREATED_BY, CREATION_DATE, RECIPE_NUM, RECIPE_VERSION)
VALUES (TS. TAG_DATA, EGO_EXTFWK_S.NEXTVAL, TS. EQUIPMENT_FK_KEY, TS. WORKORDER_FK_KEY, TS. SEGMENT_FK_KEY, TS. SHIFT_WORKDAY_FK_KEY, TS. HOUR_FK_KEY, TS. ITEM_FK_KEY, TS. READ_TIME,';
v_stmt: = v_stmt | "TS. ATTR_GROUP_ID,: 2, 1, TS.l_last_update_login, TS.l_updated_by,: 3, TS. RECIPE_NUM, TS. RECIPE_VERSION)';
V_stmt EXECUTE IMMEDIATE
With the HELP of v_colname,.
SYSDATE,
SYSDATE;
SY.
-
How to extract the names of columns in dynamic SQL
Hi all
Is it possible to extract all the names of columns in a dynamic query?
In my case according to the user selections that my query will get changed (number of the column, column name and tables that everything can vary).
So now, is it possible to retrieve all the column names of the dynamic query generated?
I am using Oracle 11g (11.2.0.4)
Thank you
Shaz
-
How to get sql % number of dynamic sql code lines
Hello
In this procedure I'm inserting and updating using dynamic sql. Now I want to go back two more PARAMETERS, the parameter should be back the number of lines inserted and how updated by stmtas of UPDATE as well as an INSERT. I'm not able to do can help you on that?
CREATE OR REPLACE PROCEDURE Sp_Array_Test( PV_TGT_NAME IN VARCHAR2, PV_SRC_NAME IN VARCHAR2, PV_PK_COLS IN VARCHAR2, PN_ERR_CD OUT NUMBER, PN_ERR_MSG OUT VARCHAR2) AS TYPE ARR_TAB IS TABLE OF VARCHAR2(4000) INDEX BY BINARY_INTEGER; --UTL_FP UTL_FILE.FILE_TYPE; LV_AN_ BLOCK VARCHAR2(32767); LN_CUR BINARY_INTEGER := DBMS_SQL.OPEN_CURSOR; LN_DESC DBMS_SQL.DESC_TAB; LN_COL_CNT PLS_INTEGER := 0; LV_SEL_UPD_STMT VARCHAR2(4000); LV_SEL_INS_STMT VARCHAR2(4000); ARR_INDX NUMBER := 1; LV_DATA_TYPE VARCHAR2(8); LN_FIND_FLAG NUMBER := 0; LN_TAB ARR_TAB; LV_COLS_ARR ARR_TAB; LV_ERR_MSG VARCHAR2(500); --PROCEDURE FILE_WRITE ( FH_IN IN UTL_FILE.FILE_TYPE, -- STRING_IN IN VARCHAR2 ) IS --BEGIN -- UTL_FILE.PUT_LINE(FH_IN,STRING_IN); -- LV_AN_BLOCK := LV_AN_BLOCK||STRING_IN; --EXCEPTION -- WHEN OTHERS THEN -- RAISE; --END FILE_WRITE; BEGIN -- UTL_FP := UTL_FILE.FOPEN('TEST_DIR', 'TEST.sql', 'W'); LV_SEL_UPD_STMT := 'SELECT A.'||REPLACE(PV_PK_COLS,',','||A.')||' PK_COLS , A.* , B.ROWID FROM '||PV_SRC_NAME||' A, '||PV_TGT_NAME||' B WHERE '; LV_SEL_INS_STMT := 'SELECT A.* FROM '||PV_SRC_NAME||' A WHERE NOT EXISTS (SELECT ''1'' FROM '||PV_TGT_NAME||' B WHERE '; LN_TAB(ARR_INDX) := 'DECLARE '; ARR_INDX := ARR_INDX + 1; LN_TAB(ARR_INDX) := 'CURSOR CUR_VIEW_UPD IS '||LV_SEL_UPD_STMT ; ARR_INDX := ARR_INDX + 1; SELECT SUBSTR(COLS,DECODE(RN,1,1,INSTR(COLS,',',1,RN-1)+1),DECODE(RN,1,INSTR(COLS,',',1,RN)-1,INSTR(COLS,',',1,RN)-INSTR(COLS,',',1,RN-1)-1)) BULK COLLECT INTO LV_COLS_ARR FROM ( SELECT RN, PV_PK_COLS||',' COLS FROM (SELECT ROWNUM RN FROM ALL_OBJECTS WHERE ROWNUM <= LENGTH(PV_PK_COLS)- LENGTH(REPLACE(PV_PK_COLS,','))+1)) ; FOR K IN 1 .. LV_COLS_ARR.COUNT LOOP LV_SEL_UPD_STMT := LV_SEL_UPD_STMT||' A.'||LV_COLS_ARR(K)||' = '; LN_TAB(ARR_INDX) := ' A.'||LV_COLS_ARR(K)||' = '; LV_SEL_UPD_STMT := LV_SEL_UPD_STMT||' B.'||LV_COLS_ARR(K) ||CASE WHEN K = LV_COLS_ARR.COUNT THEN NULL ELSE ' AND ' END; LN_TAB(ARR_INDX) := LN_TAB(ARR_INDX)||' B.'||LV_COLS_ARR(K) ||CASE WHEN K = LV_COLS_ARR.COUNT THEN ' ;' ELSE ' AND ' END; ARR_INDX := ARR_INDX + 1; END LOOP; LN_TAB(ARR_INDX) := 'CURSOR CUR_VIEW_INS IS '||LV_SEL_INS_STMT ; ARR_INDX := ARR_INDX + 1; FOR K IN 1 .. LV_COLS_ARR.COUNT LOOP LV_SEL_INS_STMT := LV_SEL_INS_STMT||' A.'||LV_COLS_ARR(K)||' = '; LN_TAB(ARR_INDX) := ' A.'||LV_COLS_ARR(K)||' = '; LV_SEL_INS_STMT := LV_SEL_INS_STMT||' B.'||LV_COLS_ARR(K) ||CASE WHEN K = LV_COLS_ARR.COUNT THEN NULL ELSE ' AND ' END; LN_TAB(ARR_INDX) := LN_TAB(ARR_INDX)||' B.'||LV_COLS_ARR(K) ||CASE WHEN K = LV_COLS_ARR.COUNT THEN ' );' ELSE ' AND ' END; ARR_INDX := ARR_INDX + 1; END LOOP; LV_ERR_MSG := 'WHILE PARSING SELECT STATEMENT -- '||LV_SEL_UPD_STMT; DBMS_SQL.PARSE(LN_CUR, LV_SEL_UPD_STMT, DBMS_SQL.NATIVE); LV_ERR_MSG := 'WHILE DESCRIBING SELECT STATEMENT -- '||LV_SEL_UPD_STMT; DBMS_SQL.DESCRIBE_COLUMNS(LN_CUR, LN_COL_CNT, LN_DESC); FOR i IN LN_DESC.FIRST .. LN_DESC.LAST LOOP IF LN_DESC(i).col_type = 2 THEN LV_DATA_TYPE := 'NUMBER'; ELSIF LN_DESC(i).col_type = 12 THEN LV_DATA_TYPE := 'DATE'; ELSE LV_DATA_TYPE := 'VARCHAR2'; END IF; LN_TAB(ARR_INDX) := ' T_'||LN_DESC(i).col_name||' DBMS_SQL.'||LV_DATA_TYPE||'_TABLE;'; ARR_INDX := ARR_INDX + 1; END LOOP; LN_TAB(ARR_INDX) := 'BEGIN '; ARR_INDX := ARR_INDX + 1; LN_TAB(ARR_INDX) := ' EXECUTE IMMEDIATE ''ALTER SESSION SET NLS_DATE_FORMAT = ''''DD-MON-YYYY HH24:MI:SS'''''';'; ARR_INDX := ARR_INDX + 1; LN_TAB(ARR_INDX) := ' OPEN CUR_VIEW_UPD;'; ARR_INDX := ARR_INDX + 1; LN_TAB(ARR_INDX) := ' LOOP'; ARR_INDX := ARR_INDX + 1; LN_TAB(ARR_INDX) := ' FETCH CUR_VIEW_UPD BULK COLLECT INTO T_'||LN_DESC(LN_DESC.FIRST).col_name||','; ARR_INDX := ARR_INDX + 1; FOR i IN LN_DESC.FIRST + 1 .. LN_DESC.LAST - 1 LOOP LN_TAB(ARR_INDX) := ' T_'||LN_DESC(i).col_name||','; ARR_INDX := ARR_INDX + 1; END LOOP; LN_TAB(ARR_INDX) := ' T_'||LN_DESC(LN_DESC.LAST).col_name||' LIMIT 500 ;'; ARR_INDX := ARR_INDX + 1; LN_TAB(ARR_INDX) := ' FORALL I IN 1 .. '||'T_'||LN_DESC(LN_DESC.LAST).col_name||'.COUNT '; ARR_INDX := ARR_INDX + 1; LN_TAB(ARR_INDX) := ' UPDATE '||PV_TGT_NAME||' SET '; ARR_INDX := ARR_INDX + 1; LN_FIND_FLAG := 0; FOR I IN LN_DESC.FIRST + 1 .. LN_DESC.LAST-1 LOOP FOR K IN 1 .. LV_COLS_ARR.COUNT LOOP LN_FIND_FLAG := 0; IF LN_DESC(I).COL_NAME = LV_COLS_ARR(K) THEN LN_FIND_FLAG := 1; EXIT; END IF; END LOOP; IF LN_FIND_FLAG = 0 THEN LN_TAB(ARR_INDX) := ' '||LN_DESC(i).col_name||' = '||'T_'||LN_DESC(i).col_name||'(I)'||CASE WHEN I = LN_DESC.LAST-1 THEN ' WHERE ' ELSE ',' END; ARR_INDX := ARR_INDX + 1; END IF ; END LOOP; LN_TAB(ARR_INDX) := ' ROWID = '||'T_'||LN_DESC(LN_DESC.LAST).col_name||'(I) ;'; ARR_INDX := ARR_INDX + 1; LN_TAB(ARR_INDX) := ' COMMIT;'; ARR_INDX := ARR_INDX + 1; LN_TAB(ARR_INDX) := ' EXIT WHEN CUR_VIEW_UPD%NOTFOUND;'; ARR_INDX := ARR_INDX + 1; LN_TAB(ARR_INDX) := ' END LOOP;'; ARR_INDX := ARR_INDX + 1; LN_TAB(ARR_INDX) := ' CLOSE CUR_VIEW_UPD;'; ARR_INDX := ARR_INDX + 1; LN_TAB(ARR_INDX) := ' COMMIT;'; ARR_INDX := ARR_INDX + 1; LN_TAB(ARR_INDX) := ' OPEN CUR_VIEW_INS;'; ARR_INDX := ARR_INDX + 1; LN_TAB(ARR_INDX) := ' LOOP'; ARR_INDX := ARR_INDX + 1; LN_TAB(ARR_INDX) := ' FETCH CUR_VIEW_INS BULK COLLECT INTO T_'||LN_DESC(LN_DESC.FIRST+1).col_name||','; ARR_INDX := ARR_INDX + 1; FOR i IN LN_DESC.FIRST + 2 .. LN_DESC.LAST - 2 LOOP LN_TAB(ARR_INDX) := ' T_'||LN_DESC(i).col_name||','; ARR_INDX := ARR_INDX + 1; END LOOP; LN_TAB(ARR_INDX) := ' T_'||LN_DESC(LN_DESC.LAST-1).col_name||' LIMIT 500 ;'; ARR_INDX := ARR_INDX + 1; LN_TAB(ARR_INDX) := ' FORALL J IN 1 .. '||'T_'||LN_DESC(LN_DESC.FIRST + 1).col_name||'.COUNT '; ARR_INDX := ARR_INDX + 1; LN_TAB(ARR_INDX) := ' INSERT INTO '||PV_TGT_NAME||' ('; ARR_INDX := ARR_INDX + 1; FOR i IN LN_DESC.FIRST + 1 .. LN_DESC.LAST - 1 LOOP LN_TAB(ARR_INDX) := ' '||LN_DESC(i).col_name||CASE WHEN I = LN_DESC.LAST - 1 THEN ' )' ELSE ',' END ; ARR_INDX := ARR_INDX + 1; END LOOP; FOR i IN LN_DESC.FIRST + 1 .. LN_DESC.LAST - 1 LOOP LN_TAB(ARR_INDX) := CASE WHEN I = LN_DESC.FIRST + 1 THEN 'VALUES (' ELSE NULL END ||' T_'||LN_DESC(i).col_name||'(J)'||CASE WHEN I = LN_DESC.LAST - 1 THEN ' ) ;' ELSE ',' END ; ARR_INDX := ARR_INDX + 1; END LOOP; LN_TAB(ARR_INDX) := ' COMMIT;'; ARR_INDX := ARR_INDX + 1; LN_TAB(ARR_INDX) := ' EXIT WHEN CUR_VIEW_INS%NOTFOUND;'; ARR_INDX := ARR_INDX + 1; LN_TAB(ARR_INDX) := ' END LOOP;'; ARR_INDX := ARR_INDX + 1; LN_TAB(ARR_INDX) := ' CLOSE CUR_VIEW_INS;'; ARR_INDX := ARR_INDX + 1; LN_TAB(ARR_INDX) := ' COMMIT;'; ARR_INDX := ARR_INDX + 1; LN_TAB(ARR_INDX) := 'END ;'; ARR_INDX := ARR_INDX + 1; FOR J IN 1 .. LN_TAB.COUNT LOOP -- DBMS_OUTPUT.PUT_LINE( LN_TAB(J)); -- FILE_WRITE(UTL_FP,LN_TAB(J)); LV_AN_BLOCK := LV_AN_BLOCK||LN_TAB(J); END LOOP; -- UTL_FILE.FCLOSE(UTL_FP); EXECUTE IMMEDIATE LV_AN_BLOCK; PN_ERR_CD := 0; PN_ERR_MSG := 'Successful Completion'; EXCEPTION WHEN OTHERS THEN PN_ERR_CD := SQLCODE; PN_ERR_MSG := LV_ERR_MSG||' -- '||SQLERRM ; END; /
Thank you all for your answers. I agree with you all. I have manged this time by adding variable bind, then run it immediately ON aid. I don't know how it's going to be training.
From now on my side no problem I don't thank you.
Maybe you are looking for
-
Cloud email &; gmail emails
Hi all Thank you for taking the time to help me here. I am a small business owner and I'm working on a system solution / 'better' for how to store my emails. I have my apple mail account that I have every email of significance to the business for the
-
How can I get rid of the Skype info at the end of outgoing emails when using firefox?
I uninstalled all the Skype on my computer. I have not Skype add in my firefox tools. But Skype information/spam continues to be added at the end of my emails.
-
Guy you have an idea why my Portege M200 laptop screen flickers in 5-10 minutes after ignition? THX
-
Tecra M11 - 17z - Question about upgrade to Win7 32 bit to 64 bit
I have a Tecra M11 - 17z with the pre-installed OS 32 - bit Windows 7 Professional.I want to change the 64-bit operating system. For this I have to do a 'clean' custom install the 64-bit system.As this is a new laptop and the software bundle that cam
-
I have windows xp cannot install uupdate kb953297 keeps failing
tried several times 1.1 service pak security update kb953297 fails each time, it downloads but fails to install