Oracle dynamic SQL: UDF within execute immediate - sql

I am trying to make some of my code dynamic. While typing the question how to use UDFs in dynamic SQL, I figured out the answer:
One can call the UDF from outside!
This works:
Update my_table
Set col1 = get_some_value(col2,col2)
Where 1 = 1;
This did not work:
Execute Immediate '
Update my_table
Set col1 = get_some_value(col2,col3)
Where 1 = 1
';
But this works:
Execute Immediate '
Update my_table
Set col1 = my_package_name.get_some_value(col2,col3)
Where 1 = 1
';
I am using Oracle Database 12c Enterprise Edition Release 12.1.0.2.0
In case you have an idea, how to skip the call from outside, feel free to let me know.
Many greeetings,
Peter

Check your grants and make sure you either include explicit schema owner in the quoted call, use a synonym, or connect directly as schema owner.
Remember that stored procedures execute normally with the privileges of the code creator, so you should make certain that the username you use to run the execute immediate has direct grant (not via a role) access to execute the function.
This works fine in Oracle 12c when logged in as schema owner:
create function myfunc(p_text in varchar2) return varchar2 is
begin
return initcap(p_text);
end;
/
begin
execute immediate 'update emp set ename = myfunc(ename)';
end;
/
select ename from emp;
Returns:
King
Blake
Clark
...
EDIT:
Based on the additional information that function and calling procedure are in the same package, it is likely that the problem is merely naming and scope.
When using execute immediate, the statement is parsed and executed at runtime, by Oracle's SQL engine, with very limited context of the surrounding code.
In short, the payload of execute immediate doesn't know it's running in a package.
Here's a demo that should clear things up a bit.
create or replace package mytest as
function public_func(p_text in varchar2) return varchar2;
procedure demo;
end;
/
create or replace package body mytest as
-------------------------------------------------------------------------------
function public_func(p_text in varchar2) return varchar2 is
begin
return initcap(p_text);
end;
-------------------------------------------------------------------------------
function private_func(p_text in varchar2) return varchar2 is
begin
return lower(p_text);
end;
-------------------------------------------------------------------------------
procedure demo is
begin
-- Test 1 should fail because the function name is not fully qualified
begin
execute immediate 'update emp set ename = public_func(ename)';
exception when others then
dbms_output.put_line('Test1: ' || SQLERRM);
end;
-- Test 2 should pass
begin
execute immediate 'update emp set ename = mytest.public_func(ename)';
exception when others then
dbms_output.put_line('Test2: ' || SQLERRM);
end;
-- Test 3 should fail because the private function is not visible
begin
execute immediate 'update emp set ename = mytest.private_func(ename)';
exception when others then
dbms_output.put_line('Test3: ' || SQLERRM);
end;
end;
end;
/
Here's the results:
SQL> set serveroutput on;
SQL> begin
2 mytest.demo;
3 end;
4 /
Test1: ORA-00904: "PUBLIC_FUNC": invalid identifier
Test3: ORA-00904: "MYTEST"."PRIVATE_FUNC": invalid identifier
PL/SQL procedure successfully completed.
SQL>
For test 1, the SQL engine is looking for something called "public_func" and can't find it. This make sense because you could have two packages that each have something called "public_func" in them. The SQL engine does not know that it is being invoked from within a package.
Test 2 is what you did, and it works as expected.
For test 3, a function is called that exists only within the package body. Normally, other procedures in the package can see private functions, but since this is interpreted at runtime and the SQL engine doesn't know it's being called within the scope of the package, this call fails as well.

Related

How to convert every Table from a specific User to JSON Format using the "trick" provided by SQL Developer

I want to convert all tables from a specific user to JSON (or XML) Format. I've read about a "trick" mentioned by SQL Developer.
In other words, I already started to create a Procedure with two parameters:
p_format: The format (in my case it will be "json")
p_user: The username
As IDE I use Oracle SQL Developer and my database is an Oracle XE Database.
At first the procedure loops though all tables of the given user and in the loop, it should execute the following:
SELECT /*p_format*/ * FROM p_user || '.' || table
Unfortunately, I cannot use this SELECT Statement as mentioned above. I need to use the command EXECUTE IMMEDIATE <Statement>.
The next problem I faced was the following: I wanted to output the result of the EXECUTE IMMEDIATE command. Therefore I used the command EXECUTE IMMEDIATE <Statement> INTO <Variable>. After compiling the procedure and executing it, I stumpled across the following Error:
"inconsistent datatypes: expected %s got %s"
This is my code of the procedure:
CREATE OR REPLACE PROCEDURE EXPORT_TABLE_TO_FORMAT_FROM(p_format VARCHAR2, p_user VARCHAR2) IS
/***************************************************************************
Author:
Class:
School:
Date:
Function - EXPORT_TABLE_TO_JSON_FROM(p_user):
Displays the data of every table from a given User as JSON
Parameter: p_user ... User
***************************************************************************/
v_tableData VARCHAR2(32767);
v_sqlStatement VARCHAR2(200);
BEGIN
FOR tablerec IN (SELECT *
FROM ALL_TABLES
WHERE OWNER = p_user)
LOOP
v_sqlStatement := 'SELECT /*' || p_format || '*/ * FROM ' || p_user || '.' || tablerec.TABLE_NAME;
EXECUTE IMMEDIATE v_sqlStatement INTO v_tableData;
DBMS_OUTPUT.PUT_LINE (v_sqlStatement);
END LOOP;
END;
You can see that I loop though all tables of a given user and created a sql statement with p_format and p_user and with tablerec.TABLE_NAME.
The desired result should look exactly like that:
{"results":[{"columns":[{"name":"COUNTRY_ID","type":"CHAR"},
{"name":"COUNTRY_NAME","type":"VARCHAR2"},{"name":"REGION_ID","type":"NUMBER"}],"items":
[
{"country_id":"AR","country_name":"Argentina","region_id":2},
{"country_id":"AU","country_name":"Australia","region_id":3},
{"country_id":"BE","country_name":"Belgium","region_id":1},
{"country_id":"BR","country_name":"Brazil","region_id":2},
{"country_id":"CA","country_name":"Canada","region_id":2},
{"country_id":"CH","country_name":"Switzerland","region_id":1},
{"country_id":"CN","country_name":"China","region_id":3},
{"country_id":"DE","country_name":"Germany","region_id":1},
{"country_id":"DK","country_name":"Denmark","region_id":1},
{"country_id":"EG","country_name":"Egypt","region_id":4},
{"country_id":"FR","country_name":"France","region_id":1},
{"country_id":"IL","country_name":"Israel","region_id":4},
{"country_id":"IN","country_name":"India","region_id":3},
{"country_id":"IT","country_name":"Italy","region_id":1},
{"country_id":"JP","country_name":"Japan","region_id":3},
{"country_id":"KW","country_name":"Kuwait","region_id":4},
{"country_id":"ML","country_name":"Malaysia","region_id":3},
{"country_id":"MX","country_name":"Mexico","region_id":2},
{"country_id":"NG","country_name":"Nigeria","region_id":4},
{"country_id":"NL","country_name":"Netherlands","region_id":1},
{"country_id":"SG","country_name":"Singapore","region_id":3},
{"country_id":"UK","country_name":"United Kingdom","region_id":1},
{"country_id":"US","country_name":"United States of America","region_id":2},
{"country_id":"ZM","country_name":"Zambia","region_id":4},
{"country_id":"ZW","country_name":"Zimbabwe","region_id":4}]}]}
The JSON hint is specific to SQL Developer and SQLcl, not the database directly. So you need to run the entire thing within these tools.
Easiest way to do that is to have your script write a script that you can then run, eg
spool /tmp/get_all_json.sql
select 'select /*json*/ * from '||table_name||';'
from user_tables;
spool off
#/tmp/get_all_json.sql

Oracle SQL Developer, using dynamic SQL in function

I have the following script which contains a function named 'myFunction'. (declaration of types named rowValueTmp and rowValueTable are also attached for your information) Basically, I need to use a table name as an input parameter for myFunction. I found that I need to use dynamic SQL in order to use the table name as a parameter (Please correct me if there are alternative ways to do this). So the following code is what I have tried so far.
create or replace type rowValueTmp as object (
month number,
year number
);
/
create or replace type rowValueTable as table of rowValueTmp;
/
create or replace FUNCTION myFunction (TABLENAME in VARCHAR2)
return rowValueTable as
v_ret rowValueTable;
begin
execute immediate '
select rowValueTmp(month, year)
bulk collect into v_ret
from '||TABLENAME;
return v_ret;
end myFunction;
/
select * from table(myFunction('SCHEMA.TEST'));
But, this code gives me an error, and I assumed that this error is occurred because of using 'bulk collect' in execute immediate block.
ORA-03001: unimplemented feature
If I replace the content of execute immediate as the following, the above script is working..
select rowValueTmp(month, year)
bulk collect into v_ret
from SCHEMA.TEST;
Question
1] Is there any way(rather than Dynamic SQL) that I can use a table name as an input parameter for myFunction?
2] If I am not allowed to use bulk collect in execute immediate block, what do you suggest?
You can return values from execute immediately into a bulk collect:
CREATE OR REPLACE FUNCTION myfunction (tablename IN VARCHAR2)
RETURN rowvaluetable AS
v_ret rowvaluetable;
v_table VARCHAR2 (61) := DBMS_ASSERT.sql_object_name (tablename);
BEGIN
EXECUTE IMMEDIATE '
select rowValueTmp(month, year)
from ' || v_table
BULK COLLECT INTO v_ret;
RETURN v_ret;
END myfunction;
/
In the interest of an abundance of caution, I'd recommend using DBMS_ASSERT to validate the table parameter as well (as shown).

Inconsistent behaviour when executing a function from within Oracle SQL*Plus?

I am new to Oracle and SQL and am trying to execute a simple test function from within SQL*Plus. My function is called tf (for Test Function) and it is defined in a file called tf.sql as follows ;
create or replace
function
tf
(
arg1 in varchar2
)
return number
as
return_value number;
begin
return_value := 0;
dbms_output.put_line('Argument 1 = ' || arg1);
return return_value;
end;
/
I am able to successfully load this function into Oracle using the following command ;
SQL> start ./tf.sql
As a result of executing this command, SQL*Plus simply states ;
Function created.
When I then execute the following command from the SQL*Plus command prompt (after I have invoked set serveroutput on) ;
SQL> exec dbms_output.put_line(SYSTEM.TF('Hello'));
I get the following output ;
Argument = Hello
0
PL/SQL procedure successfully completed.
Now, if I attempt to execute my function directly from the SQL*Plus command prompt, using the following command ;
SQL> exec SYSTEM.TF('Hello');
then I get presented with the following error message from SQL*Plus ;
BEGIN SYSTEM.TF('Hello'); END;
*
ERROR at line 1:
ORA-06550: line 1, column 7:
PLS-00221: 'TF' is not a procedure or is undefined
ORA-06550: ;ine 1, column 7
PL/SQL: Statement ignored
Is anyone able to shed any light on this for me? I can't work out why my function appears to execute successfully in the first case, but not in the second case.
If I execute the following command from the SQL*Plus command prompt ;
SQL> select * from user_objects where object_name = 'TF';
then I get the following results returned ;
OBJECT_NAME
-----------
TF
SUBOBJECT_NAME
--------------
OBJECT_ID
---------
74475
DATA_OBJECT_ID
--------------
OBJECT_TYPE
-----------
FUNCTION
CREATED
-------
05-FEB-12
LAST_DDL_
---------
05-FEB-12
TIMESTAMP
---------
2012-02-05:02:11:15
STATUS
------
VALID
T
-
N
G
-
N
S
-
N
EDITION_NAME
------------
1
Any help on this would be immensely appreciated.
Thanks in advance.
Craig
exec does not work with functions because it does not know what to do with the return value. This is similar to a regular PL/SQL statement; if you call a function you must assign the return value to something.
If you want to use a function in SQL*Plus, you should use SQL instead:
select tf('asdf') from dual;
Also, you should never create objects in SYSTEM. This can cause some really weird problems.
Following on from #jonearles answer,which highlights the difference between a function and a procedure from SQL*Plus' perspective, and #MS Stp's comment, one way to run it is:
variable rc number;
exec :rc := tf('Hello');
Argument = Hello
PL/SQL procedure successfully completed.
To see the return code you can then do:
print rc
0
exec is really just shorthand for an anonymous PL/SQL block, as you can see from the error message you got. variable lets you declare a bind variable at SQL*Plus level, rather than in the block. You can also declare the argument as bind variable, and set it with a separate exec call:
variable rc number;
variable arg varchar2(5);
exec :arg := 'Hello';
exec :rc := tf(:arg);
I often use this construct for testing an existing procedure call, e.g. something copied from Pro*C code, without having to replace the variables in that call with fixed values. It can make it easier to call repeatedly with different arguments, and you can reuse variables in several calls - so you could pass :rc to another function later.

Oracle Stored Procedure with Alter command

I am trying to build an Oracle stored procedure which will accept a table name as a parameter. The procedure will then rebuild all indexes on the table.
My problem is I get an error while using the ALTER command from a stored procedure, as if PLSQL does not allow that command.
Use the execute immediate statement to execute DDL inside PL/SQL.
create procedure RebuildIndex(index_name varchar2) as
begin
execute immediate 'alter index ' || index_name || ' rebuild';
end;
I tested this code; it works.
Documentation.
Passing Schema Object Names As Parameters
Suppose you need a procedure that
accepts the name of any database
table, then drops that table from your
schema. You must build a string with a
statement that includes the object
names, then use EXECUTE IMMEDIATE to
execute the statement:
CREATE TABLE employees_temp AS SELECT last_name FROM employees;
CREATE PROCEDURE drop_table (table_name IN VARCHAR2) AS
BEGIN
EXECUTE IMMEDIATE 'DROP TABLE ' || table_name;
END;
/
Use concatenation to build the string,
rather than trying to pass the table
name as a bind variable through the
USING clause.
In addition, if you need to call a
procedure whose name is unknown until
runtime, you can pass a parameter
identifying the procedure. For
example, the following procedure can
call another procedure (drop_table) by
specifying the procedure name when
executed.
CREATE PROCEDURE run_proc (proc_name IN VARCHAR2, table_name IN VARCHAR2) ASBEGIN
EXECUTE IMMEDIATE 'CALL "' || proc_name || '" ( :proc_name )' using table_name;
END;
/
If you want to drop a table with the
drop_table procedure, you can run the
procedure as follows. Note that the
procedure name is capitalized.
CREATE TABLE employees_temp AS SELECT last_name FROM employees;
BEGIN
run_proc('DROP_TABLE', 'employees_temp');
END;
/
Here are a couple of possibilities. First, you would have to treat the SQL as dynamic SQL. Second, Oracle DDL statements cannot be run in a transaction (or, they terminate the current transaction and cannot themselves be rolled back). This may affect whether you can use them in stored procedures, or where you can use stored procedures that contain them.
If none of the above apply at all - there could easily be something else astray - I suggest posting some code.

Execute Immediate within a stored procedure keeps giving insufficient priviliges error

Here is the definition of the stored procedure:
CREATE OR REPLACE PROCEDURE usp_dropTable(schema VARCHAR, tblToDrop VARCHAR) IS
BEGIN
DECLARE v_cnt NUMBER;
BEGIN
SELECT COUNT(*)
INTO v_cnt
FROM all_tables
WHERE owner = schema
AND table_name = tblToDrop;
IF v_cnt > 0 THEN
EXECUTE IMMEDIATE('DROP TABLE someschema.some_table PURGE');
END IF;
END;
END;
Here is the call:
CALL usp_dropTable('SOMESCHEMA', 'SOME_TABLE');
For some reason, I keep getting insufficient privileges error for the EXECUTE IMMEDIATE command. I looked online and found out that the insufficient privileges error usually means the oracle user account does not have privileges for the command used in the query that is passes, which in this case is DROP. However, I have drop privileges. I am really confused and I can't seem to find a solution that works for me.
Thanks to you in advance.
SOLUTION:
As Steve mentioned below, Oracle security model is weird in that it needs to know explicitly somewhere in the procedure what kind of privileges to use. The way to let Oracle know that is to use AUTHID keyword in the CREATE OR REPLACE statement. If you want the same level of privileges as the creator of the procedure, you use AUTHID DEFINER. If you want Oracle to use the privileges of the user currently running the stored procedure, you want to use AUTHID CURRENT_USER. The procedure declaration looks as follows:
CREATE OR REPLACE PROCEDURE usp_dropTable(schema VARCHAR, tblToDrop VARCHAR)
AUTHID CURRENT_USER IS
BEGIN
DECLARE v_cnt NUMBER;
BEGIN
SELECT COUNT(*)
INTO v_cnt
FROM all_tables
WHERE owner = schema
AND table_name = tblToDrop;
IF v_cnt > 0 THEN
EXECUTE IMMEDIATE('DROP TABLE someschema.some_table PURGE');
END IF;
END;
END;
Thank you everyone for responding. This was definitely very annoying problem to get to the solution.
Oracle's security model is such that when executing dynamic SQL using Execute Immediate (inside the context of a PL/SQL block or procedure), the user does not have privileges to objects or commands that are granted via role membership. Your user likely has "DBA" role or something similar. You must explicitly grant "drop table" permissions to this user. The same would apply if you were trying to select from tables in another schema (such as sys or system) - you would need to grant explicit SELECT privileges on that table to this user.
You should use this example with AUTHID CURRENT_USER :
CREATE OR REPLACE PROCEDURE Create_sequence_for_tab (VAR_TAB_NAME IN VARCHAR2)
AUTHID CURRENT_USER
IS
SEQ_NAME VARCHAR2 (100);
FINAL_QUERY VARCHAR2 (100);
COUNT_NUMBER NUMBER := 0;
cur_id NUMBER;
BEGIN
SEQ_NAME := 'SEQ_' || VAR_TAB_NAME;
SELECT COUNT (*)
INTO COUNT_NUMBER
FROM USER_SEQUENCES
WHERE SEQUENCE_NAME = SEQ_NAME;
DBMS_OUTPUT.PUT_LINE (SEQ_NAME || '>' || COUNT_NUMBER);
IF COUNT_NUMBER = 0
THEN
--DBMS_OUTPUT.PUT_LINE('DROP SEQUENCE ' || SEQ_NAME);
-- EXECUTE IMMEDIATE 'DROP SEQUENCE ' || SEQ_NAME;
-- ELSE
SELECT 'CREATE SEQUENCE COMPTABILITE.' || SEQ_NAME || ' START WITH ' || ROUND (DBMS_RANDOM.VALUE (100000000000, 999999999999), 0) || ' INCREMENT BY 1'
INTO FINAL_QUERY
FROM DUAL;
DBMS_OUTPUT.PUT_LINE (FINAL_QUERY);
cur_id := DBMS_SQL.OPEN_CURSOR;
DBMS_SQL.parse (cur_id, FINAL_QUERY, DBMS_SQL.v7);
DBMS_SQL.CLOSE_CURSOR (cur_id);
-- EXECUTE IMMEDIATE FINAL_QUERY;
END IF;
COMMIT;
END;
/
you could use "AUTHID CURRENT_USER" in body of your procedure definition for your requirements.
Alternatively you can grant the user DROP_ANY_TABLE privilege if need be and the procedure will run as is without the need for any alteration. Dangerous maybe but depends what you're doing :)