141 DBMS_UTILITY

Using DBMS_UTILITY

• Security Model

• Constants

• Types

• Exceptions

---

Security Model

DBMS_UTILITY runs with the privileges of the calling user for the NAME_RESOLVE Procedure and the COMPILE_SCHEMA Procedure. This is necessary so that the SQL works correctly.

The package does not run as SYS. The privileges are checked using DBMS_DDL.

---

Constants

The DBMS_UTILITY package uses the constants shown in Table 141-1, "DBMS_UTILITY Constants".

---

Types

• dblink_array

• index_table_type

• instance_record

• lname_array

• name_array

• number_array

• uncl_array

dblink_array

TYPE dblink_array IS TABLE OF VARCHAR2(128) INDEX BY BINARY_INTEGER;

Lists of database links should be stored here.

index_table_type

TYPE index_table_type IS TABLE OF BINARY_INTEGER INDEX BY BINARY_INTEGER;

The order in which objects should be generated is returned here.

instance_record

TYPE instance_record IS RECORD (
       inst_number   NUMBER,
       inst_name     VARCHAR2(60));
  TYPE instance_table IS TABLE OF instance_record INDEX BY BINARY_INTEGER;

The list of active instance number and instance name.

The starting index of instance_table is 1; instance_table is dense.

lname_array

TYPE lname_array IS TABLE OF VARCHAR2(4000) index by BINARY_INTEGER;

Lists of Long NAME should be stored here, it includes fully qualified attribute names.

name_array

TYPE name_array IS TABLE OF VARCHAR2(30) INDEX BY BINARY_INTEGER;

Lists of NAME should be stored here.

number_array

TYPE number_array IS TABLE OF NUMBER INDEX BY BINARY_INTEGER;

The order in which objects should be generated is returned here for users.

uncl_array

TYPE uncl_array IS TABLE OF VARCHAR2(227) INDEX BY BINARY_INTEGER;

Lists of "USER"."NAME"."COLUMN"@LINK should be stored here.

---

Exceptions

The following table lists the exceptions raised by DBMS_UTILITY.

Summary of DBMS_UTILITY Subprograms

---

ACTIVE_INSTANCES Procedure

This procedure returns the active instance.

Syntax

DBMS_UTILITY.ACTIVE_INSTANCES (
   instance_table   OUT INSTANCE_TABLE,
   instance_count   OUT NUMBER); 

Parameters

---

ANALYZE_PART_OBJECT Procedure

This procedure is equivalent to SQL:

"ANALYZE TABLE|INDEX [<schema>.]<object_name> PARTITION <pname> [<command_type>] [<command_opt>] [<sample_clause>] 

Syntax

DBMS_UTILITY.ANALYZE_PART_OBJECT (
   schema        IN VARCHAR2 DEFAULT NULL,
   object_name   IN VARCHAR2 DEFAULT NULL,
   object_type   IN CHAR     DEFAULT 'T',
   command_type  IN CHAR     DEFAULT 'E',
   command_opt   IN VARCHAR2 DEFAULT NULL,
   sample_clause IN VARCHAR2 DEFAULT 'sample 5 percent ');

Parameters

Usage Notes

For each partition of the object, run in parallel using job queues.

---

CANONICALIZE Procedure

This procedure canonicalizes the given string. The procedure handles a single reserved or key word (such as 'table'), and strips off white spaces for a single identifier so that ' table ' becomes TABLE.

Syntax

DBMS_UTILITY.CANONICALIZE(
   name           IN    VARCHAR2,
   canon_name     OUT   VARCHAR2,
   canon_len      IN    BINARY_INTEGER);

Parameters

Return Values

Returns the first canon_len bytes in canon_name.

Usage Notes

• If name is NULL, canon_name becomes NULL.

• If name is not a dotted name, and if name begins and ends with a double quote, remove both quotes. Alternatively, convert to upper case with NLS_UPPER. Note that this case does not include a name with special characters, such as a space, but is not doubly quoted.

• If name is a dotted name (such as a."b".c), for each component in the dotted name in the case in which the component begins and ends with a double quote, no transformation will be performed on this component. Alternatively, convert to upper case with NLS_UPPER and apply begin and end double quotes to the capitalized form of this component. In such a case, each canonicalized component will be concatenated together in the input position, separated by ".".

• Any other character after a[.b]* will be ignored.

• The procedure does not handle cases like 'A B.'

Examples

• a becomes A

• "a" becomes a

• "a".b becomes "a"."B"

• "a".b,c.f becomes "a"."B" with",c.f" ignored.

---

COMMA_TO_TABLE Procedures

These procedures converts a comma-delimited list of names into a PL/SQL table of names. The second version supports fully-qualified attribute names.

Syntax

DBMS_UTILITY.COMMA_TO_TABLE ( 
   list   IN  VARCHAR2,
   tablen OUT BINARY_INTEGER,
   tab    OUT uncl_array); 

DBMS_UTILITY.COMMA_TO_TABLE ( 
   list   IN  VARCHAR2,
   tablen OUT BINARY_INTEGER,
   tab    OUT lname_array);

Parameters

Return Values

A PL/SQL table is returned, with values 1..n and n+1 is null.

Usage Notes

The list must be a non-empty comma-delimited list: Anything other than a comma-delimited list is rejected. Commas inside double quotes do not count.

Entries in the comma-delimited list cannot include multibyte characters.

The values in tab are copied from the original list, with no transformations.

---

COMPILE_SCHEMA Procedure

This procedure compiles all procedures, functions, packages, views and triggers in the specified schema.

Syntax

DBMS_UTILITY.COMPILE_SCHEMA (
   schema          VARCHAR2,
   compile_all     BOOLEAN DEFAULT TRUE,
   reuse_settings  BOOLEAN DEFAULT FALSE);

Parameters

Exceptions

Usage Notes

• Note that this subprogram is a wrapper for RECOMP_SERIAL Procedure included with the UTL_RECOMP package.

• After calling this procedure, you should select from view ALL_OBJECTS for items with status of INVALID to see if all objects were successfully compiled.

• To see the errors associated with INVALID objects, you may use the Enterprise Manager command:

  SHOW ERRORS <type> <schema>.<name>
  

---

CREATE_ALTER_TYPE_ERROR_TABLE Procedure

This procedure creates an error table to be used in the EXCEPTION clause of the ALTER TYPE statement.

Syntax

DBMS_UTILITY.CREATE_ALTER_TYPE_ERROR_TABLE(
   schema_name     IN     VARCHAR2,
   table_name      IN     VARCHAR2);

Parameters

Exceptions

An error is returned if the table already exists.

---

CURRENT_INSTANCE Function

This function returns the current connected instance number. It returns NULL when connected instance is down.

Syntax

DBMS_UTILITY.CURRENT_INSTANCE
   RETURN NUMBER;

---

DATA_BLOCK_ADDRESS_BLOCK Function

This function gets the block number part of a data block address.

Syntax

DBMS_UTILITY.DATA_BLOCK_ADDRESS_BLOCK (
   dba NUMBER) 
  RETURN NUMBER;

Parameters

Pragmas

pragma restrict_references(data_block_address_block, WNDS, RNDS, WNPS, RNPS);

Return Values

Block offset of the block.

Usage Notes

This function should not be used with datablocks which belong to bigfile tablespaces.

---

DATA_BLOCK_ADDRESS_FILE Function

This function gets the file number part of a data block address.

Syntax

DBMS_UTILITY.DATA_BLOCK_ADDRESS_FILE (
   dba NUMBER) 
  RETURN NUMBER;

Parameters

Pragmas

pragma restrict_references (data_block_address_file, WNDS, RNDS, WNPS, RNPS);

Return Values

File that contains the block.

Usage Notes

This function should not be used with datablocks which belong to bigfile tablespaces.

---

DB_VERSION Procedure

This procedure returns version information for the database.

Syntax

DBMS_UTILITY.DB_VERSION (
   version       OUT VARCHAR2,
   compatibility OUT VARCHAR2); 

Parameters

---

EXEC_DDL_STATEMENT Procedure

This procedure executes the DDL statement in parse_string.

Syntax

DBMS_UTILITY.EXEC_DDL_STATEMENT (
   parse_string IN VARCHAR2);

Parameters

---

FORMAT_CALL_STACK Function

This function formats the current call stack. This can be used on any stored procedure or trigger to access the call stack. This can be useful for debugging.

Syntax

DBMS_UTILITY.FORMAT_CALL_STACK 
  RETURN VARCHAR2;

Pragmas

pragma restrict_references(format_call_stack,WNDS); 

Return Values

This returns the call stack, up to 2000 bytes.

---

FORMAT_ERROR_BACKTRACE Function

This procedure displays the call stack at the point where an exception was raised, even if the procedure is called from an exception handler in an outer scope. The output is similar to the output of the SQLERRM function, but not subject to the same size limitation.

Syntax

DBMS_UTILITY.FORMAT_ERROR_BACKTRACE 
  RETURN VARCHAR2;

Return Values

The backtrace string. A NULL string is returned if no error is currently being handled.

Examples

CREATE OR REPLACE PROCEDURE Log_Errors ( i_buff in varchar2 ) IS
  g_start_pos integer := 1;
  g_end_pos  integer;

  FUNCTION Output_One_Line RETURN BOOLEAN IS
  BEGIN
    g_end_pos := Instr ( i_buff, Chr(10), g_start_pos );

    CASE g_end_pos > 0
      WHEN true THEN
        DBMS_OUTPUT.PUT_LINE ( Substr ( i_buff, g_start_pos,
g_end_pos-g_start_pos ) );
        g_start_pos := g_end_pos+1;
        RETURN TRUE;

      WHEN FALSE THEN
        DBMS_OUTPUT.PUT_LINE ( Substr ( i_buff, g_start_pos,
(Length(i_buff)-g_start_pos)+1 ) );
        RETURN FALSE;
    END CASE;
  END Output_One_Line;

BEGIN
  WHILE Output_One_Line() LOOP NULL; 
  END LOOP;
END Log_Errors;
/

Set Doc Off
Set Feedback off
Set Echo Off

CREATE OR REPLACE PROCEDURE P0 IS
  e_01476 EXCEPTION; pragma exception_init ( e_01476, -1476 );
BEGIN
  RAISE e_01476;
END P0;
/
Show Errors

CREATE OR REPLACE PROCEDURE P1 IS
BEGIN
  P0();
END P1;
/
SHOW ERRORS

CREATE OR REPLACE PROCEDURE P2 IS
BEGIN
  P1();
END P2;
/
SHOW ERRORS

CREATE OR REPLACE PROCEDURE P3 IS
BEGIN
  P2();
END P3;
/
SHOW ERRORS

CREATE OR REPLACE PROCEDURE P4 IS
  BEGIN P3(); END P4;
/
CREATE OR REPLACE PROCEDURE P5 IS
  BEGIN P4(); END P5;
/
SHOW ERRORS

CREATE OR REPLACE PROCEDURE Top_Naive IS
BEGIN
  P5();
END Top_Naive;
/
SHOW ERRORS

CREATE OR REPLACE PROCEDURE Top_With_Logging IS
  -- NOTE: SqlErrm in principle gives the same info as Format_Error_Stack.
  -- But SqlErrm is subject to some length limits,
  -- while Format_Error_Stack is not.
BEGIN
  P5();
EXCEPTION
  WHEN OTHERS THEN
    Log_Errors ( 'Error_Stack...' || Chr(10) ||
      DBMS_UTILITY.FORMAT_ERROR_STACK() );
    Log_Errors ( 'Error_Backtrace...' || Chr(10) ||
      DBMS_UTILITY.FORMAT_ERROR_BACKTRACE() );
    DBMS_OUTPUT.PUT_LINE ( '----------' );
END Top_With_Logging;
/
SHOW ERRORS

--------------------------------------------------------------------------------

Set ServerOutput On
call Top_Naive()
  /*
  ERROR at line 1:
  ORA-01476: divisor is equal to zero
  ORA-06512: at "U.P0", line 4
  ORA-06512: at "U.P1", line 3
  ORA-06512: at "U.P2", line 3
  ORA-06512: at "U.P3", line 3
  ORA-06512: at "U.P4", line 2
  ORA-06512: at "U.P5", line 2
  ORA-06512: at "U.TOP_NAIVE", line 3
  */
  ;

Set ServerOutput On
call Top_With_Logging()
  /*
  Error_Stack...
  ORA-01476: divisor is equal to zero
  Error_Backtrace...
  ORA-06512: at "U.P0", line 4
  ORA-06512: at "U.P1", line 3
  ORA-06512: at "U.P2", line 3
  ORA-06512: at "U.P3", line 3
  ORA-06512: at "U.P4", line 2
  ORA-06512: at "U.P5", line 2
  ORA-06512: at "U.TOP_WITH_LOGGING", line 6
  ----------
  */
  ;

/*
  ORA-06512:
  Cause:
    Backtrace message as the stack is
    unwound by unhandled exceptions.
  Action:
    Fix the problem causing the exception
    or write an exception handler for this condition.
    Or you may need to contact your application administrator
    or database administrator.
*/

---

FORMAT_ERROR_STACK Function

This function formats the current error stack. This can be used in exception handlers to look at the full error stack.

Syntax

DBMS_UTILITY.FORMAT_ERROR_STACK 
  RETURN VARCHAR2;

Return Values

This returns the error stack, up to 2000 bytes.

Return Values

See FORMAT_ERROR_BACKTRACE Function.

---

GET_CPU_TIME Function

This function returns the current CPU time in 100th's of a second. The returned CPU time is the number of 100th's of a second from some arbitrary epoch.

Syntax

DBMS_UTILITY.GET_CPU_TIME
   RETURN NUMBER;

Return Values

Time is the number of 100th's of a second from some arbitrary epoch.

Usage Notes

This subprogram reports cycles (CPU time) used in performing work, and is unrelated to clock time or any other fixed reference. Since the base line is rubricator, and the relationship between work performed and the number generated is Operating System specific, the amount of work performed is calculated by measuring the difference between a start point and end point for a particular operation.

---

GET_DEPENDENCY Procedure

This procedure shows the dependencies on the object passed in.

Syntax

DBMS_UTILITY.GET_DEPENDENCY
   type      IN     VARCHAR2,
   schema    IN     VARCHAR2,
   name      IN     VARCHAR2);

Parameters

Usage Notes

This procedure uses the DBMS_OUTPUTpackage to display results, and so you must declare SET SERVEROUTPUT ON if you wish to view dependencies. Alternatively, any application that checks the DBMS_OUTPUT output buffers can invoke this subprogram and then retrieve the output by means of DBMS_OUTPUT subprograms such as GET_LINES.

---

GET_HASH_VALUE Function

This function computes a hash value for the given string.

Syntax

DBMS_UTILITY.GET_HASH_VALUE (
   name      VARCHAR2, 
   base      NUMBER, 
   hash_size NUMBER)
  RETURN NUMBER;

Parameters

Pragmas

pragma restrict_references(get_hash_value, WNDS, RNDS, WNPS, RNPS);    

Return Values

A hash value based on the input string. For example, to get a hash value on a string where the hash value should be between 1000 and 3047, use 1000 as the base value and 2048 as the hash_size value. Using a power of 2 for the hash_size parameter works best.

---

GET_PARAMETER_VALUE Function

This function gets the value of specified init.ora parameter.

Syntax

DBMS_UTILITY.GET_PARAMETER_VALUE (
   parnam     IN        VARCHAR2,
   intval     IN OUT    BINARY_INTEGER,
   strval     IN OUT    VARCHAR2)
  RETURN BINARY_INTEGER;

Parameters

Return Values

Parameter type:

• 0 if parameter is an INTEGER/BOOLEAN parameter

• 1 if parameter is a string/file parameter

Usage Notes

When using DBMS_UTILITY.GET_PARAMETER_VALUE, only the first parameter setting of /dir1 is returned when init.ora is set as follows:

utl_file_dir = /dir1
utl_file_dir = /dir2

However, the full comma-delimited string is returned if you are using:

utl_file_dir = /dir1, /dir2

Examples

DECLARE
  parnam VARCHAR2(256);
  intval BINARY_INTEGER;
  strval VARCHAR2(256);
  partyp BINARY_INTEGER;
BEGIN
  partyp := dbms_utility.get_parameter_value('max_dump_file_size',
                                              intval, strval);
  dbms_output.put('parameter value is: ');
  IF partyp = 1 THEN
    dbms_output.put_line(strval);
  ELSE
    dbms_output.put_line(intval);
  END IF;
  IF partyp = 1 THEN
    dbms_output.put('parameter value length is: ');
    dbms_output.put_line(intval);
  END IF;
  dbms_output.put('parameter type is: ');
  IF partyp = 1 THEN
    dbms_output.put_line('string');
  ELSE
    dbms_output.put_line('integer');
  END IF;
END;

---

GET_TIME Function

This function determines the current time in 100th's of a second. This subprogram is primarily used for determining elapsed time. The subprogram is called twice – at the beginning and end of some process – and then the first (earlier) number is subtracted from the second (later) number to determine the time elapsed.

Syntax

DBMS_UTILITY.GET_TIME 
  RETURN NUMBER;

Return Values

Time is the number of 100th's of a second from the point in time at which the subprogram is invoked.

Usage Notes

Numbers are returned in the range -2147483648 to 2147483647 depending on platform and machine, and your application must take the sign of the number into account in determining the interval. For instance, in the case of two negative numbers, application logic must allow that the first (earlier) number will be larger than the second (later) number which is closer to zero. By the same token, your application should also allow that the first (earlier) number be negative and the second (later) number be positive.

---

INVALIDATE Procedure

This procedure invalidates a database object and (optionally) modifies its PL/SQL compiler parameter settings. It also invalidates any objects that (directly or indirectly) depend on the object being invalidated.

Syntax

DBMS_UTILITY.INVALIDATE ( 
   p_object_id              NUMBER,
   p_plsql_object_settings  VARCHAR2 DEFAULT NULL,
   p_option_flags           PLS_INTEGER DEFAULT 0);

Parameters

Exceptions

Usage Notes

The object type (object_type column from ALL_OBJECTS) of the object specified by p_object_id must be a PROCEDURE, FUNCTION, PACKAGE, PACKAGE BODY, TRIGGER, TYPE, TYPE BODY, LIBRARY, VIEW, OPERATOR, SYNONYM, or JAVA CLASS. If the object is not one of these types and the flag inv_error_on_restrictions is specified in p_option_flags then the exception inv_restricted_object is raised, else no action is taken.

If the object specified by p_object_id is the package specification of STANDARD, DBMS_STANDARD, or specification or body of DBMS_UTILITY and the flag inv_error_on_restrictions is specified in p_option_flags then the exception inv_restricted_object is raised, else no action is taken.

If the object specified by p_object_id is an object type specification and there exist tables which depend on the type and the flag inv_error_on_restrictions is specified in p_option_flags then the exception inv_restricted_object is raised, else no action is taken.

Examples

Example 1

DBMS_UTILITY.INVALIDATE (1232, 'PLSQL_OPTIMIZE_LEVEL = 2 REUSE SETTINGS');

Assume that the object_id 1232 refers to the procedure remove_emp in the HR schema. Then the above call will mark the remove_emp procedure invalid and change it's PLSQL_OPTIMIZE_LEVEL compiler setting to 2. The values of other compiler settings will remain unchanged since REUSE SETTINGS is specified.

Objects that depend on hr.remove_emp will also get marked invalid. Their compiler parameters will not be changed.

Example 2

DBMS_UTILITY.INVALIDATE (40775, 'plsql_code_type = native');

Assume that the object_id 40775 refers to the type body leaf_category_typ in the OE schema. Then the above call will mark the type body invalid and change its PLSQL_CODE_TYPE compiler setting to NATIVE. The values of other compiler settings will be picked up from the current session environment since REUSE SETTINGS has not been specified.

Since no objects can depend on bodies, there are no cascaded invalidations.

Example 3

DBMS_UTILITY.INVALIDATE (40796);

Assume that the object_id 40796 refers to the view oc_orders in the OE schema. Then the above call will mark the oc_orders view invalid.

Objects that depend on oe.oc_orders will also get marked invalid.

---

IS_CLUSTER_DATABASE Function

This function finds out if this database is running in cluster database mode.

Syntax

DBMS_UTILITY.IS_CLUSTER_DATABASE 
  RETURN BOOLEAN;

Return Values

This function returns TRUE if this instance was started in cluster database mode; FALSE otherwise.

---

MAKE_DATA_BLOCK_ADDRESS Function

This function creates a data block address given a file number and a block number. A data block address is the internal structure used to identify a block in the database. This function is useful when accessing certain fixed tables that contain data block addresses.

Syntax

DBMS_UTILITY.MAKE_DATA_BLOCK_ADDRESS (
   file  NUMBER, 
   block NUMBER) 
  RETURN NUMBER;

Parameters

Pragmas

pragma restrict_references (make_data_block_address, WNDS, RNDS, WNPS, RNPS);

Return Values

Data block address.

---

NAME_RESOLVE Procedure

This procedure resolves the given name, including synonym translation and authorization checking as necessary.

Syntax

DBMS_UTILITY.NAME_RESOLVE (
   name          IN  VARCHAR2, 
   context       IN  NUMBER,
   schema        OUT VARCHAR2, 
   part1         OUT VARCHAR2, 
   part2         OUT VARCHAR2,
   dblink        OUT VARCHAR2, 
   part1_type    OUT NUMBER, 
   object_number OUT NUMBER);

Parameters

Exceptions

All errors are handled by raising exceptions. A wide variety of exceptions are possible, based on the various syntax error that are possible when specifying object names.

---

NAME_TOKENIZE Procedure

This procedure calls the parser to parse the given name as "a [. b [. c ]][@ dblink ]". It strips double quotes, or converts to uppercase if there are no quotes. It ignores comments of all sorts, and does no semantic analysis. Missing values are left as NULL.

Syntax

DBMS_UTILITY.NAME_TOKENIZE ( 
   name    IN  VARCHAR2,
   a       OUT VARCHAR2,
   b       OUT VARCHAR2,
   c       OUT VARCHAR2,
   dblink  OUT VARCHAR2, 
   nextpos OUT BINARY_INTEGER);

Parameters

For each of a, b, c, dblink, tell where the following token starts in anext, bnext, cnext, dnext respectively.

---

PORT_STRING Function

This function returns a string that identifies the operating system and the TWO TASK PROTOCOL version of the database. For example, "VAX/VMX-7.1.0.0"

The maximum length is port-specific.

Syntax

DBMS_UTILITY.PORT_STRING 
   RETURN VARCHAR2;

Pragmas

pragma restrict_references(port_string, WNDS, RNDS, WNPS, RNPS);

---

TABLE_TO_COMMA Procedures

These procedures converts a PL/SQL table of names into a comma-delimited list of names. This takes a PL/SQL table, 1..n, terminated with n+1 null. The second version supports fully-qualified attribute names.

Syntax

DBMS_UTILITY.TABLE_TO_COMMA ( 
   tab    IN  UNCL_ARRAY, 
   tablen OUT BINARY_INTEGER,
   list   OUT VARCHAR2);

DBMS_UTILITY.TABLE_TO_COMMA ( 
   tab    IN  lname_array,
   tablen OUT BINARY_INTEGER,
   list   OUT VARCHAR2);

Parameters

Return Values

A comma-delimited list and the number of elements found in the table.

---

VALIDATE Procedure

This procedure makes invalid database objects valid.

Syntax

DBMS_UTILITY.VALIDATE(
    object_id       NUMBER);

DBMS_UTILITY.VALIDATE(
   owner          VARCHAR2, 
   objname        VARCHAR2, 
   namespace      NUMBER,   edition_name   := SYS_CONTEXT ('USERENV', 'CURRENT_EDITION'));

Parameters

Usage Notes

• No errors are raised if the object does not exist or is already valid or is an object that cannot be validated.

• If the object being validated is not actual in the specified edition, the subprogram automatically switches into the edition in which the object is actual prior to validation. That is, a call to VALIDATE will not actualize the object in the specified edition.

• The INVALIDATE Procedure invalidates a database object and optionally changes its PL/SQL compiler parameter settings. The object to be invalidated is specified by its object_id. The subprogram automatically switches to the edition in which the object is actual prior to invalidation. That is, a call to INVALIDATE will not actualize the object in the current edition.