Release 8.0
A54642_01
 Library |
 Product |
 Contents |
 Index |
| Oracle8
Server Application Developer's Guide Release 8.0 A54642_01 |
|
This chapter describes the dynamic SQL package, DBMS_SQL. The following topics are described in this chapter:
You can write stored procedures and anonymous PL/SQL blocks that use dynamic SQL. Dynamic SQL statements are not embedded in your source program; rather, they are stored in character strings that are input to, or built by, the program at runtime. This permits you to create procedures that are more general purpose. For example, using dynamic SQL allows you to create a procedure that operates on a table whose name is not known until runtime.
Additionally, you can parse any data manipulation language (DML) or data definition language (DDL) statement using the DBMS_SQL package. This helps solve the problem of not being able to parse data definition language statements directly using PL/SQL. For example, you might now choose to issue a DROP TABLE statement from within a stored procedure by using the PARSE procedure supplied with the DBMS_SQL package.
To create the DBMS_SQL package, submit the DBMSSQL.SQL and PRVTSQL.PLB scripts when connected as the user SYS. These scripts are run automatically by the CATPROC.SQL script. See "Privileges Required to Execute a Procedure" on page 10-37 for information on granting the necessary privileges to users who will be executing this package.
The ability to use dynamic SQL from within stored procedures generally follows the model of the Oracle Call Interface (OCI). You should refer to the Programmer's Guide to the Oracle Call Interface, Volume I: OCI Concepts for additional information on the concepts presented in this chapter.
PL/SQL differs somewhat from other common programming languages, such as C. For example, addresses (also called pointers) are not user visible in PL/SQL. As a result, there are some differences between the Oracle Call Interface and the DBMS_SQL package. These differences include the following:
A sample usage of the DBMS_SQL package is shown below. For users of the Oracle Call Interfaces, this code should seem fairly straightforward. Each of the functions and procedures used in this example is described later in this chapter.
A more detailed example, which shows how you can use the DBMS_SQL package to build a query statement dynamically, begins on page 14-28. This example does not actually require the use of dynamic SQL, because the text of the statement is known at compile time. However, it illustrates the concepts of this package.
/* The DEMO procedure deletes all of the employees from the EMP
* table whose salaries are greater than the salary that you
* specify when you run DEMO. */
CREATE OR REPLACE PROCEDURE demo(salary IN NUMBER) AS
cursor_name INTEGER;
rows_processed INTEGER;
BEGIN
cursor_name := dbms_sql.open_cursor;
dbms_sql.parse(cursor_name, 'DELETE FROM emp WHERE sal > :x',
dbms_sql);
dbms_sql.bind_variable(cursor_name, ':x', salary);
rows_processed := dbms_sql.execute(cursor_name);
dbms_sql.close_cursor(cursor_name);
EXCEPTION
WHEN OTHERS THEN
dbms_sql.close_cursor(cursor_name);
END;
The typical flow of procedure calls is shown in Figure 14-1. A general explanation of these procedures follows.
Each of these procedures is described in greater detail starting on page 14-7.
To process a SQL statement, you must have an open cursor. When you call the OPEN_CURSOR function, you receive a cursor ID number for the data structure representing a valid cursor maintained by Oracle. These cursors are distinct from cursors defined at the precompiler, OCI, or PL/SQL level, and are used only by the DBMS_SQL package.
Every SQL statement must be parsed by calling the PARSE procedure. Parsing the statement checks the statement's syntax and associates it with the cursor in your program. See Oracle8 Server Concepts: Volume II for an explanation of how SQL statements are parsed.
You can parse any data manipulation language or data definition language statements. Data definition language statements are executed on the parse, which performs the implied commit.
Attention: When parsing a data definition language statement to drop a package or a procedure, a deadlock can occur if a procedure in the package is still in use by you. After a call to a procedure, that procedure is considered to be in use until execution has returned to the user side. Any such deadlock will timeout after five minutes.
Many data manipulation language statements require that data in your program be input to Oracle. When you define a SQL statement that contains input data to be supplied at runtime, you must use placeholders in the SQL statement to mark where data must be supplied.
For each placeholder in the SQL statement, you must call one of the bind procedures, BIND_VARIABLE or BIND_ARRAY, to supply the value of a variable in your program (or the values of an array) to the placeholder. When the SQL statement is subsequently executed, Oracle uses the data that your program has placed in the output and input, or bind, variables.
DBMS_SQL can execute a DML statement multiple times - each time with a different bind variable. The BIND_ARRAY procedure allows you to bind an array of scalars, each value of which will be used as an input variable once per EXECUTE. This is similar to the array interface supported by the OCI.
The columns of the row being selected in a SELECT statement are identified by their relative positions as they appear in the select list, from left to right. For a query, you must call one of the define procedures (DEFINE_COLUMN, DEFINE_COLUMN_LONG, or DEFINE_ARRAY) to specify the variables that are to receive the SELECT values, much the way an INTO clause does for a static query.
You use the DEFINE_COLUMN_LONG procedure to define LONG columns, in the same way that DEFINE_COLUMN is used to define non-LONG columns. You must call DEFINE_COLUMN_LONG before using the COLUMN_VALUE_LONG procedure to fetch from the LONG column.
You use the DEFINE_ARRAY procedure to define a PL/SQL array into which you want to fetch rows in a single SELECT statement. You must call DEFINE_ARRAY before using the COLUMN_VALUE procedure to fetch the rows.
Call the EXECUTE function to execute your SQL statement.
Call the FETCH_ROWS function to retrieve the rows that satisfy the query. Each successive fetch retrieves another row, until the fetch is unable to retrieve anymore rows. Instead of calling EXECUTE and then FETCH_ROWS, you may find it more efficient to call EXECUTE_AND_FETCH if you are calling EXECUTE for a single iteration.
For queries, call COLUMN_VALUE to determine the value of a column retrieved by the FETCH_ROWS call. For anonymous blocks containing calls to PL/SQL procedures, call VARIABLE_VALUE to retrieve the values assigned to the output variables of the PL/SQL procedures when they were executed.
To fetch just part of a LONG database column (which can be up to two gigabytes in size), you use the COLUMN_VALUE_LONG procedure. You can specify the offset (in bytes) into the column value, and the number of bytes to fetch.
When you no longer need a cursor for a session, close the cursor by calling CLOSE_CURSOR. If you are using an Oracle Open Gateway, you may need to close cursors at other times as well. Consult your Oracle Open Gateway documentation for additional information.
If you neglect to close a cursor, the memory used by that cursor remains allocated even though it is no longer needed.
This section describes the security domain for DBMS_SQL procedures when you are using the Oracle Server or Trusted Oracle Server.
Any DBMS_SQL procedures called from an anonymous PL/SQL block are executed using the privileges of the current user. Any DBMS_SQL procedures called from a stored procedure are executed using the privileges of the owner of the stored procedure.
Therefore, if a user creates a procedure and grants EXECUTE privilege on it to a second user, the second user must also be granted privileges explicitly for any DML operations performed in the body of the stored procedure. (These privileges cannot be granted through a role, because roles are disabled inside stored procedures.) Not granting privileges for the DML operations may result in the error message ORA-01031: "insufficient privileges" at runtime.
Any DBMS_SQL procedures called from an anonymous PL/SQL block are executed using the privileges of the current user. Any DBMS_SQL procedures called from a stored procedure are executed using the discretionary access control (DAC) and system privileges of the owner of the stored procedure and the union of the mandatory access control (MAC) privileges granted to the stored procedure and the current user.
Table 14-1 provides a brief description of each of the procedures and functions associated with the DBMS_SQL package, which are described in detail later in this chapter. For examples of how these functions can be used, see "Examples of Using DBMS_SQL" on page 14-28.
Call OPEN_CURSOR to open a new cursor. When you no longer need this cursor, you must close it explicitly by calling CLOSE_CURSOR.
You can use cursors to execute the same SQL statement repeatedly or to execute a new SQL statement. When a cursor is reused, the contents of the corresponding cursor data area are reset when the new SQL statement is parsed. It is never necessary to close and reopen a cursor before reusing it.
The OPEN_CURSOR function returns the cursor ID number of the new cursor. The syntax for this function is:
DBMS_SQL.OPEN_CURSOR RETURN INTEGER;
Call PARSE to parse the given statement in the given cursor. All statements are parsed immediately. This may change in future versions; you should not rely on this behavior.
The parameters for the PARSE procedure are described in Table 14-2. The syntax for this procedure is:
DBMS_SQL.PARSE(
c IN INTEGER,
statement IN VARCHAR2,
language_flag IN INTEGER);
The size limit for parsing SQL statements with the syntax above is size_t, which is the largest possible contiguous allocation on the machine being used. The PARSE procedure also supports the following syntax for large SQL statements:
DBMS_SQL.PARSE(
c IN INTEGER,
statement IN VARCHAR2S,
lb IN INTEGER,
ub IN INTEGER,
lfflg IN BOOLEAN,
language_flag IN INTEGER);
The procedure concatenates elements of a PL/SQL table statement and parses the resulting string. You can use this procedure to parse a statement that is longer than the limit for a single VARCHAR2 variable by splitting up the statement.
To parse SQL statements larger than 32 KB, the DBMS_SQL package makes use of PL/SQL tables to pass a table of strings to the PARSE procedure. These strings are concatenated and then passed on to the Oracle Server.
You can declare a local variable as the VARCHAR2S table-item type, and then use the PARSE procedure to parse a large SQL statement as VARCHAR2S.
The definition of the VARCHAR2S datatype is:
type varchar2s is table of varchar2(256) index by binary_integer;
Call DBMS_SQL.PARSE to parse a large SQL statement in the given cursor. All statements are parsed immediately.
Call BIND_VARIABLE or BIND_ARRAY to bind a given value or set of values to a given variable in a cursor, based on the name of the variable in the statement. If the variable is an IN or IN/OUT variable or an IN array, the given bind value must be valid for the variable or array type. Bind values for OUT variables are ignored.
The bind variables or arrays of a SQL statement are identified by their names. When binding a value to a bind variable or bind array, the string identifying it in the statement must contain a leading colon, as shown in the following example:
SELECT emp_name FROM emp WHERE SAL > :X;
For this example, the corresponding bind call would look similar to
BIND_VARIABLE(cursor_name, ':X', 3500);
The parameters for the BIND_VARIABLE procedures and functions are described in Table 14-3. The syntax for these procedures and functions is shown below. Notice that BIND_VARIABLE is overloaded to accept different datatypes.
DBMS_SQL.BIND_VARIABLE(
c IN INTEGER,
name IN VARCHAR2,
value IN <datatype>);
where <datatype> can be any one of the following types:
NUMBER DATE MLSLABEL VARCHAR2 CHARACTER SET ANY_CS BLOB CLOB CHARACTER SET ANY_CS BFILE
(Chapter 6, "Large Objects (LOBs)" describes the BLOB, CLOB, and BFILE datatypes.)
The following syntax is also supported for BIND_VARIABLE. The square brackets [] indicate an optional parameter for the BIND_VARIABLE function.
DBMS_SQL.BIND_VARIABLE(
c IN INTEGER,
name IN VARCHAR2,
value IN VARCHAR2 CHARACTER SET ANY_CS
[,out_value_size IN INTEGER]);
To bind CHAR, RAW, and ROWID data, you can use the following variations on the syntax:
DBMS_SQL.BIND_VARIABLE_CHAR(
c IN INTEGER,
name IN VARCHAR2,
value IN CHAR CHARACTER SET ANY_CS
[,out_value_size IN INTEGER]);
DBMS_SQL.BIND_VARIABLE_RAW(
c IN INTEGER,
name IN VARCHAR2,
value IN RAW
[,out_value_size IN INTEGER]);
DBMS_SQL.BIND_VARIABLE_ROWID(
c IN INTEGER,
name IN VARCHAR2,
value IN ROWID);
Bulk selects, inserts, updates, and deletes can enhance the performance of applications by bundling many calls into one. The DBMS_SQL package allows you to work on arrays of data using the PL/SQL table type.
Table items are unbounded homogeneous collections. In persistent storage, they are like other relational tables and have no intrinsic ordering. But when a table item is brought into the workspace (either by querying or by navigational access of persistent data), or when it is created as the value of a PL/SQL variable or parameter, its elements are given subscripts that can be used with array-style syntax to get and set the values of elements.
The subscripts of these elements need not be dense, and can be any number including negative numbers. For example, a table item can contain elements at locations -10, 2, and 7 only.
When a table item is moved from transient workspace to persistent storage, the subscripts are not stored; the table item is unordered in persistent storage.
At bind time the table is copied out from the PL/SQL buffers into local DBMS_SQL buffers (the same as for all scalar types) and then the table is manipulated from the local DBMS_SQL buffers. Therefore, if you change the table after the bind call, that change will not affect the way the execute behaves.
You can declare a local variable as one of the following table-item types, which are defined as public types in DBMS_SQL.
type Number_Table is table of Number index by binary_integer; type Varchar2_Table is table of varchar2(2000) index by binary_integer; type Date_Table is table of Date index by binary_integer; type Blob_Table is table of Blob index by binary_integer; type Clob_Table is table of Clob index by binary_integer; type Bfile_Table is table of Bfile index by binary_integer;
The parameters for the BIND_ARRAY procedure to bind an entire array or a range of an array are described in Table 14-4. The syntax for this procedure is shown below. Notice that the BIND_ARRAY procedure is overloaded to accept different datatypes.
DBMS_SQL.BIND_ARRAY(
c IN INTEGER,
name IN VARCHAR2,
<table_variable> IN <datatype>
[,index1 IN INTEGER,
index2 IN INTEGER)] );
where the <table_variable> and its corresponding <datatype> can be any one of the following matching pairs:
<num_tab> Number_Table <vchr2_tab> Varchar2_Table <date_tab> Date_Table <blob_tab> Blob_Table <clob_tab> Clob_Table <bfile_tab> Bfile_Table
For binding a range, the table must contain the elements that specify the range - tab(index1) and tab(index2) - but the range does not have to be dense. Index1 must be less than or equal to index2. All elements between tab(index1) and tab(index2) will be used in the bind.
If you do not specify indexes in the bind call and two different binds in a statement specify tables that contain a different number of elements, the number of elements actually used is the minimum number between all tables. This is also the case if you specify indexes - the minimum range is selected between the two indexes for all tables.
Not all bind variables in a query have to be array binds. Some can be regular binds and the same value will be used for each element of the arrays in expression evaluations (and so forth).
For examples of how to bind arrays, see "Examples 3, 4, and 5: Bulk DML" on page 14-30.
If you are using dynamic SQL to process a query, you must perform the following steps:
This procedure is only used with SELECT cursors. Call DEFINE_COLUMN to define a column to be selected from the given cursor. The column being defined is identified by its relative position in the SELECT list of the statement in the given cursor. The type of the COLUMN value determines the type of the column being defined.
The parameters for the DEFINE_COLUMN procedure are described in Table 14-5. The syntax for this procedure is shown below. Notice that DEFINE_COLUMN is overloaded to accept different datatypes.
DBMS_SQL.DEFINE_COLUMN(
c IN INTEGER,
position IN INTEGER
column IN <datatype>);
where <datatype> can be any one of the following types:
NUMBER DATE MLSLABEL BLOB CLOB CHARACTER SET ANY_CS BFILE
(Chapter 6, "Large Objects (LOBs)" describes BLOB, CLOB, and BFILE.)
The following syntax is also supported for the DEFINE_COLUMN procedure:
DBMS_SQL.DEFINE_COLUMN(
c IN INTEGER,
position IN INTEGER,
column IN VARCHAR2 CHARACTER SET ANY_CS,
column_size IN INTEGER);
To define columns with CHAR, RAW, and ROWID data, you can use the following variations on the procedure syntax:
DBMS_SQL.DEFINE_COLUMN_CHAR(
c IN INTEGER,
position IN INTEGER,
column IN CHAR CHARACTER SET ANY_CS,
column_size IN INTEGER);
DBMS_SQL.DEFINE_COLUMN_RAW(
c IN INTEGER,
position IN INTEGER,
column IN RAW,
column_size IN INTEGER);
DBMS_SQL.DEFINE_COLUMN_ROWID(
c IN INTEGER,
position IN INTEGER,
column IN ROWID);
Call DEFINE_ARRAY to define the appropriate table variable into which you want to fetch rows (with a FETCH_ROWS call). This procedure allows you to do batch fetching of rows from a single SELECT statement. A single fetch call brings over a number of rows into the PL/SQL aggregate object.
When you fetch the rows, they are copied into DBMS_SQL buffers until you execute a COLUMN_VALUE call, at which time the rows are copied into the table that was passed as an argument to the COLUMN_VALUE call.
You can declare a local variable as one of the following table-item types, and then fetch any number of rows into it using DBMS_SQL. (These are the same types as you can specify for the BIND_ARRAY procedure.)
type Number_Table is table of Number index by binary_integer; type Varchar2_Table is table of varchar2(2000) index by binary_integer; type Date_Table is table of Date index by binary_integer; type Blob_Table is table of Blob index by binary_integer; type Clob_Table is table of Clob index by binary_integer; type Bfile_Table is table of Bfile index by binary_integer;
The parameters for the DEFINE_ARRAY procedure are described in Table 14-6. The syntax for this procedure is shown below. Notice that the DEFINE_ARRAY procedure is overloaded to accept different datatypes.
DBMS_SQL.DEFINE_ARRAY(
c IN INTEGER,
position IN INTEGER,
<table_variable> IN <datatype>,
count IN INTEGER,
indx IN INTEGER);
where the <table_variable> and its corresponding <datatype> can be any one of the following matching pairs:
<num_tab> Number_Table <vchr2_tab> Varchar2_Table <date_tab> Date_Table <blob_tab> Blob_Table <clob_tab> Clob_Table <bfile_tab> Bfile_Table
This procedure defines an appropriate "table" for the column at position "position" for cursor "c". The subsequent FETCH_ROWS call will fetch "count" rows. When the COLUMN_VALUE call is made, these rows will be placed in positions indx, indx+1, indx+2, and so on. While there are still rows coming, the user keeps issuing FETCH_ROWS/COLUMN_VALUE calls. The rows keep accumulating in the table specified as an argument in the COLUMN_VALUE call.
The "count" has to be an integer greater than zero, otherwise an exception is raised. The "indx" can be positive, negative, or zero. A query on which a DEFINE_ARRAY call was issued cannot contain array binds.
For examples of how to define arrays, see "Examples 6 and 7: Defining an Array" on page 14-32.
Call this procedure to define a LONG column for a SELECT cursor. The column being defined is identified by its relative position in the SELECT list of the statement for the given cursor. The type of the COLUMN value determines the type of the column being defined.
The parameters of DEFINE_COLUMN_LONG are described in Table 14-7. The syntax is:
DBMS_SQL.DEFINE_COLUMN_LONG(
c IN INTEGER,
position IN INTEGER);
Call EXECUTE to execute a given cursor. This function accepts the ID number of the cursor and returns the number of rows processed. The return value is only valid for INSERT, UPDATE, and DELETE statements; for other types of statements, including DDL, the return value is undefined and should be ignored.
The syntax for the EXECUTE function is:
DBMS_SQL.EXECUTE (
c IN INTEGER)
RETURN INTEGER;
Call EXECUTE_AND_FETCH to execute the given cursor and fetch rows. This function provides the same functionality as calling EXECUTE and then calling FETCH_ROWS. Calling EXECUTE_AND_FETCH instead, however, may cut down on the number of network round-trips when used against a remote database.
The EXECUTE_AND_FETCH function returns the number of rows actually fetched. The parameters for this procedure are described in Table 14-8, and the syntax is shown below.
DBMS_SQL.EXECUTE_AND_FETCH(
c IN INTEGER,
exact IN BOOLEAN DEFAULT FALSE)
RETURN INTEGER;
Call FETCH_ROWS to fetch a row from a given cursor. You can call FETCH_ROWS repeatedly as long as there are rows remaining to be fetched. These rows are retrieved into a buffer, and must be read by calling COLUMN_VALUE, for each column, after each call to FETCH_ROWS.
The FETCH_ROWS function accepts the ID number of the cursor to fetch, and returns the number of rows actually fetched. The syntax for this function is shown below.
DBMS_SQL.FETCH_ROWS(
c IN INTEGER)
RETURN INTEGER;
This procedure returns the value of the cursor element for a given position in a given cursor. This procedure is used to access the data fetched by calling FETCH_ROWS.
The parameters for the COLUMN_VALUE procedure are described in Table 14-9. The syntax for this procedure is shown below. The square brackets [] indicate optional parameters.
DBMS_SQL.COLUMN_VALUE(
c IN INTEGER,
position IN INTEGER,
value OUT <datatype>
[,column_error OUT NUMBER]
[,actual_length OUT INTEGER]);
where <datatype> can be any one of the following types:
NUMBER DATE MLSLABEL VARCHAR2 CHARACTER SET ANY_CS BLOB CLOB CHARACTER SET ANY_CS BFILE
(Chapter 6, "Large Objects (LOBs)" describes the BLOB, CLOB, and BFILE datatypes.)
The following syntax is also supported for the COLUMN_VALUE procedure:
DBMS_SQL.COLUMN_VALUE(
c IN INTEGER,
position IN INTEGER,
<table_variable> IN <datatype>);
where the <table_variable> and its corresponding <datatype> can be any one of these matching pairs:
<num_tab> Number_Table <vchr2_tab> Varchar2_Table <date_tab> Date_Table <blob_tab> Blob_Table <clob_tab> Clob_Table <bfile_tab> Bfile_Table
For columns containing CHAR, RAW, and ROWID data, you can use the following variations on the syntax:
DBMS_SQL.COLUMN_VALUE_CHAR(
c IN INTEGER,
position IN INTEGER,
value OUT CHAR CHARACTER SET ANY_CS
[,column_error OUT NUMBER]
[,actual_length OUT INTEGER]);
DBMS_SQL.COLUMN_VALUE_RAW(
c IN INTEGER,
position IN INTEGER,
value OUT RAW
[,column_error OUT NUMBER]
[,actual_length OUT INTEGER]);
DBMS_SQL.COLUMN_VALUE_ROWID(
c IN INTEGER,
position IN INTEGER,
value OUT ROWID
[,column_error OUT NUMBER]
[,actual_length OUT INTEGER]);
This procedure returns the value of the cursor element for a given position, offset, and size in a given cursor. This procedure is used to access the data fetched by calling FETCH_ROWS.
The parameters of the COLUMN_VALUE_LONG procedure are described in Table 14-10. The syntax of the procedure is:
DBMS_SQL.COLUMN_VALUE_LONG(
c IN INTEGER,
position IN INTEGER,
length IN INTEGER,
offset IN INTEGER,
value OUT VARCHAR2,
value_length OUT INTEGER);
This procedure returns the value of the named variable for a given cursor. It is also used to return the values of bind variables inside PL/SQL blocks.
The parameters for the VARIABLE_VALUE procedure are described in Table 14-11. The syntax for this procedure is:
DBMS_SQL.VARIABLE_VALUE(
c IN INTEGER,
name IN VARCHAR2,
value OUT <datatype>);
where <datatype> can be any one of the following types:
NUMBER DATE MLSLABEL VARCHAR2 CHARACTER SET ANY_CS BLOB CLOB CHARACTER SET ANY_CS BFILE
For variables containing CHAR, RAW, and ROWID data, you can use the following variations on the syntax:
DBMS_SQL.VARIABLE_VALUE_CHAR(
c IN INTEGER,
name IN VARCHAR2,
value OUT CHAR CHARACTER SET ANY_CS);
DBMS_SQL.VARIABLE_VALUE_RAW(
c IN INTEGER,
name IN VARCHAR2,
value OUT RAW);
DBMS_SQL.VARIABLE_VALUE_ROWID(
c IN INTEGER,
name IN VARCHAR2,
value OUT ROWID);
If you are using dynamic SQL to process an INSERT, UPDATE, or DELETE, you must perform the following steps:
The IS_OPEN function returns TRUE if the given cursor is currently open.
The IS_OPEN function accepts the ID number of a cursor, and returns TRUE if the cursor is currently open, or FALSE if it is not. The syntax for this function is:
DBMS_SQL.IS_OPEN(
c IN INTEGER)
RETURN BOOLEAN;
Call this procedure to describe the columns for a cursor opened and parsed through DBMS_SQL.
The DBMS_SQL package declares the DESC_REC record type as follows:
type desc_rec is record (
col_type binary_integer := 0,
col_max_len binary_integer := 0,
col_name varchar2(32) := '',
col_name_len binary_integer := 0,
col_schema_name varchar2(32) := '',
col_schema_name_len binary_integer := 0,
col_precision binary_integer := 0,
col_scale binary_integer := 0,
col_charsetid binary_integer := 0,
col_charsetform binary_integer := 0,
col_null_ok boolean := TRUE);
The parameters of DESC_REC are described in Table 14-12.
The DESC_TAB type is a PL/SQL table of DESC_REC records:
type desc_tab is table of desc_rec index by binary_integer;
You can declare a local variable as the PL/SQL table type DESC_TAB, and then call the DESCRIBE_COLUMNS procedure to fill in the table with the description of each column. All columns are described; you cannot describe a single column.
The parameters of DESCRIBE_COLUMNS are described in Table 14-13. The syntax is:
DBMS_SQL.DESCRIBE_COLUMNS(
c IN INTEGER,
col_cnt OUT INTEGER,
desc_t OUT DESC_TAB);
| Parameter | Mode | Description |
|---|---|---|
c
|
IN |
The ID number of the cursor for the columns being described. |
col_cnt
|
OUT |
|
desc_t
|
OUT |
|
For an example of how to use DESCRIBE_COLUMNS, see "Example 8: Describe Columns" on page 14-34.
Call CLOSE_CURSOR to close a given cursor.
The parameter for the CLOSE_CURSOR procedure is described in Table 14-14. The syntax for this procedure is:
DBMS_SQL.CLOSE_CURSOR(
c IN OUT INTEGER);
There are additional functions in the DBMS_SQL package for obtaining information about the last referenced cursor in the session. The values returned by these functions are only meaningful immediately after a SQL statement is executed. In addition, some error-locating functions are only meaningful after certain DBMS_SQL calls. For example, you call LAST_ERROR_POSITION immediately after a PARSE.
Returns the byte offset in the SQL statement text where the error occurred. The first character in the SQL statement is at position 0.
DBMS_SQL.LAST_ERROR_POSITION RETURN INTEGER;
Call this function after a PARSE call, before any other DBMS_SQL procedures or functions are called.
Returns the cumulative count of the number of rows fetched.
DBMS_SQL.LAST_ROW_COUNT RETURN INTEGER;
Call this function after a FETCH_ROWS or an EXECUTE_AND_FETCH call. If called after an EXECUTE call, the value returned will be zero.
Returns the ROWID of the last row processed.
DBMS_SQL.LAST_ROW_ID RETURN ROWID;
Call this function after a FETCH_ROWS or an EXECUTE_AND_FETCH call.
Returns the SQL function code for the statement. These codes are listed in the Programmer's Guide to the Oracle Call Interface, Volume II: OCI Reference.
DBMS_SQL.LAST_SQL_FUNCTION_CODE RETURN INTEGER;
You should call this function immediately after the SQL statement is executed; otherwise, the return value is undefined.
This section provides example procedures that make use of the DBMS_SQL package.
The following sample procedure is passed a SQL statement, which it then parses and executes:
CREATE OR REPLACE PROCEDURE exec(STRING IN varchar2) AS
cursor_name INTEGER;
ret INTEGER;
BEGIN
cursor_name := DBMS_SQL.OPEN_CURSOR;
--DDL statements are executed by the parse call, which
--performs the implied commit
DBMS_SQL.PARSE(cursor_name, string, DBMS_SQL);
ret := DBMS_SQL.EXECUTE(cursor_name);
DBMS_SQL.CLOSE_CURSOR(cursor_name);
END;
Creating such a procedure allows you to perform the following operations:
For example, after creating this procedure, you could make the following call:
exec('create table acct(c1 integer)');
You could even call this procedure remotely, as shown in the following example. This allows you to perform remote DDL.
exec@hq.com('CREATE TABLE acct(c1 INTEGER)');
The following sample procedure is passed the names of a source and a destination table, and copies the rows from the source table to the destination table. This sample procedure assumes that both the source and destination tables have the following columns:
ID of type NUMBER NAME of type VARCHAR2(30) BIRTHDATE of type DATE
This procedure does not specifically require the use of dynamic SQL; however, it illustrates the concepts of this package.
CREATE OR REPLACE PROCEDURE copy(source IN VARCHAR2,
destination IN VARCHAR2) is
-- This procedure copies rows from a given source table to a
-- given destination table assuming that both source and
-- destination tables have the following columns:
-- - ID of type NUMBER,
-- - NAME of type VARCHAR2(30),
-- - BIRTHDATE of type DATE.
id NUMBER;
name VARCHAR2(30);
birthdate DATE;
source_cursor INTEGER;
destination_cursor INTEGER;
ignore INTEGER;
BEGIN
-- prepare a cursor to select from the source table
source_cursor := dbms_sql.open_cursor;
DBMS_SQL.PARSE(source_cursor,
'SELECT id, name, birthdate FROM ' || source,
DBMS_SQL);
DBMS_SQL.DEFINE_COLUMN(source_cursor, 1, id);
DBMS_SQL.DEFINE_COLUMN(source_cursor, 2, name, 30);
DBMS_SQL.DEFINE_COLUMN(source_cursor, 3, birthdate);
ignore := DBMS_SQL.EXECUTE(source_cursor);
-- prepare a cursor to insert into the destination table
destination_cursor := DBMS_SQL.OPEN_CURSOR;
DBMS_SQL.PARSE(destination_cursor,
'INSERT INTO ' || destination ||
' VALUES (:id, :name, :birthdate)',
DBMS_SQL);
-- fetch a row from the source table and
-- insert it into the destination table
LOOP
IF DBMS_SQL.FETCH_ROWS(source_cursor)>0 THEN
-- get column values of the row
DBMS_SQL.COLUMN_VALUE(source_cursor, 1, id);
DBMS_SQL.COLUMN_VALUE(source_cursor, 2, name);
DBMS_SQL.COLUMN_VALUE(source_cursor, 3, birthdate);
-- bind the row into the cursor that inserts into the
-- destination table
-- You could alter this example to require the use of
-- dynamic SQL by inserting an if condition before the
-- bind.
DBMS_SQL.BIND_VARIABLE(destination_cursor, 'id', id);
DBMS_SQL.BIND_VARIABLE(destination_cursor, 'name', name);
DBMS_SQL.BIND_VARIABLE(destination_cursor, 'birthdate',
birthdate);
ignore := DBMS_SQL.EXECUTE(destination_cursor);
ELSE
-- no more row to copy
EXIT;
END IF;
END LOOP;
-- commit and close all cursors
COMMIT;
DBMS_SQL.CLOSE_CURSOR(source_cursor);
DBMS_SQL.CLOSE_CURSOR(destination_cursor);
EXCEPTION
WHEN OTHERS THEN
IF DBMS_SQL.IS_OPEN(source_cursor) THEN
DBMS_SQL.CLOSE_CURSOR(source_cursor);
END IF;
IF DBMS_SQL.IS_OPEN(destination_cursor) THEN
DBMS_SQL.CLOSE_CURSOR(destination_cursor);
END IF;
RAISE;
END;
This series of examples shows how to use bulk array binds (table items) in the SQL DML statements DELETE, INSERT, and UPDATE.
In a DELETE statement, for example, you could bind in an array in the WHERE clause and have the statement be executed for each element in the array:
declare
stmt varchar2(200);
dept_no_array dbms_sql.Number_Table;
c number;
dummy number;
begin
dept_no_array(1) := 10; dept_no_array(2) := 20;
dept_no_array(3) := 30; dept_no_array(4) := 40;
dept_no_array(5) := 30; dept_no_array(6) := 40;
stmt := 'delete from emp where deptno = :dept_array';
c := dbms_sql.open_cursor;
dbms_sql.parse(c, stmt, dbms_sql.native);
dbms_sql.bind_array(c, ':dept_array', dept_no_array, 1, 4);
dummy := dbms_sql.execute(c);
dbms_sql.close_cursor(c);
exception when others then
if dbms_sql.is_open(c) then
dbms_sql.close_cursor(c);
end if;
raise;
end;
/
In the example above, only elements 1 through 4 will be used as specified by the bind_array call. Each element of the array will potentially delete a large number of employees from the database.
Here is an example of a bulk INSERT statement:
declare
stmt varchar2(200);
empno_array dbms_sql.Number_Table;
empname_array dbms_sql.Varchar2_Table;
c number;
dummy number;
begin
for i in 0..9 loop
empno_array(i) := 1000 + i;
empname_array(I) := get_name(i);
end loop;
stmt := 'insert into emp values(:num_array, :name_array)';
c := dbms_sql.open_cursor;
dbms_sql.parse(c, stmt, dbms_sql.native);
dbms_sql.bind_array(c, ':num_array', empno_array);
dbms_sql.bind_array(c, ':name_array', empname_array);
dummy := dbms_sql.execute(c);
dbms_sql.close_cursor(c);
exception when others then
if dbms_sql.is_open(c) then
dbms_sql.close_cursor(c);
end if;
raise;
end;
/
When the execute takes place, all 10 of the employees are inserted into the table.
Finally, here is an example of an bulk UPDATE statement.
declare
stmt varchar2(200);
emp_no_array dbms_sql.Number_Table;
emp_addr_array dbms_sql.Varchar2_Table;
c number;
dummy number;
begin
for i in 0..9 loop
emp_no_array(i) := 1000 + i;
emp_addr_array(I) := get_new_addr(i);
end loop;
stmt := 'update emp set ename = :name_array
where empno = :num_array';
c := dbms_sql.open_cursor;
dbms_sql.parse(c, stmt, dbms_sql.native);
dbms_sql.bind_array(c, ':num_array', empno_array);
dbms_sql.bind_array(c, ':name_array', empname_array);
dummy := dbms_sql.execute(c);
dbms_sql.close_cursor(c);
exception when others then
if dbms_sql.is_open(c) then
dbms_sql.close_cursor(c);
end if;
raise;
end;
/
Here when the execute call happens, the addresses of all employees are updated in one shot. The two arrays are always stepped in unison. If the WHERE clause returns more than one row, all those employees will get the address the "addr_array" happens to be pointing to at the time.
The following examples show how to use the DEFINE_ARRAY procedure:
declare
c number;
d number;
n_tab dbms_sql.Number_Table;
indx number := -10;
begin
c := dbms_sql.open_cursor;
dbms_sql.parse(c, 'select n from t order by 1', dbms_sql);
dbms_sql.define_array(c, 1, n_tab, 10, indx);
d := dbms_sql.execute(c);
loop
d := dbms_sql.fetch_rows(c);
dbms_sql.column_value(c, 1, n_tab);
exit when d != 10;
end loop;
dbms_sql.close_cursor(c);
exception when others then
if dbms_sql.is_open(c) then
dbms_sql.close_cursor(c);
end if;
raise;
end;
/
Each time the example above does a FETCH_ROWS call, it fetches 10 rows that are kept in DBMS_SQL buffers. When the COLUMN_VALUE call is executed, those rows move into the PL/SQL table specified (in this case n_tab), at positions -10 to -1, as specified in the DEFINE statements. When the second batch is fetched in the loop, the rows go to positions 0 to 9; and so on.
A current index into each array is maintained automatically. This index is initialized to "indx" at EXECUTE and keeps getting updated every time a COLUMN_VALUE call is made. If you re-execute at any point, the current index for each DEFINE is re-initialized to "indx".
In this way the entire result of the query is fetched into the table. When FETCH_ROWS cannot fetch 10 rows, it returns the number of rows actually fetched (if no rows could be fetched it returns zero) and exits the loop.
Here is another example of using the DEFINE_ARRAY procedure:
Consider a table MULTI_TAB defined as:
create table multi_tab (num number,
dat1 date,
var varchar2(24),
dat2 date)
To select everything from this table and move it into four PL/SQL tables, you could use the following simple program:
declare
c number;
d number;
n_tab dbms_sql.Number_Table;
d_tab1 dbms_sql.Date_Table;
v_tab dbms_sql.Varchar2_Table;
d_tab2 dbms_sql.Date_Table;
indx number := 10;
begin
c := dbms_sql.open_cursor;
dbms_sql.parse(c, 'select * from multi_tab order by 1', dbms_sql);
dbms_sql.define_array(c, 1, n_tab, 5, indx);
dbms_sql.define_array(c, 2, d_tab1, 5, indx);
dbms_sql.define_array(c, 3, v_tab, 5, indx);
dbms_sql.define_array(c, 4, d_tab2, 5, indx);
d := dbms_sql.execute(c);
loop
d := dbms_sql.fetch_rows(c);
dbms_sql.column_value(c, 1, n_tab);
dbms_sql.column_value(c, 2, d_tab1);
dbms_sql.column_value(c, 3, v_tab);
dbms_sql.column_value(c, 4, d_tab2);
exit when d != 5;
end loop;
dbms_sql.close_cursor(c);
/*
Here the four tables can be used for anything at all. One usage
might be to use bind_array to move the rows to another table
by using a query such as 'insert into some_t values(:a, :b, :c, :d);
*/
exception when others then
if dbms_sql.is_open(c) then
dbms_sql.close_cursor(c);
end if;
raise;
end;
/
This can be used as a substitute to the SQL*Plus DESCRIBE call by using a SELECT * query on the table that you want to describe.
declare
c number;
d number;
col_cnt integer;
f boolean;
rec_tab dbms_sql.desc_tab;
col_num number;
procedure print_rec(rec in dbms_sql.desc_rec) is
begin
dbms_output.new_line;
dbms_output.put_line('col_type = '
|| rec.col_type);
dbms_output.put_line('col_maxlen = '
|| rec.col_max_len);
dbms_output.put_line('col_name = '
|| rec.col_name);
dbms_output.put_line('col_name_len = '
|| rec.col_name_len);
dbms_output.put_line('col_schema_name = '
|| rec.col_schema_name);
dbms_output.put_line('col_schema_name_len = '
|| rec.col_schema_name_len);
dbms_output.put_line('col_precision = '
|| rec.col_precision);
dbms_output.put_line('col_scale = '
|| rec.col_scale);
dbms_output.put('col_null_ok = ');
if (rec.col_null_ok) then
dbms_output.put_line('true');
else
dbms_output.put_line('false');
end if;
end;
begin
c := dbms_sql.open_cursor;
dbms_sql.parse(c, 'select * from scott.bonus', dbms_sql);
d := dbms_sql.execute(c);
dbms_sql.describe_columns(c, col_cnt, rec_tab);
/*
* Following loop could simply be for j in 1..col_cnt loop.
* Here we are simply illustrating some of the PL/SQL table
* features.
*/
col_num := rec_tab.first;
if (col_num is not null) then
loop
print_rec(rec_tab(col_num));
col_num := rec_tab.next(col_num);
exit when (col_num is null);
end loop;
end if;
dbms_sql.close_cursor(c);
end;
/
|
 Copyright © 1997 Oracle Corporation. All Rights Reserved. |
|