Oracle7 Server Migration Guide Go to Product Documentation Library
Library
Go to books for this product
Product
Go to Contents for this book
Contents
Go to Index
Index



Go to previous file in sequence Go to next file in sequence

Summary of Changes in Oracle7, Release 7.2


This appendix provides an overview of the changes in the Oracle7 Server, Release7.2. The topics included in this appendix are:


Functionality Enhancements

This section contains the following topics:

Cursor Variables

PL/SQL, Release 2.2 provides a new and convenient declarative mechanism to return multi-row result sets from stored procedures. Using a cursor variable, a calling program can obtain a handle that enables the retrieval of multiple rows from a stored procedure. Cursor variables allow you to pass cursors as parameters in your application.

Cursor variables are like C or Pascal pointers, which hold the memory location (address) of some object instead of the object itself, allowing you to allocate memory for the object dynamically.

The cursor variables feature results in reduced coding complexity by providing a structured way to encapsulate queries in a stored procedure.

For more information, see the PL/SQL User's Guide and Reference.

CREATE TABLE...AS SELECT

Table creation and population for CREATE TABLE...AS <subquery> statements can now be performed in parallel. This new functionality is an extension to the parallel query option of Oracle7, Release 7.1. Parallel execution of the CREATE TABLE...AS SELECT operation permits the insertion of the results of a query into a new table in parallel, in addition to executing the query itself in parallel.

The Create Table As Select feature permits the following:

For more information about CREATE TABLE...AS SELECT, see Oracle7 Server SQL Reference and "SQL Changes" [*].

UNRECOVERABLE

A new keyword, UNRECOVERABLE, allows you to avoid redo logging during the creation of tables or indexes (regardless of whether they are created in parallel). Objects created with the UNRECOVERABLE keyword cannot be recovered in the event of media failure, but the creation time will be significantly shorter.

For more information about UNRECOVERABLE, see Oracle7 Server SQL Reference and "SQL Changes" [*].

CREATE INDEX

CREATE INDEX and CREATE TABLE AS SELECT operations can now be performed faster. Release 7.2 offers you the option to disable logging for the CREATE INDEX and CREATE TABLE AS SELECT operations. The new Release 7.2 option is available for parallel and serial execution of the CREATE INDEX and CREATE TABLE AS SELECT operations.

The syntax for the CREATE INDEX command is

CREATE INDEX <index name> ON <tablename> (column name)

For more information about CREATE INDEX, see Oracle7 Server Tuning and Oracle7 Server SQL Reference and "SQL Changes" [*].


Recovery Enhancements

This section contains the following topics:

Checksums

A new feature in Release 7.2 facilitates earlier detection of database corruption. Timely detection of database corruption also aids in diagnosing the causes of corruption.

A new initialization parameter, LOG_BLOCK_CHECKSUM, can be set to write a checksum in the header of every block of the redo log as it is written out of the log buffer. The checksums, if present, are validated when the log is read for archiving or recovery. Similarly, data blocks may also have checksums to detect corruption more easily.

The initialization parameter DB_BLOCK_CHECKSUM enables the checksum option.

For more information, see the Oracle7 Server Administrator's Guide.

Media Recovery

A new mechanism avoids performing media recovery if a database crash occurs while one or more files is in online backup mode.

For more information, see the Oracle7 Server Administrator's Guide.


Security Enhancements

This section contains the following topics:

Network Security Enhancements

The Oracle7 external authentication mechanism now extends to include support for network authentication services. Distributed installations can now validate database access and roles by taking advantage of network authentication services.

Specific network security features are

For more information, see Oracle7 Server Concepts.

Trusted Stored Procedures

The release 7.2, Trusted Stored Procedures enables application developers to

For more information, see the Trusted Oracle7 Server Administrator's Guide.

The PL/SQL Wrapper feature consists of two components:

PL/SQL Wrapper offers the same portability and flexibility as the PL/SQL source code format. In particular, PL/SQL Wrapper provides

For more information, see the PL/SQL User's Guide and Reference.

Secure Connections with Encrypted Passwords

To better protect the confidentiality of your password, Oracle7 releases, 7.1 and higher, can be configured to use encrypted passwords for client/server and server/server connections.

You can require that the password used to verify a connection always be encrypted by setting the following values.

If enabled at both the client and server, passwords will not be sent across the network "in the clear", but will be encrypted using a modified DES (Data Encryption Standard) algorithm.

The DBLINK_ENCRYPT_LOGIN parameter is used for connections between two Oracle servers (for example, when performing distributed queries). If you are connecting from a client, Oracle checks the ORA_ENCRYPT_LOGIN environment variable.

Oracle releases before release 7.1 are not designed to accept encrypted passwords. If you know that you will attempt connections from a release 7.1 server to earlier versions of the Oracle Server (for example, if you are using a database link for distributed queries) you must set DBLINK_ENCRYPT_LOGIN to FALSE for the connection to succeed. If you will be connecting from a new client to an older version of the server (for example, if you run Server Manager against an earlier version of Oracle) you must set ORA_ENCRYPT_LOGIN to FALSE.

For more information, see Oracle7 Server Reference and the Oracle7 Server Administrator's Guide.


Performance Enhancements

This section contains the following topics:

Operational Efficiency and Oracle7 Parallel Query Features

Oracle7 offers new parallel features with the parallel query option. These new features include the following:

For best performance, the Oracle Corporation recommends that your system meet the following requirements:

For best performance with SQL*Loader direct parallel load, place database files to be loaded on a separate disk or on striped volumes/file systems. If you are not implementing disk striping, the Oracle Corporation recommends that you only start as many direct parallel load sessions as you have datafiles. For striped disk implementation, start as many direct parallel load sessions as possible, but without causing CPU bottlenecks or I/O saturation.

For best performance with CREATE INDEX ... PARALLEL operations, the previous suggestions also apply. For some SMP systems, the degree of parallelism can be twice the number of CPUs, provided your volume or file system are on separate drives or are striped.

For a full discussion on all of the Oracle7 parallel query options, see Oracle7 Server Tuning.

Hash Clusters

Release 7.2 provides you with improved control over the creation of hash clusters. You can now specify an application-specific SQL expression, with some restrictions, as the hash function for a cluster. Choice of appropriate hash functions reduces collisions, resulting in better performance.

For more information, see Oracle7 Server Concepts.

Loadable Character Sets (NLS)

Release 7.2 loads character sets upon first reference. Instead of linking all character sets as static data, each character set is read into dynamic memory upon first reference. The size of the executable is thus reduced by eliminating character set data not in use during execution.

There are three new utility programs that facilitate the use of the NLS loadable character set. They are the NLS Data Installation Utility, the NLS Configuration Utility, and the NLS Calendar Utility.

The NLS Data Installation Utility

A default set of NLS objects is included with the NLS loadable character set you have purchased. At some future time, you may receive updates to the data in the form of text files, or you may create your own text-form object files. Such files must be converted into binary format and merged into the existing NLS object set. The NLS Installation Utility allows you to perform the necessary conversion to binary format.

The NLS Configuration Utility

Along with the binary object files, a boot file is also generated by the Data Object Installer. The boot file is used by NLS modules to identify and locate all the NLS objects that need to be loaded. There are three types of boot files:

The NLS Configuration Utility allows you to configure your boot files such that NLS loads only the NLS objects that you require.

NLS Calendar Utility

NLS provides support for a number of calendars in addition to the Gregorian calendar. All supported calendars are defined with data linked directly into the NLS. However, some of the calendars may require the addition of ruler eras (in the case of imperial calendars) or deviation days (in the case of lunar calendars) in the future. The NLS Calendar Utility enables you to define additional eras or deviation days in an external file without waiting for a new NLS release. The external file will be automatically loaded by NLS when calendar functions are executed.

For more information about the NLS, see Oracle7 Server Reference.

National Language Support Enhancements

Release 7.2 supports certain national language parameters as environment variables that can be altered by issuing appropriate operating system commands. Greater flexibility for multi-lingual applications is, thereby, provided by allowing more granular specification of NLS parameters. The environment variables include

NLS_DATE_FORMAT NLS_CREDIT
NLS_DATE_LANGUAGE NLS_MONETARY_CHARACTERS
NLS_SORT NLS_CALENDAR
NLS_LIST_SEPARATOR NLS_DISPLAY
NLS_DEBIT
The UNICODE encoding scheme, UTF2, a variable-width, multi-byte format, is supported with Oracle7, Release 7.2 to support multi-byte character sets.

Release 7.2 supports five additional calendar systems: Japanese Imperial, ROC (Republic of China) Official, Thai Buddha, Arabic Hijrah, and the Arabic/Hebrew display character set.

For more information about NLS environment variables, see Oracle7 Server Reference.

Application Registration

A new package, DBMS_APPLICATION_INFO, provides a mechanism for registering, with the database, the name and activity of the application currently running.

For more information, see Oracle7 Server Tuning.

Index Fixed Tables

Certain V$ views now have a fast access method that can greatly improve query performance on these views. In previous versions, queries of V$SESSION, V$SESSTAT, and V$SQLAREA could cause performance bottlenecks due to very large sort and scan operations. With Release 7.2, queries using the numeric columns of these tables as the search key can perform significantly faster due to an "index" on the fixed tables' numeric columns.

You can determine which columns have had indexes created on them by using the V$INDEXED_FIXED_COLUMN view. For more information, see Oracle7 Server Reference.

Parallel Server Enhancements

The Oracle7 Parallel Server has new performance enhancements. Descriptions of the functionality of the new features follow.

Space Management

Batch Allocate at Insert: Batch allocation of all of the blocks covered by a PCM (Parallel Cache Management) lock during insert eliminates the need to pre-allocate space for free list groups. Two situations can be significantly affected:

Freelist Group Size Determination: Release 7.2 allows you to determine how much of the space you have pre-allocated to a freelist has been used. The allocation of more space than is required is, therefore, prevented. A new package, DBMS_SPACE (in DBMSUTIL.SQL), returns the number of blocks that a table has for each freelist group.

Indexes Support Freelist Groups: Indexes now support freelist groups, but extents cannot be pre-allocated to the freelist groups.

Choosing a Freelist Group for Inserts: A new ALTER SESSION command allows a session to override the value of the INSTANCE parameter.

Locking Issues

Table Locks: A new command, ALTER TABLE <table_name> DISABLE TABLE LOCK, disables all table locks on a per-table basis. A corresponding ENABLE command re-enables table locking. The command must wait for all transactions that are active in the system to COMMIT or ROLLBACK.

Lock Recovery: Pre-Release 7.2 procedures recovered locks after an instance crash, one by one and in serial order, using the SMON process. Locks are now recovered by the lock processes after an instance crashes, which is a much faster process than the one used in pre-Release 7.2 databases. Also, multiple lock processes can operate in parallel.

New and Modified Views and Dynamic Performance Views: The following views and dynamic performance views were modified or added to allow better monitoring of PCM (Parallel Cache Monitoring).

FILE_LOCK V$FALSE_PING
FILE_PING V$LOCK_ELEMENT
V$BH V$LOCK_WITH_COLLISIONS
V$CACHE V$PING
V$_CACH_ACTIVITY

Ping Efficiency

Early Release of Lock on Block: During a ping, the lock on a block is now released once the block has been written; this feature is now implemented in the DBWR. The buffer is available for reuse immediately after the write of the block completes, instead of waiting for the write of the entire batch of blocks to complete.

Ping Fairness: The number of changes that can be made, per second, to very hot blocks is now greatly increased. Previously, only one access could be made to a block in an instance before the block pinged out. Now, multiple accesses can be made to a block before it is pinged out.

Tuning Tools for the Parallel Server

Ping Per File Statistics: There is now a generic implementation of the ping per file statistics.

Other Caches: Release 7.2 provides statistics on pings in the library and dictionary cache.

Fixed Table Change: The lock element pointer is now added to V$BH (the buffer cache header fixed table). You may now determine which locks cover which blocks. You may then determine when false pinging occurs and which blocks are being falsely pinged.

For more information on Oracle7 Parallel Server enhancements, see Oracle7 Parallel Server Concepts & Administration.


Administration Enhancements

This section contains the following topics:

Resizeable Datafiles

The resizeable datafiles feature allows datafiles to automatically increase in size when more space is needed. The database administrator is thus provided with additional flexibility in the design, operation, and management of datafiles.

Benefits of the resizeable datafiles feature are

The Export Utility and Resizeable Datafiles

The Export Utility does not export FILEXT$ and does not preserve any information about the automatic extension feature of Resizeable Datafiles. Since FILE$ is exported, the exported size of the datafile reflects the last datafile resize operation. After the datafiles have been imported, the database administrator (DBA) must issue SQL commands to turn on the auto-extension feature.

For more information, see the Oracle7 Server Administrator's Guide and Oracle7 Server SQL Reference.

Job Queues

Release 7.2 provides a new feature, job queues, that enables you to schedule periodic execution of PL/SQL code. For example, job queues are used by Oracle's automatic snapshot refresh.

A new package, DBMS_JOB, can schedule execution of stored PL/SQL procedures and programs written with the Oracle Call Interface, Oracle Precompilers, SQL*Module, or within a stored PL/SQL subprogram.

The DBMS_JOB package allows you to submit, change, and remove jobs in a queue. Job queues are used by the symmetric application facility to refresh snapshots automatically and to schedule communication between master sites. You can use job queues to schedule administrative procedures to be performed at periodic intervals.

For more information, see the Oracle7 Server Administrator's Guide.

Displaying Space Information

This new Release 7.2 feature enables you to determine the precise manner in which extents are actually used, from a space management perspective. The additional space management information is provided through a set of PL/SQL procedures.

A new PL/SQL package, called DBMS_SPACE, contains the following two procedures:

The DBMS_SPACE package is declared in the DBMSUTIL script that is installed when all other Release 7.2 utility packages are installed. Therefore, both upgrade and downgrade operations (in other words, forward and backward migrations, respectively) are accomplished by running the DBMSUTIL script.

Authorization required to use the DBM_SPACE package is the same as required to perform an ANALYZE command. The caller's authorization performs the security checks.

A complete set of definitions for all of the options in the UNUSED_SPACE and FREE_BLOCKS procedures is found in the Oracle7 Server Administrator's Guide.

Network Authentication

Sites migrating OS roles to their network service must remember to use the global database name instead of the SID when assigning names. However, no change to the database-defined roles is necessary.

Sites that wish to move OS authenticated users to their network service must use the same external name. For example, if a database user, SCOTT, is defined as IDENTIFIED EXTERNALLY and the operating system username is SCOTT, the network username cannot be defined as JANE; it must remain SCOTT to match the database username that already exists. If sites are creating new database accounts, the network identity need not match the operating system identity. However, the database username must correspond to the network identity for network authentication to take place.

For more information, see Oracle7 Server Utilities and the Oracle7 Server Administrator's Guide.

Activate SQL_TRACE for Remote Sessions

A portable mechanism now enables a database administrator to activate SQL_TRACE for a session other than her/his own.

For more information, see Oracle7 Server Tuning.


Oracle7, Release 7.2 Datatype Changes

The following table gives the rules for input and output host variables of different Oracle 7.2 datatypes that are national language variables:

Oracle 7.2 External Datatypes On Input On Output
ANSI Fixed CHAR (External datatype 96: Pro*COBOL-mode = ANSI Input host variables are stripped of trailing double-byte spaces.? Pro*COBOL variables that are declared as PIC N take this datatype with mode=ANSI. Output host variables are blank padded with double-byte spaces Pro*COBOL variables that are declared as PIC N take this datatype with mode=ANSI.
CHARZ (External datatype 97: Pro*C-DBMS = V7.2 Input host variables are stripped of trailing double-byte spaces and must contain a null terminator at the end of the data.? Pro*C CHAR variables take this datatype if the variable is typed to be a NLS variable (using the NLS_CHAR option). Output host variables are double-byte blank padded and null terminated at the end of the buffer. Pro*C/C++ CHAR variables take this datatype if the variable is typed to be a NLS variable (using the NLS_CHAR option).
VARCHAR (External datatype 9) Input host variables are not stripped of double-byte spaces. The length component is assumed to be the length of the data in characters, not bytes. Output host variables are not padded at all. The length of the buffer is set to the length in characters, not bytes.
? If the input national language host variable contains only spaces, one double byte will be left in the buffer to act as a sentinel.


SQL Changes

This section contains the following topics:

ALTER DATABASE CLEAR

A new SQL command, ALTER DATABASE CLEAR, can reinitialize and optionally not archive an online log. The new command is the equivalent of a drop and add of the log file, except that it is allowed even if there are only two logs for the thread. The command may also be done to the current log of a closed thread. The syntax of the new command is

ALTER DATABASE CLEAR [UNARCHIVED] LOGFILE <log_name> [<log_name>,...] [UNRECOVERABLE DATAFILE]

ALTER DATABASE DATAFILE datafile END BACKUP

It is now unnecessary to do complete media recovery when a database was shutdown without finishing an online backup with the ALTER DATABASE...END BACKUP command.

The following command takes the named datafile(s) out of backup mode:

ALTER DATABASE DATAFILE `<name>' [`<name>'...] END BACKUP

ALTER ROLLBACK SEGMENT SHRINK

It is now possible to shrink a rollback segment to an optimum size using the following command:

ALTER ROLLBACK SEGMENT name SHRINK TO size ;

ALTER SESSION SET INSTANCE

In a parallel server environment while connected to one instance it is now possible to mimic that the session is connected to another instance. For example:

ALTER SESSION SET INSTANCE = 3;

ALTER TABLE

It is now possible to allow or disallow users to use a table lock using the following commands:

ALTER TABLE table_name DISABLE TABLE LOCK;
ALTER TABLE table_name ENABLE TABLE LOCK;

This feature is of most use in a parallel server environment where a table lock can affect system performance.

CREATE CLUSTER ... HASH IS

The HASH IS option in the CREATE CLUSTER command is changed. The old HASH option specified a column as the hash function for a hash cluster. The new HASH option specifies a SQL expression as the hash function for a hash cluster. It is now possible to use your own PL/SQL functions to calculate the hash key. For example:

CREATE CLUSTER cloudy (deptno number(2))
	HASHKEY 20 HASH IS my_hash(deptno);

CREATE CLUSTER ... PARALLEL

It is now possible to use parallelism when creating a cluster and when querying the table after creation. For example:

CREATE CLUSTER cluz (empno number(6))
	PARALLEL (DEGREE 3);

CREATE TABLE ... PARALLEL

It is now possible to use parallelism when creating a table and when querying the table after creation. For example:

CREATE TABLE tmp_emp
	PARALLEL (DEGREE 3)
	AS SELECT * FROM emp WHERE deptno = 10;

CREATE TABLE UNRECOVERABLE

Release 7.2 provides support for unrecoverable logging during CREATE INDEX and CREATE TABLE AS SELECT operations. If you specify UNRECOVERABLE, log records are not generated for the operation. If you specify RECOVERABLE, log records are generated so that the newly created object can be recovered from database backups taken before the object was created. The syntax for the CREATE TABLE AS SELECT and CREATE INDEX commands is

CREATE TABLE `<name>' [`<name>'...] AS SELECT UNRECOVERABLE
CREATE INDEX `<index_name>'
	ON `<table_name>'(column_name)
	UNRECOVERABLE

CREATE INDEX ... PARALLEL

It is now possible to use parallelism when creating an index. For example:

CREATE INDEX idx ON emp(ename)
	PARALLEL (DEGREE 3);

expr

It is now possible to use a user defined PL/SQL function in the same manner as a SQL expression. For example:

SELECT my_fun(ename) FROM emp;

INSERT INTO subquery

It is now possible to use a subquery in the INTO clause of an insert statement similar to how views are used. For example:

INSERT INTO (SELECT * FROM dept)
	VALUES (50, 'DEVELOPMENT', 'BELMONT');

TO_CHAR

A number format model using '9's now returns a zero for the value zero. For example:

SELECT TO_CHAR(0,'999') num FROM DUAL;

NUM
----
   0

UPDATE subquery

It is now possible to use a subquery in an updated statement similar to how views are used. For example:

UPDATE (SELECT * FROM dept)
	SET deptno = 50
	WHERE deptno = 60

Subquery in FROM Clause

Release 7.2 offers the capability to use a subquery in the FROM clause of a SQL query. Such subqueries provide an "immediate view", specified directly in the query. The need to write relatively more complex queries that would otherwise require a view is thereby eliminated. The syntax is

SELECT <column_list>
FROM <table or subquery> [alias] [<table or subquery> [alias]]

For more information on new Release 7.2 SQL commands, see Oracle7 Server SQL Reference and the Oracle7 Server Administrator's Guide.


New Features of the Oracle Precompilers, Release 1.7

This section contains the following topics:

Release 1.7 of the Oracle Precompilers is introduced with Oracle7, Release 7.2 and incorporates many new features designed to enhance precompiler functionality and application performance, as well as allow your applications to access new Oracle7, Release 7.2 capabilities. Most users will upgrade to Release 1.7 at the same time they upgrade to Oracle7, Release 7.2. Use of Releases 1.3, 1.4 1.5, and 1.6 of the Oracle Precompilers is supported against Oracle7, Release 7.2. However, if you wish to use the enhanced functionality of Oracle7, Release 7.2, you must use Release 1.7 of Pro*COBOL, Pro*FORTRAN, and Pro*ADA.

For more information, see the Programmer's Guide to the Oracle Pro*C/C++ Precompiler, the Pro*COBOL Supplement to the Oracle Precompilers Guide, and the SQL*Module User's Guide and Reference.

The following new features are introduced with the Oracle Precompilers, Release 1.7.

Cursor Variable Support

The new cursor variable feature in PL/SQL, Release 2.2 is fully supported in Pro*C, Release 2.1. PL/SQL,

VARCHAR Recognition, C structs, and COBOL Group Items

Pro*C/C++, Release 2.1 and Pro*COBOL, Release 1.7, with the VARCHAR command-line option enabled, recognize two unique data structures as VARCHAR host variables. These data structures consist of a length field and a string field, as shown in the following examples.

The following C struct is recognized by the Pro*C precompiler as a VARCHAR host variable:

struct {
        short     len;
        char      arr[length];
}  data_name

where data_name is a VARCHAR host variable and length is expressed in single-byte characters. The variable-length string is then referenced as data_name and the length field (len) is referenced as data_name.len.

In addition to single-byte character data, the Pro*COBOL precompiler also supports group items with string fields declared as multi-byte NLS datatypes (PIV N). The following COBOL group item is recognized by the Pro*COBOL precompiler as a VARCHAR host variable:

WORKING-STORAGE SECTION.
...
01     data_name.
       49 data_name-LEN PIC S9(4) {COMP|COMP-5}.
       49 data_name-ARR PIC {X(length)|N(length)}.

where data_name is a VARCHAR host variable and length is expressed in single-byte (X) or double-byte (N) characters. The variable-length string is then referenced as data_name and the field length (len) is referenced as data_name-LEN. "COMP" and "COMP-5" are used to declare integer datatypes.

The Oracle Precompiler Pro*C/C++, Release 2.1 offers enhanced functionality, performance, level of standards compliance for embedded SQL application development.


Pro*C/C++, Release 2.1 New Features

This section contains the following topics:

C++ Support and Embedded SQL Programming

Pro*C/C++, Release 2.1, which is available with the Oracle7 Server, Release 7.2, permits SQL and PL/SQL blocks to be embedded in C++ programs. A new PARSE command line option is available in Pro*C/C++, Release 2.1. Parse functionality permits you to decide how much of a program, outside the EXEC SQL DECLARE SECTION, is parsed. You may choose between parsing the entire program, parsing preprocessor directives and C declarations within an EXEC SQL DECLARE SECTION, or not parsing at all.

The extension to Pro*C/C++ has nothing to do with C++ intrinsically, but rather with grammatical extensions to C. An example of such an extension is that the Microsoft C compiler allows declarations to be amended with the (new) keywords near and far. Pro*C/C++, Release 2.1 has mechanisms that allow porters to handle special cases, but for cases not handled, you could invoke the Pro*C/C++ option as a workaround.

For more information, see the Programmer's Guide to the Oracle Pro*C/C++ Precompiler.

PL/SQL, Release 2.2 provide a new and convenient declarative mechanism to return multi-row result sets from stored procedures. Cursors can be used as regular (bind) variables, passed as parameters in functions and procedures, and returned from functions. This flexible and powerful solution allows you to fetch from a cursor at the client side, while the SQL statement belonging to the cursor is defined in a stored procedure.

For more information, see the PL/SQL User's Guide and Reference.

Pro*C/C++, Release 2.1 and Multi Byte NLS Character Data

In Pro*C, Release 2.1, your applications can use national language characters according to the Nippon Telegraph and Telephone Corporation (NTT) Multivendor Integration Architecture (MIA) specifications. (See "Standards Compliance" [*].)

The NTT MIA enhancement requires runtime support in SQLLIB. All applications that previously used Pro*COBOL 1.7 or Pro*C/C++ 2.1 need to link against the current version of SQLLIB.

For more information, see the Programmer's Guide to the Oracle Pro*C/C++ Precompiler, the Pro*COBOL Supplement to the Oracle Precompilers Guide, and the SQL*Module User's Guide and Reference.


SQL*Module Default WITH INTERFACE Clause

SQL*Module can now generate a default WITH INTERFACE clause for client-side remote procedure calls (RPC) to PL/SQL stored procedures.

The uses of the default WITH INTERFACE clause are as follows:

For more information, see the SQL*Module User's Guide and Reference.


Non-blocking Oracle Call Interface (OCI)

Release 7.2 enhances responsiveness of client/server applications by overcoming the limitations of synchronous, Oracle Call Interface (OCI) calls. Non-blocking OCI calls are now possible.

A new mechanism allows an OCI call to yield control back to the application if the call is blocked. Applications may now respond to events such as mouse clicks or timers while a call is being processed by Oracle, resulting in improved responsiveness.

Non-blocking calls are especially useful in client/server environments in which events from multiple sources, not generated in predictable order, must be handled.

Client side applications that existed before migration to Release 7.2 need to be modified if the non-blocking feature is to be used with the new Release 7.2 database.

For more information on non-blocking OCI, see the Programmer's Guide to the Oracle Call Interface.


Initialization Parameter Changes

The following table lists the initialization parameters that are obsolete and new in Release 7.2. See Oracle7 Server Reference for a complete description of all initialization parameters.

Obsolete Parameters
LOG_ALLOCATION INIT_SQL_FILES
New Parameters
CPU_COUNT SD_ALL
DB_BLOCK_CHECKSUM SD_INHIBIT
LOG_BLOCK_CHECKSUM SORT_DIRECT_WRITES
If you use Version 6 National Language Support, you must alter your Oracle7 parameter file to include the appropriate initialization parameters. Additionally, the specification of the NLS_SORT parameter has changed. See Oracle7 Server Reference for a complete description of National Language Support.


Data Dictionary Changes

This section contains the following topics:

This section describes the changes to the Oracle7, Release 7.2 data dictionary. See Oracle7 Server Reference for descriptions of all standard tables and views that are available to developers.

Views

Oracle7, Release 7.2 introduces changes and additions to data dictionary views. Some pre-Release 7.2 views have been changed. New views have been introduced. The following tables summarize the changes and additions.

Changed Views
ALL_INDEXES LOADER_CONSTRAINT_INFO
ALL_TABLES USER_CLUSTERS
DBA_CLUSTERS USER_INDEXES
DBA_INDEXES USER_TABLES
DBA_TABLES
New Views
ALL_CLUSTERS
ALL_CLUSTERS_HASH_EXPRESSIONS
DBA_CLUSTER_HASH_EXPRESSIONS
FILE_LOCK
FILE_PING
USER_CLUSTER_HASH_EXPRESSIONS

Dynamic Performance Tables

This section lists the changes and additions to the dynamic performance tables.

Obsolete Views
V$BACKUP V$XATRANS$
V$PENDING_XATRANS$
Changed Views
V$LATCHHOLDER V$SESSION
V$OPEN_CURSOR V$SQLAREA
V$FIXED_TABLE
New Views
V$BACKUP_STATE V$INSTANCE
V$BH V$LOCK_ACTIVITY
V$CACHE V$LOCK_ELEMENT
V$CACHE_LOCK V$LOCK_WITH_COLLISIONS
V$PING V$OBJECT_DEPENDENCY
V$DB_PIPES V$PWFILE_USERS
V$EVENT_NAME V$SQL
V$FALSE_PING V$SQLTEXT
V$FIXED_VIEW_DEFINITION V$SQLTEXT_WITH_NEWLINES
V$INDEXED_FIXED_COLUMN FILEXT$

Trusted Oracle7 Views

The following views and dynamic tables are now in Trusted Oracle7:

New Views
PROCEDURE_MAC_PRIVS


Standards Compliance

Multi-byte NLS

Pro*C/C++, Release 2.1 and Pro*COBOL, Release 1.7 are now compatible with some aspects of the Multi-vendor Integration Architecture (MIA) specification developed by the Nippon Telegraph and Telephone Corporation (NTT), Japan. MIA is an attempt by NTT to achieve greater SQL compatibility among its vendors in the Japanese market. Developers can create double-byte National Language Character (NLC) character applications using NLC literals and variables.

For more information, see the Programmer's Guide to the Oracle Pro*C/C++ Precompiler, the Pro*COBOL Supplement to the Oracle Precompilers Guide, and the SQL*Module User's Guide and Reference.


Release 7.2 Backward Compatibility

Release 7.2 is completely backward compatible with Version 6, and Releases 7.0 and 7.1 of the Oracle7 Server. As a maintenance release, Release 7.2 is the normal maintenance path for users running 7.0 and 7.1 applications.

Warning: If the SQL*DBA COMPATIBILITY parameter is set to Version 6, the functionality of certain Release 7.2 features will not be available. For example, if COMPATIBLE is set to Version 6, the only constraint that will be enforced automatically is NOT NULL. The PRIMARY KEY constraint will not be enforced automatically unless it is indexed. All other constraints will not be enforced.

For more information about backward compatibility and downgrading, see Chapter 9, "Upgrading and Downgrading between Oracle7 Releases".


New, Changed, and Obsolete SQL Scripts

New scripts are added and existing scripts are modified or made obsolete to support the extended functionality of Oracle7 Release 7.2. The new and changed scripts are listed below.

By convention, scripts that begin with CAT are intended to be run by SYS. They create views and tables in the data dictionary. Scripts that begin with DBMS are also run by SYS. They create stored procedures, functions, and packages. Scripts that begin with UTL are intended to be run by individual users. For more information, see Oracle7 Server Reference.

Obsolete Scripts
UTLOPCRX.SQL UTLOPEXP.SQL
UTLOPDRX.SQL UTLOPIMP.SQL
UTLOPENV.SQL UTLOPTAB.SQL
Changed Scripts
DBMSUTIL.SQL| TKPQCON*.SQL
TKPQSUIT.TSC TKPQKPMM.TSC
TKPQSPRF.TSC TKPQABD.SQL
TKTTAB.TSC
| The DBMS_SPACE package was added to this script.
New Scripts
CAT70103.SQL SAMPLE1.SQL
CAT7201.SQL SAMPLE2.SQL
CAT7202.SQL SAMPLE3.SQL
DBMSPROBE.SQL SAMPLE4.SQL
PBREAK.SQL TKTCTAS.SQL
PBRPH.SQL TKPQCTAS.TAS
PBUTL.SQL


Other Changes

XA Library

The Oracle XA library is included with Release 7.2. The Oracle XA library is an external interface that allows global transactions to be coordinated by a transaction manager other than Oracle7. This allows inclusion of non-Oracle entities called resource managers (RM) in distributed, two-phase commit transactions. The Oracle XA library conforms to the X/Open Distributed Transaction Processing (DTP) software architecture's XA interface specification. For more information, see Oracle7 Server Distributed Systems, Volume I.




Go to previous file in sequence Go to next file in sequence
Prev Next
Oracle
Copyright © 1996 Oracle Corporation.
All Rights Reserved.
Go to Product Documentation Library
Library
Go to books for this product
Product
Go to Contents for this book
Contents
Go to Index
Index