Re: [Q] V7.0 vs. V7.1?

From: David Bath <dtb_at_hera.bf.rmit.edu.au>
Date: 23 Dec 1994 16:37:32 GMT
Message-ID: <3deucc$2co_at_aggedor.rmit.EDU.AU>


singerap_at_powergrid.electriciti.com (Paul Singer) writes:

>>From: bpearc1_at_abacus.tis.tandy.com (Byron Pearce)
>>Subject: [Q] V7.0 vs. V7.1?
>>V7.1 RDBMS?  As I understand it, the only difference is the the 
>>parallel server option and all of the internals, structures, et al.
>>remain the same.
 

>>I would appreciate any information that The following taken from the FAQ I maintain available from

        http://www.bf.rmit.edu.au/~dtb/orafaq/v7_1.html See the general FAQ announcement I made earlier        

                              ORACLE VERSION 7.1
                                       
   
   
   $Date: 1994/12/15 10:43:00 $

   $Revision: 1.5 $
   $State: Exp $
   Revision Log        

   This section covers some of the features announced for version 7.1 of    the Oracle RDBMS.

  • Back To Oracle FAQ Contents
  • ANSI SQL + FIPS ANSI-SQL Compliance Levels + FIPS Compliance Checking + Column AS Alias
  • DBA Issues + New Server Manager + Read-Only Tablespaces + Support for CONNECT INTERNAL + Oracle Names + Parallel Processing + Blank Trimming initialization parameter
  • General Issues + Package for Dynamic SQL + Multiple triggers + Order by column alias + PL/SQL Functions in SQL
   

FIPS ANSI-SQL Compliance Levels        

   Oracle RDBMS 7.1 permits FULL compliance with FIPS SQL-89 and    ENTRY-level compliance with FIPS SQL-92. Some aspects are pushing into    TRANSITION-level compliance.    

   Top of File


   

FIPS Compliance Checking        

   A parameter called FIPS is available as a precompiler construct that    when turned on for a particular Pro*C file will bitch about any    non-ANSI construct.    

   This will permit you to validate your applications for standards, for    example example if you want to migrate an Oracle application to a    non-Oracle RDBMS. (But why would you want to do something that    stupid).    

   As some of the best parts of Oracle (such as DECODE) are non-ANSI,    this is usually going to be annoying when turned on.    

   Top of File


   

Column AS Alias        

   From 7.1 (and later versions of 7.0) you can use the ANSI method of    declaring column aliases, which you will be familiar with if you are    using products such as Sybase or MS-Access. The "old" Oracle method    was as follows:
   select ENAME, SAL / 12 MONTHLY_SALARY from EMP;    

   The ANSI method, which I actually prefer because it makes the parser    in my head work better is as follows:    select ENAME, SAL / 12 as MONTHLY_SALARY from EMP;    

New Server Manager        

   A new all-singing all-dancing GUI replacement for sqldba is being    introduced. Support for sqldba (especially lmode=y is gradually being    withdrawn. However a line mode database administration tool will    remain available, whether or not it is sqldba, the server manager, or    something else.    

   Top of File


   

Read-Only Tablespaces        

   Read-only tablespaces are being introduced. When you make a tablespace    read only, you can improve performance, prevent most writes to the    data (objects can be dropped, of course, because this only requires    writes to the data dictionary in the SYSTEM tablespace) and will    probably eventually support databases on CD-ROM. The control files    note the latest logfile number required for the tablespace.    

   Top of File


   

Support for CONNECT INTERNAL        

   Support for connect internal is being altered. It should work across    networks. Let's hope that none of your machines lie about your network    user id. Guess you should NEVER assign dba privileges to root, except    perhaps during an installation. Too many machines have security holes.    

   Top of File


   

Oracle Names        

   Oracle Names is a new facility being introduced for distributed    environments that allows a distributed name server, in pretty much the    same way as the UNIX yp (yellow pages) works. (Now why do the best    ideas come from the UNIX environment? :-).)    

   You can have more than one name server running in your environment    should one fall over (or, more likely, have the rug pulled out from    under it by a disk crash).    

   Top of File


   

Parallel Processing        

   Utilities such as recovery, export/import and loading will become    parallelized with some architectures. Communication will apparently    occur through the SGA so you'd better make this MUCH bigger.    

   Top of File


   

Blank Trimming initialization parameter        

   A new initialization parameter BLANK_TRIMMING which can take the    values true or false permits assignment of string values to a short    target from a longer source when TRUE. If this parameter is FALSE such    assignments are disallowed.    

   Top of File


   

Package for Dynamic SQL        

   A new DBMS package permitting the generation of SQL on the fly will be    available, probably called DBMS_SQL. I am told this will permit not    only EXEC SQL type statements (such as update and delete) but also    select.    

   Top of File


   

Multiple triggers        

   Whereas with v7 only one trigger of any type can exist on a table,    with 7.1 this restriction is removed. For example, you could have two    post-insert triggers, each updating another audit table. This, in my    opinion, should be AVOIDED. The order in which the triggers fire is    not guaranteed (and why should it?) and I think it will only lead to    confusion. I'd stick to only allowing one trigger of each type in MY    databases, whether Oracle permits more or not, at least of the    user-written variety.    

   The main advantage of this is to permit table replication to be done    easily - and this uses triggers quite a bit.    

   Top of File


   

Order by column alias        

   With 7.1 you can use a column alias as an expression in the order by    clause of an SQL statement. This means you do not have to use the    column number for ordering. (You should still use column number when    you are using set operations such as union - apart from this you are    discouraged from using column number.)    

   An SQL statement using this facility looks like this:    select EMP_ID, (EMP_SAL * EMP_BONUS) as EMP_REMUNERATION    from EMP
   order by EMP_REMUNERATION;    

   Top of File


   

PL/SQL Functions in SQL        

   Database functions written in PL/SQL can be used in SQL statements.    This is a big advantage. This could be particularly useful in    combination with the new dynamic SQL package and DBMS pipes.    Application design and control can be really tightened up. As a    trivial example, you could write a function that decodes a    particularly nasty code. (This could be the encoding of system object    types in the OBJ# internal table, for example). Simply write the    function once, document it and make the doco available to other    application designers and then everyone can easily use it in their    SQL. And when you change the core function (and recompile it)    everything flows through. Bliss.    

   However, I can see some restrictions. If Oracle have been lazy, then    you won't be able to select from within such a function. I would    complain about this. However, if they merely prohibit database updates    from inside such functions, then that is fair enough, and possibly a    good thing. Such "side-effects" could be considered pathological.    

   Special pragmas are available to control the "purity" of such    functions. It uses commands like
   WRITE [NO] DATABASE
   WRITE [NO] PACKAGE            If you declare the pragma as writing to the database, you cannot use    the function in SQL statements. If you declare the pragma as writing    to packages (well, really variables in those packages), then you    cannot use the function in parallel.    

   Top of File


   

Revision Log

   Back to top of file

$Log: v7_1.html,v $

# Revision 1.5  1994/12/15  10:43:00  dtb
# Minor fiddling
#
# Revision 1.4  1994/12/01  00:34:23  dtb
# Editing after responses from Oracle about line mode dba
#
# Revision 1.3  1994/11/29  10:44:06  dtb
# Some reformatting, at versus at-at symbols.
#
# Revision 1.2  1994/11/24  09:46:37  dtb
# Corrections and additions
#
# Revision 1.1  1994/11/23  16:55:38  dtb
# Initial revision
#

   
   Back to top of file

_________________________________________________________________
--- David T. Bath | dtb_at_ftp.bf.rmit.edu.au (mail=dtb_at_bf.rmit.edu.au) 3/175 Wattletree Rd | Snr Tech Consultant, Global Consulting Services MALVERN 3144 VIC | 179 Grattan St, Carlton, Vic 3053, Australia +61 3 500 9337 | Ph:+61 3 3477411 Fx:+61 3 3470182 Mb: +61 15 824171
WWW pages incl. ORA FAQ: lynx http://www.bf.rmit.edu.au/~dtb
--
---
David T. Bath       | dtb_at_ftp.bf.rmit.edu.au (mail=dtb_at_bf.rmit.edu.au)
3/175 Wattletree Rd | Snr Tech Consultant, Global Consulting Services
MALVERN 3144 VIC    | 179 Grattan St, Carlton, Vic 3053, Australia
+61 3 500 9337      | Ph:+61 3 3477411  Fx:+61 3 3470182 Mb: +61 15 824171
WWW pages incl. ORA FAQ: lynx http://www.bf.rmit.edu.au/~dtb
Received on Fri Dec 23 1994 - 17:37:32 CET

Original text of this message