Sybase NNTP forums - End Of Life (EOL)

The NNTP forums from Sybase - forums.sybase.com - are now closed.

All new questions should be directed to the appropriate forum at the SAP Community Network (SCN).

Individual products have links to the respective forums on SCN, or you can go to SCN and search for your product in the search box (upper right corner) to find your specific developer center.

wish list

7 posts in Product Futures Discussion Last posting was on 2002-04-03 21:44:27.0Z
Iwish Posted on 2002-04-03 07:26:47.0Z
From: "Iwish" <iwish@example.com>
Subject: wish list
Date: Wed, 3 Apr 2002 16:26:47 +0900
Lines: 63
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 5.50.4522.1200
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4522.1200
Message-ID: <Q0nAfLu2BHA.216@forums.sybase.com>
Newsgroups: sybase.public.ase.product_futures_discussion
NNTP-Posting-Host: 158.77.52.194
Path: forums-1-dub!forums-master.sybase.com!forums.sybase.com
Xref: forums-1-dub sybase.public.ase.product_futures_discussion:571
Article PK: 94097

Hi

My wish lists.............

[ASE]
- clustering (like a Oracle Parallel Server)
- Very Large Database support(Over 1TB)
- partitioning(range, hash, key)
- job scheduler in ASE(like a crond of UNIX)
- transaction without logging
- notify option(like a recovery percent at startup)
- dbcc performance enhancement
- resizable database size(can be shrunken)
- rename object(user table, user stored procedure, database)
- vdevno of device(automatic increase)
- alias(object, database) <- not a user alias
- temporary set option(not a session based option but a query based option)

[T-SQL]
- set rowcount 100 skip 90 (can select a portion of data)
- inline view(derived table)
- rownum function (or number(*) in asa)
- user defined function(T-sql)
- rank,replace function
- simple encryption/decryption function
- date type(not a datetime)
- variable binding in dynamic sql
- cursor Behavior(forward, backward, jump) & improve cursor performance
- global nextval(not a identity, in memory, no physical i/o)
- column alias can be used in having, group by

[bcp]
- conditional bcp out(not using a view)

[isql]
- reexecution (ie '/' of ORACLE)
- formatting prompt (ie ase12:master 1> )

[Sybase Central]
- can view data(like a powerbuilder).
- clean & neat design
- support win32 version continuously

[PC Client]
- sql advantage
- need a basic OLAP tool

[Backup/Restore]
- backup/restore a single object

[ESQLC]
- array fetch to C Struct Type(ESQLC)
- array insert

[3rd party support]
- SAP support
- Tivoli support


Thanks


Marc Zampetti Posted on 2002-04-03 16:16:08.0Z
Message-ID: <3CAB2AC8.9070707@aol.com>
Date: Wed, 03 Apr 2002 11:16:08 -0500
From: Marc Zampetti <zampmarc@aol.com>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:0.9.4) Gecko/20011128 Netscape6/6.2.1
X-Accept-Language: en-us
MIME-Version: 1.0
To: Iwish <iwish@example.com>
Subject: Re: wish list
References: <Q0nAfLu2BHA.216@forums.sybase.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Newsgroups: sybase.public.ase.product_futures_discussion
Lines: 82
NNTP-Posting-Host: h-64-236-138-1.aoltw.net 64.236.138.1
Path: forums-1-dub!forums-master.sybase.com!forums.sybase.com
Xref: forums-1-dub sybase.public.ase.product_futures_discussion:568
Article PK: 94096

See my commpents inline.

Iwish wrote:

> Hi
>
> My wish lists.............
>
> [ASE]
> - clustering (like a Oracle Parallel Server)
> - Very Large Database support(Over 1TB)

This is already here. I've personally run databases over 2TB is size
without issue. I think the limit is 8TB in 12.5. If you really want VLDB
for DSS/DW apps, then check out IQ. It is MUCH better at that job.


> - partitioning(range, hash, key)
> - job scheduler in ASE(like a crond of UNIX)


Would want it as a seperate server. Honestly, the OS provides this much
better than Sybase can. Not high on my list.


> - transaction without logging


This can be useful. Should change to be "minimally logged transactions".
Allow insert/update/delete to be minamally logged. Just enough to allow
recovery of the database, without corrupting it. Document that if the
server crashes before checkpoint can write out the data pages, then oh
well. Poor user!


> - notify option(like a recovery percent at startup)


Why? A single script would handle this too.


> - dbcc performance enhancement


No argument here.


> - resizable database size(can be shrunken)


Would be nice. Not too high on my list, but something to work for.


> - rename object(user table, user stored procedure, database)


Checkout sp_rename for objects, and sp_renamedb for databases. This has
been here for ever.


> - vdevno of device(automatic increase)


Think its gone in 12.5. Simple to write a sproc to avoid this as well.


> - alias(object, database) <- not a user alias


Why?


> - temporary set option(not a session based option but a query based option)


I would be interested in understanding this. Why not use set xxx, run
query, and then run set xxx to reset?


>
> [T-SQL]
> - set rowcount 100 skip 90 (can select a portion of data)


Seen a lot of this. Would be nice, but at what cost?


> - inline view(derived table)

> - rownum function (or number(*) in asa)
> - user defined function(T-sql)
> - rank,replace function
> - simple encryption/decryption function
> - date type(not a datetime)
> - variable binding in dynamic sql
> - cursor Behavior(forward, backward, jump) & improve cursor performance
> - global nextval(not a identity, in memory, no physical i/o)
> - column alias can be used in having, group by
>
> [bcp]
> - conditional bcp out(not using a view)
>
> [isql]
> - reexecution (ie '/' of ORACLE)
> - formatting prompt (ie ase12:master 1> )


Use sqsh.


>
> [Sybase Central]
> - can view data(like a powerbuilder).
> - clean & neat design
> - support win32 version continuously
>
> [PC Client]
> - sql advantage


Check out 12.5.0.1. Its back!


> - need a basic OLAP tool


Definitely NOT! Why make Sybase spend its time writing these tools, when
there are LOTS of people doing it already. Just get one of those.


>
> [Backup/Restore]
> - backup/restore a single object
>
> [ESQLC]
> - array fetch to C Struct Type(ESQLC)
> - array insert
>
> [3rd party support]
> - SAP support
> - Tivoli support


Pretty sure Tivoli supports Sybase, to some extent.


>
>
> Thanks
>
>
>
>

Marc Zampetti


Pablo Sanchez Posted on 2002-04-03 17:25:36.0Z
From: "Pablo Sanchez" <pablo@dev.null>
References: <Q0nAfLu2BHA.216@forums.sybase.com> <3CAB2AC8.9070707@aol.com>
Subject: Re: wish list
Date: Wed, 3 Apr 2002 10:25:36 -0700
Lines: 20
Organization: High-Performance Database Engineering
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2600.0000
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000
Message-ID: <UeEAtWz2BHA.204@forums.sybase.com>
Newsgroups: sybase.public.ase.product_futures_discussion
NNTP-Posting-Host: 207.225.105.222
Path: forums-1-dub!forums-master.sybase.com!forums.sybase.com
Xref: forums-1-dub sybase.public.ase.product_futures_discussion:567
Article PK: 94095


"Marc Zampetti" <zampmarc@aol.com> wrote in message
news:3CAB2AC8.9070707@aol.com...
> See my commpents inline.
>
> > [isql]
> > - reexecution (ie '/' of ORACLE)
> > - formatting prompt (ie ase12:master 1> )
>
>
> Use sqsh.

Or EMACS ... start a SHELL, and run 'isql' within it. Very spiffy!
:)
--
Pablo Sanchez, High-Performance Database Engineering
mailto:pablo@hpdbe.com
Available for short-term and long-term contracts


Matt Posted on 2002-04-03 09:01:23.0Z
From: "Matt" <matt@fanhome.com>
References: <Q0nAfLu2BHA.216@forums.sybase.com>
Subject: Re: wish list
Date: Wed, 3 Apr 2002 04:01:23 -0500
Lines: 78
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2600.0000
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4910.0300
Message-ID: <Rzaq62u2BHA.153@forums.sybase.com>
Newsgroups: sybase.public.ase.product_futures_discussion
NNTP-Posting-Host: 24.164.69.119
Path: forums-1-dub!forums-master.sybase.com!forums.sybase.com
Xref: forums-1-dub sybase.public.ase.product_futures_discussion:570
Article PK: 94098

I am in no means employed by Sybase, but as a customer like yourself I would
express my concern with selected items:

"Iwish" <iwish@example.com> wrote in message
news:Q0nAfLu2BHA.216@forums.sybase.com...
> [ASE]
> - clustering (like a Oracle Parallel Server)

I don't think clustering is the 'be all, end all' to everyone's problems.
Oracle makes it seem like the silver-bullet but in practice it is anything
but what they claim.

> - Very Large Database support(Over 1TB)
I like ASE being a OLTP powerhouse. Use IQ if you want a data warehouse
solution. One reason why Oracle is so bloated is that it's two products in
one! We don't have a need for a warehouse solution, so why should we pay
for it?

> - job scheduler in ASE(like a crond of UNIX)
This was brought up on another thread; I'd want it to be a separate product
(again to keep ASE nice and clean).

> - transaction without logging
How would it roll back? I guess you'd just have it quit if it breaks, a la
MySQL... so it wouldn't be much of a transaction, then?

> - dbcc performance enhancement
Yes. It is currently prohibitively slow and CPU hogging to run the most
extensive option on big tables/DBs. Would be nice if there was a way to run
it without locking the whole table or DB.

> - resizable database size(can be shrunken)
YES. Especially transaction log segments!

> - rename object(user table, user stored procedure, database)
Doesn't sp_rename accomplish all but DB?

> - vdevno of device(automatic increase)
I thought 12.5 eliminated the user even needing to know the dev number... ?

> - temporary set option(not a session based option but a query based
option)
What does this mean?

> [T-SQL]
> - set rowcount 100 skip 90 (can select a portion of data)
I would like that as well (a la MySQL's LIMIT numrows, offset)

> - rownum function (or number(*) in asa)
If you use identity columns (or set the auto_ident flag on) you basically
have a row number. However, I'm not sure why such a thing would be needed.

> - cursor Behavior(forward, backward, jump) & improve cursor performance
Bleh. Cursors are not good from a set-theory (e.g. relational) perspective.
Never needed 'em. :)

> - global nextval(not a identity, in memory, no physical i/o)

> - column alias can be used in having, group by
That would be neat. You can already ORDER BY it, why not GROUP BY?

> [Sybase Central]
> - can view data(like a powerbuilder).
> - clean & neat design
> - support win32 version continuously

> [Backup/Restore]
> - backup/restore a single object
This too would be nice. BCP runs into problems if you're stuck on a 32 bit
environment and say you have a 2GB file-max limit. Plus, the native backup
is neater/cleaner since it's not plaintext. Still, there would be
significant problems in trying to LOAD your table and indexes if you used
the current backup architecture (I would think). Maybe a BCP-like dump
(e.g. just data and no indexes, etc.) but utilizing the current backupserver
technology (so I can COMPRESS, STRIPE ON, etc.).

--
Matt


Pablo Sanchez Posted on 2002-04-03 16:02:32.0Z
From: "Pablo Sanchez" <pablo@dev.null>
References: <Q0nAfLu2BHA.216@forums.sybase.com> <Rzaq62u2BHA.153@forums.sybase.com>
Subject: Re: wish list
Date: Wed, 3 Apr 2002 09:02:32 -0700
Lines: 73
Organization: High-Performance Database Engineering
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2600.0000
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000
Message-ID: <ySlQSoy2BHA.206@forums.sybase.com>
Newsgroups: sybase.public.ase.product_futures_discussion
NNTP-Posting-Host: 207.225.105.222
Path: forums-1-dub!forums-master.sybase.com!forums.sybase.com
Xref: forums-1-dub sybase.public.ase.product_futures_discussion:569
Article PK: 94099


"Matt" <matt@fanhome.com> wrote in message
news:Rzaq62u2BHA.153@forums.sybase.com...
>
> > - Very Large Database support(Over 1TB)
>
> I like ASE being a OLTP powerhouse.

It's very possible to have a 1TB+ OLTP application.

Even if the DB was used for DSS, one shouldn't have to purchase IQ and
tool the shop for it either.

> > - transaction without logging
> How would it roll back? I guess you'd just have it quit if it
breaks, a la
> MySQL... so it wouldn't be much of a transaction, then?

Based on the requirements, that is, if Sybase really wanted to
implement this the user group should be polled to determine the
requirements, it would dictate how the feature would be implemented.

It's possible to have an in-memory log device that acts as a scratch
pad but it's never flushed to disk on COMMIT. It might be possible to
see if all operations can become 'minimally logged' etc etc. Language
could be written in the docs to warn all about the evils and let it be
'user beware.'

> > - dbcc performance enhancement
> Yes. It is currently prohibitively slow and CPU hogging to run the
most
> extensive option on big tables/DBs. Would be nice if there was a
way to run
> it without locking the whole table or DB.

Seems like we might want is an option to run 'dbcc' at the same
priority as 'free checkpoints' -- during idle periods. It'd have to
be clever to 'hurry' until a resource is unlocked, then pause itself
if the machine got busy. Anyway, just throwing out random ideas ...

> > - temporary set option(not a session based option but a query
based
> option)
> What does this mean?

I believe the ability to set a 'set'able option only during the
duration of the next DML statement. I'm not sure the worth of this
one since it's just as easy to switch it off at the end of the DML.
But that's just my opinion. :)

> > - rownum function (or number(*) in asa)
> If you use identity columns (or set the auto_ident flag on) you
basically
> have a row number. However, I'm not sure why such a thing would be
needed.

'Find me the 10th through the 20th top rows' -- especially if it's an
impromptu request so you're hacking someting on 'isql'

> > - cursor Behavior(forward, backward, jump) & improve cursor
performance
> Bleh. Cursors are not good from a set-theory (e.g. relational)
perspective.
> Never needed 'em. :)

I concur. The unfortunate thing is that a lot of developers are more
comfortable thinking row-at-a-time than set based.
--
Pablo Sanchez, High-Performance Database Engineering
www.hpdbe.com
Available for short-term and long-term contracts


Matt Rogish Posted on 2002-04-03 20:18:10.0Z
From: "Matt Rogish" <matt@fanhome.com>
References: <Q0nAfLu2BHA.216@forums.sybase.com> <Rzaq62u2BHA.153@forums.sybase.com> <ySlQSoy2BHA.206@forums.sybase.com>
Subject: Re: wish list
Date: Wed, 3 Apr 2002 15:18:10 -0500
Lines: 101
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2600.0000
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000
Message-ID: <VTZ26202BHA.153@forums.sybase.com>
Newsgroups: sybase.public.ase.product_futures_discussion
NNTP-Posting-Host: ip134-053-064-186.s64.muohio.edu 134.53.64.186
Path: forums-1-dub!forums-master.sybase.com!forums.sybase.com
Xref: forums-1-dub sybase.public.ase.product_futures_discussion:566
Article PK: 94091


"Pablo Sanchez" <pablo@dev.null> wrote in message
news:ySlQSoy2BHA.206@forums.sybase.com...
> It's very possible to have a 1TB+ OLTP application.
>
> Even if the DB was used for DSS, one shouldn't have to purchase IQ and
> tool the shop for it either.

That's true.. Although as was said in the other post in the thread it can go
up to what, 18TB? That's quite a bit... But I like to subscribe to the "Do
one (or a few) thing and do it well" rather than Oracle and MS SQL's "Let's
try and cram as much as we can in there" when it is obvious total quality
suffers. If ASE can support 100TB DBs without any impact on my OLTP
application (and without any impact on future development and bug fixes)
then I have no problem with it... But it seems to me they already did that
in IQ (why reinvent the wheel?)... Although I have no IQ experience -- how
similar is it to ASE? I would think there would be, to make IQ worth it,
significant interoperability between ASE and IQ (so that your ASE OLTP app
can easily shuffle data over to your IQ data warehouse for aggregation).

> > > - transaction without logging
> Based on the requirements, that is, if Sybase really wanted to
> implement this the user group should be polled to determine the
> requirements, it would dictate how the feature would be implemented.
>
> It's possible to have an in-memory log device that acts as a scratch
> pad but it's never flushed to disk on COMMIT. It might be possible to
> see if all operations can become 'minimally logged' etc etc. Language
> could be written in the docs to warn all about the evils and let it be
> 'user beware.'

Sounds fine to me as well. I know we do have 'non-critical' updates /
deletes (mainly to maintain denormalized data) which if they failed, eh, in
the grand scheme of things they can easily be regenerated (e.g. instead of
sum()'ing up rows store a counter in a table. If the counter fails to
increment, we can always run batches at night to reinitialize the counters
based upon the actual sum). If we can sqeek out some more performance by
ignoring rollbacks or logging (and again as was stated in the other post
keep my DB from corrupting) then I'd use it.

> > > - dbcc performance enhancement
> Seems like we might want is an option to run 'dbcc' at the same
> priority as 'free checkpoints' -- during idle periods. It'd have to
> be clever to 'hurry' until a resource is unlocked, then pause itself
> if the machine got busy. Anyway, just throwing out random ideas ...

Would it also make sense to perform row-locking cleanups during idle times
as well? e.g. performing reorgs on DOL tables when they're not being hit
very hard could pay huge dividends for the 24/7 app which uses the DOL table
'occationally' but is large enough that you don't want it to run all at once
(and don't have the time / patience to use the stop/start functionality). I
think this treads into the MS SQL method of 'doing things for you' so I'm
still on the fence. We don't have any significant tables with DOL locking
but I could see that there could be shops who could utilize this
functionality.

> > > - temporary set option(not a session based option but a query
> I believe the ability to set a 'set'able option only during the
> duration of the next DML statement. I'm not sure the worth of this
> one since it's just as easy to switch it off at the end of the DML.
> But that's just my opinion. :)

Ah I see. Yeah, that doesn't make much sense to me as well:
SET ROWCOUNT 10 FOR THIS QUERY ONLY
SELECT..

If you embed it in a stored procedure there's no real pain to go:
SET ROWCOUNT 10
SELECT
SET ROWCOUNT 0

> > > - rownum function (or number(*) in asa)
> 'Find me the 10th through the 20th top rows' -- especially if it's an
> impromptu request so you're hacking someting on 'isql'

That's true -- although I'd think that would only be useful in certain
occations -- e.g. without where or order by or group by clauses since the
numbers wouldn't be contiguous... Applying them to a result set
(pseudocolumn) instead of the underlying table themselves makes more sense
to me..
SELECT ...
FROM t1
WHERE ..
AND result_set_row_num BETWEEN 10 and 20

> > > - cursor Behavior(forward, backward, jump) & improve cursor
> I concur. The unfortunate thing is that a lot of developers are more
> comfortable thinking row-at-a-time than set based.

:(

> --
> Pablo Sanchez, High-Performance Database Engineering
> www.hpdbe.com
> Available for short-term and long-term contracts

--
Matt


Pablo Sanchez Posted on 2002-04-03 21:44:27.0Z
From: "Pablo Sanchez" <pablo@dev.null>
References: <Q0nAfLu2BHA.216@forums.sybase.com> <Rzaq62u2BHA.153@forums.sybase.com> <ySlQSoy2BHA.206@forums.sybase.com> <VTZ26202BHA.153@forums.sybase.com>
Subject: Re: wish list
Date: Wed, 3 Apr 2002 14:44:27 -0700
Lines: 131
Organization: High-Performance Database Engineering
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2600.0000
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000
Message-ID: <WkVmWn12BHA.206@forums.sybase.com>
Newsgroups: sybase.public.ase.product_futures_discussion
NNTP-Posting-Host: 207.225.105.222
Path: forums-1-dub!forums-master.sybase.com!forums.sybase.com
Xref: forums-1-dub sybase.public.ase.product_futures_discussion:565
Article PK: 94092


"Matt Rogish" <matt@fanhome.com> wrote in message
news:VTZ26202BHA.153@forums.sybase.com...
>
> "Pablo Sanchez" <pablo@dev.null> wrote in message
> news:ySlQSoy2BHA.206@forums.sybase.com...
> > It's very possible to have a 1TB+ OLTP application.
> >
> > Even if the DB was used for DSS, one shouldn't have to purchase IQ
and
> > tool the shop for it either.
>
> That's true.. Although as was said in the other post in the thread
it can go
> up to what, 18TB? That's quite a bit... But I like to subscribe to
the "Do
> one (or a few) thing and do it well" rather than Oracle and MS SQL's
"Let's
> try and cram as much as we can in there" when it is obvious total
quality
> suffers.

Methinks you've been listening to the rah-rah machine too much. :)
I'm not trying to defend Oracle nor MS SQL Server, however, they are
both very good products. I have worked with Oracle for the past
couple of years and recently have been doing a lot of hacking in
PL/SQL. There are some features in PL/SQL that I'd love to see T-SQL
support (associative arrays so they can in turn be used to feed
dynamic SQL, etc). Of course Oracle has some very blatant warts and
overall, it's not as clean as Sybase (in my opinion). But that
doesn't matter. What matters is paying the bills and after all, a
DBMS really is a DBMS is a DBMS.

Having said the above, I also agree with the Unix philosophy (which is
what you're paraphrasing) of creating finely tuned building blocks,
and stringing (piping? :) them along.

I do believe that Sybase can do a bit better about trying to suspend
larger I/O's rather than a whole bunch of little ones. For example,
Oracle, using the DB_FILE_MULTIBLOCK_READ_COUNT, can suspend very
large I/Os - on IRIX we had the option of sucking down 4MB in one
read(). Yeows! For DSS applications where you're mining through tons
of information, this is a must.

Based on Oracle's architecture it tends to consume a lot more
resources to get the same OLTP output than Sybase/MS SQL Server. If
you get a chance, look at the www.tpc.org 'C's (OLTP) benchmarks for
some interesting information.

Anyway, each does have its pro's and con's.

I am saddened that Sybase blew its lead in the early 90's.

> If we can sqeek out some more performance by
> ignoring rollbacks or logging (and again as was stated in the other
post
> keep my DB from corrupting) then I'd use it.

Bingo!

> > > > - dbcc performance enhancement
> > Seems like we might want is an option to run 'dbcc' at the same
> > priority as 'free checkpoints' -- during idle periods. It'd have
to
> > be clever to 'hurry' until a resource is unlocked, then pause
itself
> > if the machine got busy. Anyway, just throwing out random ideas
...
>
> Would it also make sense to perform row-locking cleanups during idle
times
> as well? e.g. performing reorgs on DOL tables when they're not
being hit
> very hard could pay huge dividends for the 24/7 app which uses the
DOL table
> 'occationally' but is large enough that you don't want it to run all
at once
> (and don't have the time / patience to use the stop/start
functionality). I
> think this treads into the MS SQL method of 'doing things for you'
so I'm
> still on the fence. We don't have any significant tables with DOL
locking
> but I could see that there could be shops who could utilize this
> functionality.

Sure, this totally makes sense ... the DBA can schedule so they're in
control as opposed to behind the scenes types of activities.

As a side note and really just my opinion so chuck it if need be, it's
good to look at what the other vendors are doing to bring the best-of
back into Sybase. You know that Sybase Engineering strives to write
tight code. This means that if we can help them get features that
rock, they might get more market share and ... anyway, you know where
I'm going with it.

> > > > - rownum function (or number(*) in asa)
> > 'Find me the 10th through the 20th top rows' -- especially if it's
an
> > impromptu request so you're hacking someting on 'isql'
>
> That's true -- although I'd think that would only be useful in
certain
> occations -- e.g. without where or order by or group by clauses
since the
> numbers wouldn't be contiguous... Applying them to a result set
> (pseudocolumn) instead of the underlying table themselves makes more
sense
> to me..
> SELECT ...
> FROM t1
> WHERE ..
> AND result_set_row_num BETWEEN 10 and 20

Agreed!

> > > > - cursor Behavior(forward, backward, jump) & improve cursor
> > I concur. The unfortunate thing is that a lot of developers are
more
> > comfortable thinking row-at-a-time than set based.
>
> :(

Look on the bright side, areas to tune and make the application ...
high-performance! <g>
--
Pablo Sanchez, High-Performance Database Engineering
mailto:pablo@hpdbe.com
Available for short-term and long-term contracts