RE: sql developer vs sqlplus sql and plsql syntax

From: Mark W. Farnham <mwf_at_rsiz.com>
Date: Tue, 15 Dec 2020 10:54:41 -0500
Message-ID: <1b9101d6d2fa$9a5fb9b0$cf1f2d10$_at_rsiz.com>



This entire kerfuffle was in fact specified to vendors by the MOSES protocol published circa 1995. MOSES was Massive Open Systems Environment Standards. Pioneers of using UNIX on a large commercial scale founded it to create a semblance of order in the chaotic world of Unix. John Black of USWEST New Vector noted two key things: 1) From shop to shop in the MVS [that was a mainframe opsys from IBM] world things vary in operations by one or two percent, but in UNIX shops things vary from shop to shop by 50%. 2) The power of UNIX is that there are dozens of ways to do pretty much anything; the weakness of UNIX is that there are dozens of ways to do pretty much anything.  

The second one is probably not original to John. I think the first on was.  

The key that was envisaged by MOSES was driving once via GUI, generating a line operation equivalent as both a log of operations AND an editable execution script. Then you could edit the variables for things like different ip addresses and machine names and have a repeatable, schedulable, testable script for each machine that had been built by driving the GUI with all the interactive checks and hints and so forth.  

Sigh. Once upon a time that was a useful document, and whether or not vendors followed all the specifications, that one was the big deal: Everything you can make a change with in GUI, you can log and script.  

Good luck. This is important. However you handle it, you have to remember that eventually someone will hire an inexperienced person with limited intellectual horizons to push the buttons. Or an automaton will do it without any judgement at all.  

OH – and auditors deprecate material GUI actions that are not comprehensively logged. So it is best if the tool does it as a builtin.  

mwf  

From: oracle-l-bounce_at_freelists.org [mailto:oracle-l-bounce_at_freelists.org] On Behalf Of Ls Cheng Sent: Tuesday, December 15, 2020 6:55 AM To: William Robertson
Cc: oracle-l; Jeff Smith
Subject: Re: sql developer vs sqlplus sql and plsql syntax  

Hi  

I was, am and still shocked too that they never used SQLPLUS, they don't even know the name. And these are not junior developers, I wonder where they have been living...! The quality of developers nowadays......  

Thanks      

On Tue, Dec 15, 2020 at 12:25 PM William Robertson <william_at_williamrobertson.net> wrote:

I'm frankly shocked at "But the developers have never used SQLPLUS so..." Are these not Oracle developers?  

Probably just making sure they "Run as script" would go most of the way, but obviously they really should make the effort to test it using the agreed deployment method.  

Jeff, how about a post about setting up "Run with SQL*Plus" as an external tool in SQL Developer? Then they would just have to press one button, if that's not too much effort for them.  

William    

On 14 Dec 2020, at 20:18, Ls Cheng <exriscer_at_gmail.com> wrote:  

Hi  

but SQLCl is CLI correct?  

Thanks  

On Mon, Dec 14, 2020 at 8:58 PM Jeff Smith <jeff.d.smith_at_oracle.com> wrote:

Sure it does, SQLcl and SQLDev use the exact same script engine.  

From: Ls Cheng <exriscer_at_gmail.com>
Sent: Monday, December 14, 2020 2:39 PM
To: Jeff Smith <jeff.d.smith_at_oracle.com> Cc: Oracle Mailinglist <oracle-l_at_freelists.org> Subject: Re: sql developer vs sqlplus sql and plsql syntax

Hi

That does make any difference, developers wants to use GUI, not CLI unfortunately

BR  

On Mon, Dec 14, 2020 at 5:59 PM Jeff Smith <jeff.d.smith_at_oracle.com> wrote:

Or…ask the DBAs to meet their developers half way and start using SQLcl to do the pushes to prod.

There are SET commands you can use to define what statement delimiters you want, but again, at the end of the day, if you’re not testing it exactly as it’s going to run in production, that’s FAIL.  

From: Ls Cheng <exriscer_at_gmail.com>
Sent: Monday, December 14, 2020 10:09 AM To: Jeff Smith <jeff.d.smith_at_oracle.com> Cc: Oracle Mailinglist <oracle-l_at_freelists.org> Subject: Re: sql developer vs sqlplus sql and plsql syntax  

Hi  

"If you know the DBAs are using SQL*Plus to roll out updates, then the developers need to USE SQL*Plus for testing their scripts. It’s that simple."  

Yes, that is being asked, the developers MUST test their scripts with SQLPLUS otherwise there is no way to automate updates. But the developers have never used SQLPLUS so it is very hard to ask them to move from GUI to CLI. So I was thinking if SQL developer could configure in such a way so semicolon and backslash can be forced...  

BR  

On Mon, Dec 14, 2020 at 2:41 PM Jeff Smith <jeff.d.smith_at_oracle.com> wrote:

If you ever want to talk to me, feel free to contact me directly

Jeff.d.smith_at_oracle.com  

DBA team and the developers because the developers write codes in SQL developer then they upload them as script for the DBA to run in the test/prod environment which all fails (and plsql objects not created) obecause DBA's uses SQLPLUS and because most SQL execution request are batched with SQLPLUS.  

Wrong, wrong, wrong. Bad, bad, bad. Developers and DBAs need to agree on the requirements for production rollouts/promoting code.  

If you know the DBAs are using SQL*Plus to roll out updates, then the developers need to USE SQL*Plus for testing their scripts. It’s that simple.  

The DBAs should have figured out by now what was going on and had some checks in place to prevent this from happening accidently. This is more of a communication issue than a tooling issue.  

Jeff    

From: Ls Cheng <exriscer_at_gmail.com>
Sent: Sunday, December 13, 2020 1:19 PM
To: Oracle Mailinglist <oracle-l_at_freelists.org> Subject: sql developer vs sqlplus sql and plsql syntax  

Hi  

This is for Jeff Smith! :-)  

I would like to know why SQL Developer allow people to execute SQL code without semicolon and PLSQL code with backslash? (OK TOAD does the same thing but it is not made by Oracle)  

For example code such as this works in SQL developer  

select * from dba_users  

  • backslash missing

declare
l_a number;
begin
select 1 into l_a from dual;
end;  

  • backslash missing

create or replace procedure p_test
is
l_a number;
begin
select 1 into l_a from dual;
end;    

This is causing some serious conflicts between DBA team and the developers because the developers write codes in SQL developer then they upload them as script for the DBA to run in the test/prod environment which all fails (and plsql objects not created) obecause DBA's uses SQLPLUS and because most SQL execution request are batched with SQLPLUS.  

The developers claim if the code runs in SQL developer it must run in SQLPLUS because it is a tool from Oracle. Is there a way to configure SQL Developer so it can have the same behaviour as SQLPLUS?  

Thanks    

--
http://www.freelists.org/webpage/oracle-l
Received on Tue Dec 15 2020 - 16:54:41 CET

Original text of this message