Re: SQL Command list history on Linux

From: Jeff Chirco <backseatdba_at_gmail.com>
Date: Thu, 4 Jan 2018 09:14:25 -0800
Message-ID: <CAKsxbLoKf_-QPGZqegi8YzbCMsyiRTzEG6nm5MgkuYJvVCFw7w_at_mail.gmail.com>



I have 12.2 instance and using Java that came with it, with sqlcl 17.3.0 and it takes about 4 seconds every time to do a sqlcl /nolog. sqlplus /nolog is instant. This is all on Oracle Linux 7

On Thu, Jan 4, 2018 at 8:00 AM, Jeff Smith <jeff.d.smith_at_oracle.com> wrote:

> It’s all good, I’m not upset or worried.
>
>
>
> My job is to help build tools, and help people be successful using said
> tools.
>
>
>
> Most of our users never touch the servers, they only connect to them from
> their desktops. So that’s the primary audience I’m trying to reach.
>
>
>
> Also, I tend to open SQLcl (and SQL Developer) and never close it. Also
> for Java, the first time you open a program, which will kick off the jvm,
> you’ll see a perf hit. The 2nd time you open it, not so much. So you in
> general can’t assume if it takes 4 seconds to open once, that I’ll take 4
> seconds to open the 2nd time.
>
>
>
> The DB comes with Java, so it’s already there for when you get to 12c and
> higher installs.
>
>
>
> *From:* Neil Chandler [mailto:neil_chandler_at_hotmail.com]
> *Sent:* Thursday, January 04, 2018 10:57 AM
> *To:* Jeff Smith <jeff.d.smith_at_oracle.com>; oracle-l_at_freelists.org;
> gogala.mladen_at_gmail.com
>
> *Subject:* Re: SQL Command list history on Linux
>
>
>
> Jeff,
>
>
>
> The main traction against sqlci isn't the several seconds to start rather
> than pretty much instant sqlplus (this becomes a factor when calling
> sqlplus multiple times in a loop), nor the fact it's far superior to
> sql*plus, but the red tape, security clearance, patching, etc, required to
> get new software (especially the Java component) onto the all of the DB
> servers. That pretty much nobody is using 12.2 in Production yet compounds
> it, so the red tape is required. There's nothing you can do about this that
> time won't cure except make the sqlci case increasingly compelling.
>
>
>
> I use vi and sqlplus because they are guaranteed to be there. If I was a
> perm DBA in a company long-term, I'd probably raise the paperwork to get it
> onto every DB server purely for the automatic column size adjustments if
> nothing else. Love that.
>
>
>
> Neil Chandler
>
>
>
>
> ------------------------------
>
> *From:* Jeff Smith <jeff.d.smith_at_oracle.com>
> *Sent:* 04 January 2018 15:11
> *To:* neil_chandler_at_hotmail.com; oracle-l_at_freelists.org;
> gogala.mladen_at_gmail.com
> *Subject:* RE: SQL Command list history on Linux
>
>
>
> Slow to start? Maybe a second, or 3? If it’s taking longer than that,
> please let me know.
>
>
>
> Java 8u50 or higher will be ok, which is about 2-2.5 years old?
>
>
>
> It’s a 20mb or so download, so takes less than a minute to install if you
> don’t have it. Less time than it takes maybe to setup rlwrap…esp if you’re
> on Windows.
>
>
>
> *From:* Neil Chandler [mailto:neil_chandler_at_hotmail.com
> <neil_chandler_at_hotmail.com>]
> *Sent:* Thursday, January 04, 2018 7:59 AM
> *To:* oracle-l_at_freelists.org; gogala.mladen_at_gmail.com
> *Subject:* Re: SQL Command list history on Linux
>
>
>
> Mladen,
>
>
>
> What has that go to do with the original posters question?
>
> If you are developing PL/SQL in SQL*Plus, you're probably not doing it
> efficiently. SQL Developer is the better (free) tool for that, regardless
> of platform.
>
> SQLCI is great, but it's slow to start, requires a fairly recent Java
> release, and only natively available from 12.2.
>
> If you want Windows-like editing on a Linux SQL*Plus (or dgmgrl, rman or
> whatever), use rlwrap.
>
>
>
> regards
>
>
>
> Neil Chandler.
>
> Database Guy.
>
>
> ------------------------------
>
> *From:* oracle-l-bounce_at_freelists.org <oracle-l-bounce_at_freelists.org> on
> behalf of Mladen Gogala <gogala.mladen_at_gmail.com>
> *Sent:* 04 January 2018 10:57
> *To:* oracle-l_at_freelists.org
> *Subject:* Re: SQL Command list history on Linux
>
>
>
> On 01/02/2018 04:51 PM, Neil Chandler wrote:
>
> Jeff,
>
>
>
> My preferred solution is to install "rlwrap" [ e.g. "yum install rlwrap" -
> check out https://github.com/hanslub42/rlwrap
> <https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_hanslub42_rlwrap&d=DwMFAw&c=RoP1YumCXCgaWHvlZYR8PZh8Bv7qIrMUB65eapI_JnE&r=N2hWu5HFsaIjmMkjQbnlokJ7uinNZMgPVk8rqPT9esM&m=0bds7QIoykh2KktiFwI5q5NtAunsmhvpPbDLFAGKbBY&s=cRMpNNnqbVGwt9R-SUJkdE4mCdC-1KZy0LkLX77BvUY&e=>
> ] it provides similar functionality to a windows CMD environment. You just
> call the command you want to use via rlwrap
>
>
>
> e.g. rlwrap sqlplus, or rlwrap dgmgrl
>
>
>
> To keep thing simple, you can alias it
>
>
>
> alias rsql='rlwrap sqlplus'
>
> alias rdg='rlwrap dgmgrl'
>
>
>
> regards
>
>
>
> Neil Chandler
>
>
>
> Hi Neil,
> Rlwrap is not an equivalent solution. The "rlwrap" trick breaks editing of
> PL/SQL code, which can be annoying. If you write a multi-line PL/SQL
> snippet and something goes wrong, "ed" will only return a part of the
> snippet. SQLCl will do the right thing and give you back the entire
> procedure. It has a primitive screen editor built into the code because
> arrow keys also work when editing the history commands. Also, SQLCl has
> "REPEAT" command which turns your SQLCl session into an instant monitor, if
> necessary. In addition to all that, there is Jeff Smith, who has been doing
> an outstanding job with the product and explaining it on this list. Jeff is
> extremely helpful and a very nice guy. And no, I don't work for Oracle
> Corp. and I don't plan on working for Oracle anytime soon.
> Regards
>
> --
>
> Mladen Gogala
>
> Oracle DBA
>
> http://mgogala.freehostia.com <https://urldefense.proofpoint.com/v2/url?u=http-3A__mgogala.freehostia.com&d=DwMFAw&c=RoP1YumCXCgaWHvlZYR8PZh8Bv7qIrMUB65eapI_JnE&r=N2hWu5HFsaIjmMkjQbnlokJ7uinNZMgPVk8rqPT9esM&m=0bds7QIoykh2KktiFwI5q5NtAunsmhvpPbDLFAGKbBY&s=PcMyzBvpXVEJ8OtXmrHXL-cnBgMOPb4mBf88-fufLHU&e=>
>
>

--
http://www.freelists.org/webpage/oracle-l
Received on Thu Jan 04 2018 - 18:14:25 CET

Original text of this message