]]> http://jes.blogs.shellprompt.net/2012/09/03/odtug-apexposed-2012/feed/ 0 http://jes.blogs.shellprompt.net/2012/03/27/odtug-apex-plug-in-competition/ http://jes.blogs.shellprompt.net/2012/03/27/odtug-apex-plug-in-competition/#comments Tue, 27 Mar 2012 07:14:21 +0000 John Scott http://jes.blogs.shellprompt.net/?p=501 As Martin has already announced, there is a very exciting APEX Plugin Competition hosted by ODTUG.

Here’s the official announcement (taken from the ODTUG page):

ODTUG is holding its first ever world wide APEX Plug-ins competition! This is your opportunity to write an APEX plug-in and have it viewed by the entire international APEX community and maybe crowned 2012 ODTUG APEX Plug-in Developer of the Year! Even better, there are some really great prizes with the grand prize being a free pass to Kscope13 or a Jawbone Jambox!

This is an opportunity for you to not only win some cool prizes but also to show just what is possible with APEX plugins (and don’t worry if you’re a beginner in APEX, there is a separate Experts category to make sure that everyone is judged as equally as possible).

It’s been around 3 years since I wrote Pro Oracle Application Express (with a chapter contributed by Scott Spendolini), and like most first time authors I said “That was a great experience, but I’m in no hurry to write another book again!”.

Writing a technical book is a very rewarding experience, but trying to do it alongside your regular day job is tough (tougher than I ever imagined). Trying to meet publishers deadlines while still trying to  keep your (paying!) clients happy is very tough indeed.

However, a couple of events made me rethink my “never again” policy, firstly the very sad passing of Carl Backstrom and then the equally sad and unexpected passing of Scott Spadafore.

These two people helped me immensely during my time with Oracle Application Express and since I knew them both personally I wondered what I could do to help the families of two people who I admired immensely and also considered friends.

That’s when I came up with the idea of writing a collaborative book, with each chapter written by a different author who knew both Carl and Scott.

At ODTUG last year I approached people who I knew felt the same way about Carl and Scott as I did and I felt an enormous sense of pride that every single one of the authors didn’t hesitate in agreeing to contribute a chapter, so for that I am eternally grateful to:

There is some great technical content in this book, written by some of the worlds best known Oracle APEX experts. Even if you don’t think you need this book, buy it…PLEASE.

NOTE: ALL AUTHOR ROYALTIES ARE SPLIT 50/50 BETWEEN THE TWO MEMORIAL FUNDS.

A PLEA FROM ME: If you are only going to buy 1 book on APEX this year…please make it this one, if you are going to buy more than 1…well make sure one of them is this book!

More information is available on Joels blog, and you can download APEX 4.0 as always from the main OTN APEX page here

If you haven’t heard about it yet, it’s a website where you can pit your wits against a daily question by none other than Steven Feuerstein (yes that Steven Feuerstein!) and Finn Ellebaek Nielsen. Wait it’s even cooler than that, you can win cash (and other!) prizes and if that’s not enough to tempt you then you can see how you rank against other PL/SQL developers all over the world (and who wouldn’t want to know that!).

Steven asked APEX Evangelists (what seems like a long time ago – before we formed Sumneva) to help him create the front-end for the PL/SQL Challenge website (we previously created the I Love PL/SQL site for him) and we (myself, Dimitri Gielis and Paul Broughton) have worked in our ‘spare time’ to help to bring the site to life.

We of course used Oracle Application Express to create the front-end application (all of the back-end is PL/SQL code, what else did you expect from Steven?!), and even though I’m biased I have to say this makes for a dream development environment.

I have to say it’s been a pleasure to work with two professionals such as Steven and Finn. It really has been a baptism by fire for Paul working on such a project (how many interns get the chance to work with such giants as Steven?) but he more than rose to meet the challenge, thanks Paul!

So what are you waiting for? Go along…sign up…play the quiz…and see how well you can do (someone is going to win those prizes!).

If you didn’t already see the news, the EA2 release of APEX4.0 has just been opened up.

Head over to www.tryapexnow.com, sign up for a workspace and try out all the cool new features.

EA2 includes the following key features -

• Websheets (this will be huge!)
• New collection enhancements
• Tree region
• Debugging has been rewritten

and many more (see Joel’s blog post for a full description).

APEX 4.0 is going to be a huge release and will have (arguably) the biggest impact of any APEX release, it’s certainly interesting times being involved in the APEX community.

For example, would you rather write the query -

SELECT
  ename,
  department_name
FROM
  emp_vw

or

SELECT
  e.ename,
  d.dname
FROM
  emp e,
  dept d
WHERE
  e.deptno = d.deptno;

It’s a rhetorical question so please don’t say “I prefer the second”…otherwise there’s no point me continuing this post. Anyway, that’s quite a simple example, imagine the view hides all sorts of complexity (analytics, subquerys etc).

The view gives us a single ‘point of truth’, we can tell developers “use that view, it contains all the information you need, you don’t need to understand all the logic yourself”, views are a very powerful tool for code re-usability (and simplification).

However, what if someone codes a view like this -

CREATE OR REPLACE VIEW empdept_vw AS
SELECT
  e.ename,
  d.dname,
  e.sal
FROM
  emp e,
  dept d
WHERE
  e.deptno = d.deptno
ORDER BY d.dname;

Notice the order by there? They probably/possibly created that view to use directly in a report, adding the ORDER BY into the view so they could just write -

SELECT * FROM empdept_vw

rather than

SELECT * FROM empdept_vw edv ORDER BY edv.dname

Well that’s fine, but what is someone else later on uses that view in this way -

SELECT * from empdept_vw edv ORDER BY edv.sal DESC

Well we’re now potentially making the database do more work because the query is sorting the data in a different way to the view sorts it (so therefore the sorting in the view is essentially redundant). The key point here is that the database cost based optimizer might be smart enough to figure out what you’re doing and eliminate the redundant sort, however there might be lots of cases where that doesn’t happen (consider the scenario where you have views built on top of views, with ORDER BY’s thrown in there at a deep level).

The cost-based optimizer gets smarter with every release of the Oracle datbase, however it will never (and I’m reasonably confident about this!) understand your data and requirements as much as you do (or rather as much as you should).

So, I would argue anytime you see a view with an ORDER BY in it you should really question why you’re doing that (rather than putting the ORDER BY in the query that uses the view). We can even make our job easier and look for any views that have an ORDER BY defined in them by querying DBA_VIEWS or USER_VIEWS. Unfortunately those views contain the text definition of the view in a LONG column rather than a CLOB, which means we can’t do a simple LIKE predicate -

JES@dbtest> select * from user_views where upper(text) like '%ORDER BY%';
                                     *
ERROR at line 1:
ORA-00932: inconsistent datatypes: expected NUMBER got LONG

We can get round this in a number of ways (such as creating a copy of the DBA_VIEWS or USER_VIEWS in a local table converting the LONG column to a CLOB etc), or we can use a different approach like this -

JES@dbtest> select
  *
from
  user_views
where
  dbms_xmlgen.getxml('select text from user_views where view_name = ''' || view_name || '''') like '%ORDER BY%'

(note unfortunately I can’t remember who to credit with the DBMS_XMLGEN trick above, I saw it a long time ago and it’s been lost in the mists of time).

and yes…before you ask…in the schema I ran that query in I did find a view that had an unnecessary ORDER BY in ;)

I’d be interested to hear your thoughts on if you find ORDER BY’s in views a problem/issue/area-of-concern, or if it’s just me?

However, a question does come up from time to time, often along the lines of “I am using a collection but it is slow…”.

Typically, you might create a collection in the following way -

APEX_COLLECTION.CREATE_COLLECTION_FROM_QUERY(
  p_collection_name => 'COLL1',
  p_query           => 'select * from emp',
  p_generate_md5    => 'NO');

This would create a collection (called ‘COLL1′) and populate it with the results of the query (select * from emp).

So, let’s take a quick example. Firstly I want to use a sizable table

JES@dbtest> create table my_all_objects as select * from all_objects;

Table created.

JES@dbtest> select count(*) from my_all_objects;

  COUNT(*)
----------
     55853

So now I’m going to create an APEX application, which has the following code in a process:

APEX_COLLECTION.CREATE_COLLECTION_FROM_QUERY(
  p_collection_name => 'COLL1',
  p_query           => 'select * from my_all_objects',
  p_generate_md5    => 'NO');

It’s not really important to see the APEX application as such, what is important is to see how long it takes to run this process. There are lots of ways I can achieve that (putting in my own instrumentation etc), but APEX comes with a nice debug mode, which shows me timing information for everything that it does when you run a page. Running that page in debug mode I see something similar to -

0.03: ...Process "Create 1": PLSQL (AFTER_SUBMIT) APEX_COLLECTION.CREATE_COLLECTION_FROM_QUERY( p_collection_name => 'COLL1', p_query => 'SELECT * FROM my_all_objects', p_generate_md5 => 'NO');
18.57: ...Do not run process "Create_B", process point=AFTER_SUBMIT, condition type=, when button pressed=CREATE_B

The important bit here is the difference in timings (18.57 – 0.03) between the two lines, so that means my process took around 18.54 seconds to create the collection. 18 seconds is a long time for an user to wait for something to happen in a web application.

So, what can we do to improve things? Well if you look in the documentation for APEX_COLLECTION you’ll see there is another routine called CREATE_COLLECTION_FROM_QUERY_B. From the documentation -

The CREATE_COLLECTION_FROM_QUERY_B method offers significantly faster performance than the CREATE_COLLECTION_FROM_QUERY method by performing bulk SQL operations, but has the following limitations:

No column value in the select list of the query can be more than 2,000 bytes. If a row is encountered that has a column value of more than 2,000 bytes, an error will be raised during execution.

The MD5 checksum will not be computed for any members in the collection.

Ok, so if we can live with those limitations (the restriction about not computing the checksum is not as significant as it might at first seem, since we could also manually calculate it ourselves once the collection has been populated), then it’s worth trying it, so changing our code from -

APEX_COLLECTION.CREATE_COLLECTION_FROM_QUERY(
  p_collection_name => 'COLL1',
  p_query           => 'SELECT * FROM my_all_objects',
  p_generate_md5    => 'NO');

to

APEX_COLLECTION.CREATE_COLLECTION_FROM_QUERY_B(
  p_collection_name => 'COLL1',
  p_query           => 'SELECT * FROM my_all_objects');

as you can see, it’s a relatively minor change. So let’s run the page in debug mode again and see if there is an improvement:

0.03: ...Process "Create_B": PLSQL (AFTER_SUBMIT) APEX_COLLECTION.CREATE_COLLECTION_FROM_QUERY_B( p_collection_name => 'COLL1', p_query => 'SELECT * FROM my_all_objects');
5.25: Branch point: AFTER_PROCESSING

So this time it took around 5.2 seconds (5.25 – 0.03) to create the collection. (Note that I ran these tests multiple times to remove any caching issues).

Whilst I’d say 5.2 seconds is still a long time for a user to wait for a webpage, it’s a significant improvement over 18 seconds. Note your own results might vary with the difference in timings.

Kris Rice blogged about the features and fixes available in this EA, I just quickly tested out the XLS upload (turns it into an APEX collection….very very cool!).

The APEX Listener is definitely introducing some interesting features (like the ability to record and replay sessions), so it will very soon outshine the Oracle HTTP Server in my opinion.

The answer is usually ‘yes I’d love to’, since -

1) I love to talk about and promote APEX
2) I love to travel

So, if you have an Oracle usergroup and you’re interested in promoting Oracle Application Express or looking for an APEX speaker, please feel free to drop me an email (john@sumneva.com) and hopefully we can arrange something.

With that in mind, the first public training I’m doing this year is 3 day Introduction to Oracle Application Express in Manchester on the 22nd March.

Here’s an except from the course description -

Introduction To Oracle APEX I

Introduction to Oracle Application Express is designed for those who have little to no interaction with APEX or for those looking to formalize their understanding of APEX. This course starts out with a review of the architecture of APEX, how it works, and some example projects that have been built with APEX. It then describes the Utilities and SQL Workshop facilities of APEX. The bulk of the class is spent on the application builder. Most of the core components are covered in Intro to APEX I, with hands-on exercises used to reinforce the concepts.

Numbers are strictly limited for this course, which you can book online now by clicking on the calendar.

Two and half years ago, I formed APEX Evangelists together with Dimitri Gielis because we both had a passion about Oracle Application Express. For the last two and half years we have been extremely successful providing training and consultancy around the Oracle Application express product. In the United States, Scott Spendolini and Doug Gault have walked a parallel path.

After many discussions between the four of us, we all decided it made sense to combine our respective talents to create Sumneva (the name being a combination of the company names).

I’ll post more information soon about the company and our ideas, but right now we’re all very excited about the plans and possibilities ahead of us, with 4 of the top world experts in Oracle Application Express our combined experience and talents will give our clients even better service.

My first impressions are very positive indeed. Whilst I’d seen quite a lot of it already, it’s not quite the same as being able to use it yourself.

Perhaps one of the first immediate changes, is the newly designed Application Builder look and feel

APEX 4.0 also has a new Tree View in the Application Builder, which is very different to the ‘traditional’ view which has remained more or less the same across all versions of APEX.

I must admit, when I first heard about and saw this feature (Patrick Wolf very kindly showed me a preview of it at a conference) I had my reservations about it. In fact I think I might have even said to Patrick that I wasn’t a fan of it. However, I now take that back. Now that I’ve used it (rather than just seeing it), I can see how it’s going to not only speed up my productivity (fewer clicks etc), but it is also going to be a big help to people new to APEX to understand the structure and flow of a page.

I’ve also had a play around with Dynamic Actions (which is really going to level the playing field for people who are not comfortable with custom javascript coding) and also Plugins to a limited degree. Both of these features are real game changers for APEX since they suddenly make the ‘sure we can do that, but it will take a while’ into ‘sure we can do that, there you go’.

I’m drilling down into a more detailed look at the features (future blog posts), but for now I have to say…I’m very very impressed (but then I already knew I would be).

As you’d expect it does everything you’d want from a browser, I’m pretty much a Firefox/Safari person (they complement each other pretty well particularly with APEX development).

If however you’re coming from an IE background, then Chrome adds a lot of nice features, for example I tried out Chrome against one of my local APEX instances and I had to say it *felt* pretty fast and I didn’t find anything in APEX that didn’t work in Chrome (which is pretty impressive considering APEX isn’t certified against Chrome yet, but I have heard that is something the Oracle APEX development team are looking at for the future).

Even SQL Workshop (which I’ve had a few issues in Safari with seemed to work nicely).

From a development perspective there are some very nice features, such as the “Developers Tools” panel which shows you detailed information about the pages you’re browsing. For example, it is now very easy to determine how long it takes to download and render all the resources in your webpage, or whether a particular Javascript routine is responsible for consuming all the time etc.

If you’ve already used Safari then most of this will already be familiar to you (as it appears to be the same code from WebKit). I’ve been using this functionality in Safari, together with the YSlow and Firebug plugins for Firefox to get the same information, but it’s always nice to have more options.

Usually at most conferences I attend Dimitri Gielis and myself try to arrange a very informal gathering of APEX-minded people in a pub. It’s a great time to network with old and new friends away from the conference halls, particularly if it’s one of your first events.

So, if you’re interested in attending an APEX Meetup, add a comment below (including your preferred night) and we’ll try and sort out an evening which suits most people.

However, how many people really use an 800×600 resolution these days? Out of interest I took a look at the statistics from my blog (the one you’re reading right now) for yesterday (I posted a new post yesterday so it would be interesting to see the stats for the new views just for yesterday rather than over the entire lifetime of the site).

So, whilst I’m not trying to pretend that these figures are representative of any other sites out there. They highlight a couple of things to me.

1) Nobody visited my site using 800×600 resolution.

2) The smallest resolution was 320×396 (probably an iPhone or other mobile device). After that the next smallest was 1024×768.

3) There are quite a lot of variations in browser resolution, more than I expected really.

So, what’s my point here? Well…I don’t really have one….other than -

1) Needing to support 800×600 is very different to designing the site in 800×600. In other words it’s no bad thing to have your site gracefully degrade to working nicely in 800×600, but I certainly wouldn’t make the starting design 800×600, otherwise you’d spending effort where there seems to be least reward.

2) 1024×768 is now becoming over-shadowed by higher resolutions. Most laptops these days support higher resolutions than that, let alone desktops. Most developers I know have multiple monitors too.

So, don’t be afraid of challenging the requirement of “We must design for 800×600″…how many people are *really* using it? Let’s stamp it out now…

Today, I received an email from someone who attended that presentation. I have asked their permission to post their message -

“I just thought I would let you know of the performance improvements I have got since your presentation yesterday”

and what were those improvements?

“The updateable grid has always been slow to run and was taking 1min 2seconds ( I couldn’t think how to improve it without completely redesigning it), with the changes this has been reduced to 6 seconds.”

and the difference in the production system -

That’s quite an improvement from the users perspective, something which used to take over 30 seconds is now well under a second.

From my perspective, it’s great to know that some people were inspired to use what they learned to make some significant changes which will ultimately make their end users happier (and we’re all end users of one system or another right?).

At the UKOUG conference in a few weeks time, I’m actually going to be presenting a 2 hour Masterclass on the same topic (which allows me to go into more depth, more demos and more aspects than an hour allows).

Secondly, time flies…UKOUG is right around the corner now (30th November – 2nd December). There are lots of great APEX sessions planned -

If you didn’t spot it, I’m doing 2 sessions, Dispelling Myths about Application Express and also a 2 hour masterclass on Building Scalable Applications with Oracle Application Express (which I believe should be interesting to *anyone* developing with APEX, not just those people who are developing large scale applications).

We are also going to hold an APEX round-table on Tuesday (1st December) at 12:55-13:55, where you can come along and ask any (APEX related!) questions you have.

It’s interesting to see that Oracle are taking this event seriously too, with David Peake, Anthony Rayner and Hilary Farrell from the APEX Development team all attending (and presenting).

Now, I just need to book my hotel!

Whilst I’m still on the fence about the whole ‘certification thing’, I do see this as another commitment by Oracle to make Application Express a serious contender amongst some of the more publicised (shall we say) development tool offerings they have.

You can get directly to the pilot page here (but I advise reading Joel’s blog first).