In the last article, we created a simple APEX application fetching data about English football from the Football Web Pages site which provides an authenticated REST API.

However, all I really want to do is to quickly look at Kingstonian's forthcoming fixtures for the next month. Fortunately, there is an FWP API providing that information.

Endpoint: https://football-web-pages1.p.rapidapi.com/fixtures-results.json

• Matches - The current list of matches for a competition/team
• The following parameters may be set:
  • comp - The ID of the competition (note: one of "comp" or "team" is required)
  • team - The ID of the team (note: one of "comp" or "team" is required)

We already have created an APEX report listing all the available Competitions (including the numeric ID values) so it would be useful to have a similar report listing all the Teams.

Teams¶

Endpoint: https://football-web-pages1.p.rapidapi.com/teams.json

• Teams - A list of the teams covered
• The following parameters may be set:
  • comp - The ID of the competition

Create a new REST Data Source called 'FWP Teams' with this endpoint using the 'Football Web Pages' authentication method.

Create a new page 'Teams' with an Interactive Report fetching data from this REST Data Source.

Search for 'Kingstonian' to find out the numeric identifier for this club (236)

Fixtures¶

REST Data Source¶

The Fixtures API can take a query parameter which is either 'competition' or 'team'.

Thus far, we have only used a 'Header' to supply the credentials to access the API.

API query parameters are name-value pairs introduced by a '?'. This is analogous to database queries which filter the results returned by a SQL query in the 'WHERE' clause.

select *
from emp
where name = 'JONES';

The APEX fixtures page should support both of these API query parameters ('competition' and 'team') and enforce that either one or the other is supplied.

Create a REST Data Source for 'Fixtures'

• Rest Source Name: FWP-Fixtures
• Endpoint: https://football-web-pages1.p.rapidapi.com/fixtures-results.json
• Authentication: Football Web Pages

If you click 'Discover', APEX will return an error

ORA-20987: No data found in uploaded file.

This is because the REST Data Source is expecting a parameter and we have not specified one. To fix this, click 'Advanced' and add the required parameter.

• Parameter Type: Query String variable
• Parameter Name: team
• Value: 236 (Kingstonian)

Leave the HTTP method as the default (GET) and click 'Discover'. APEX should display a list of Kingstonian's results and forthcoming fixtures.

Add a second query parameter named 'comp' with no value.

Apply the changes to this REST Data Source.

Fixtures Page¶

Navigate back to Application Builder and create a blank page called 'Fixtures'.

Search region¶

Create a region named 'Search'. Change the template of the 'Search' region to 'Collapsible'. The region template is in the 'Appearance' section.

Create two page items on the Search region

1. Competition
2. Team (uncheck 'Start New Row')

Create a button named 'Go' with the default action of 'Submit Page'. Check the 'Hot' checkbox under 'Appearance'.

Report region¶

Create a new region called 'Fixtures'. This is simply an Interactive report based on the REST source 'FWP Fixtures'.

Run the page. Hopefully, your screen should vaguely resemble this.

Clearly, this is just a checkpoint and not fully functional yet but I like to 'release early and release often' (if only to myself).

Joking apart, this approach is actually useful to demonstrate to an end user what the APEX UI will look like and how you can use standard APEX functionality to search and filter within the Interactive report.

No need to spend days mocking up wire frames of the proposed UI. In APEX, you can present a meaningful prototype using APEX early on.

Plus you can endlessly argue over the title, size, placement and colour of the 'Go' button.

Return to the 'Fixtures' page in Application Builder. Expand the 'Fixtures' region. You will see that APEX has helpfully added a section called 'Parameters'. Expand this and you will see the two query parameters for this REST Data Source exposed here.

Edit the 'comp' parameter. Under 'Value', change the Type to 'Item' and select 'P4_COMPETITION' from the pop-up menu.

Similarly, change the value for the 'team' parameter to the APEX page item P4_TEAM.

Save and run the page. You get an error but this is expected as we have not supplied any APEX page parameters (yet).

Enter '236' into the Search region for 'Team' and click 'Go'. You will see Kingstonian's results and fixtures displayed.

Check this is actually working by changing the team to '1'. This will display results and fixtures for 'Arsenal' in the English Premier League.

This looks promising. APEX truly is a low code solution.

Now let's look at the 'Competition' parameter.

Enter '2' for 'Competition'. This is for the English Championship (neither Kingstonian nor Arsenal play in this league).

Nullify the 'Team' parameter and click 'Go'.

List of Values¶

This report is improving but needs more work. End users typically don't know that they have to enter '1' to get data for Arsenal. The 'Competition' and 'Team' parameters are clearly List of Values so we will implement that now.

Navigate to 'Shared Components' and add the following List of Values for 'Competitions'

• Name: Competitions
• Type: Dynamic
• Data Source: REST Data Source
• Rest Data Source: FWP Competitions
• Return Column: ID
• Display Column: FULL_NAME
• Default Sort: ID

Once an APEX application has a REST Data Source available, it is available to all components (LOV's, reports etc) - just like a conventional local database table.

Create a second LOV for 'Teams'

• Name: Teams
• Type: Dynamic
• Data Source: REST Data Source
• Rest Data Source: FWP Teams
• Return Column: ID
• Display Column: FULL_NAME
• Default Sort: FULL_NAME

Navigate to the 'Fixtures' page and change the page items to use the newly created LOV's.

Change the type of P4_COMPETITION to 'Popup Lov'

Under 'List of Values', select

• Type: Shared Component
• List of Values: FWP-Competitions
• Display Extra Values: Unchecked
• Display Null Value: Checked
• Null Display Value: - Select -

Repeat this process for the P4_TEAM page item using FWP-TEAMS as the LOV.

Run the page. This looks better. Now we can select a Competition and a Team correctly.

You gleefully share your V2 prototype with a colleague for peer code review and her feedback is as follows:

1. When clicking 'Fixtures', I get 'ORA-20999: REST Data Source returned an HTTP error: HTTP 400: Bad request'
2. If you enter a Competition only, it works fine.
3. If you enter a Team only, it works fine.
4. If you enter both a 'Competition' and 'Team', the results look weird. Should 'Team' be a cascading LOV based on the 'Competition' ?
5. The column names and labels need tidying up. There are a lot of meaningless ID fields displayed.
6. It would be nice to have the option to review past results separately from fixtures in the future.
7. Performance - the Popup LOV's for Competition and Team are sluggish. Why are they so S L O W ?
8. The navigation menu looks chaotic and ugly.

In the next article, we will try to address this valid feedback.

Introduction

The last article provided a quick introduction to REST APIs. Now we will use a simple REST API to develop an APEX application using a real world example.

Football Web Pages

I enjoy watching football (soccer). My local team are Kingstonian FC, a non-league team in South West London. Kingstonian play in the seventh tier of English football. Kingstonian's players are semi-professional so the players hold down jobs and train and play part-time.

Football Web Pages (FWP) is an excellent site for all things related to football. The site includes news, fixtures, results for all English leagues (including non league) and the European leagues. I recently noticed FWP provides a REST API.

FWP API¶

Reviewing the FWP API, the first thing to note is whether the API is public (i.e. freely available) which it is and whether it requires authentication (it does).

To access our data you must subscribe to one of our pricing plans (which include a free plan) via Rapid API at the following address:

rapidapi.com/football-web-pages1-football-web-pages-default/api/football-web-pages1

Authentication

When you subscribe via Rapid API you will be given a key, and you must provide this in a header named "X-RapidAPI-Key" with every request.

A lot of API's provided by larger sites offer a facility to issue API calls directly on the site. This enables the developer to examine the specification of the API and experiment with different headers, query parameters and examine the response data in various formats.

FWP doesn't offer this functionality but it's a relatively simple API so we can use Insomnia to experiment with the API.

Normally, I choose the simplest API available - one with no query parameters or headers (other than required for authentication).

For FWP, the 'Competitions' API looks like a decent candidate

Competitions

A list of the competitions covered

The following parameters may be set:

include: One or both of: rounds, teams (default: neither)\ Endpoint: https://football-web-pages1.p.rapidapi.com/competitions.json

I'm lazy and so are you so you just enter this endpoint into Firefox. You are thwarted.

The FWP REST API does indeed require authentication so we need Insomnia.

Firstly, we create a folder to store all our FWP API requests. Name the folder 'Football Web Pages'.

Select the newly created folder and click 'click to add first request'.

Double click on the 'New HTTP Request' on the panel on the LHS. Rename this request to 'Competitions'.

Now enter the FWP API endpoint into the GET section in the middle panel. The endpoint (URL) is:

https://football-web-pages1.p.rapidapi.com/competitions.json

Click 'Send'. You get the same authentication error. You feel thwarted and disappointed but this is OK. You haven't provided your credentials yet but the endpoint is correct and the FWP server correctly responded with a '401 - Unauthorized' error.

This API requires that the API key (password, credentials) are supplied in the 'Header' of the API request.

Click on the 'Header' tab in the middle section

Add 'X-RapidAPI-Key' as the 'New Header'. Then add your private API key as the 'Value'. Remember that API headers are simply Name-Value pairs.

Click 'Send'. There is no need to explicitly save the changes to the Headers.

Finally. Success !

Look at the results in the panel on the RHS.

The API request returned a status of '200' (success). The elapsed time for the API request was 213 milliseconds and returned 10KB of data.

FWP APEX application

This demo was created and tested on Oracle's AlwaysFree tier. However, it should also work fine on Oracle's hosted APEX service on apex.oracle.com or a local APEX instance.

Navigate to App Builder

• Click 'Create a new App'
• Click 'New Application'
• Name the application 'Football Web Pages'
• Accept all the default options.

First, we need to configure the Web credentials in APEX to access the FWP REST API's

In APEX, Web Credentials are shared across the workspace. Click 'App Builder - Workspace Utilities - All Workspace Utilities'

Click 'Web Credentials'

Click 'Create'

Enter the following values for the parameters

• Name: Football Web Pages
• Static Identifier: FWP
• Authentication Type: HTTP header
• Credential Name: X-RapidAPI-Key
• Credential Secret: secretapikey
• Comments - FWP API key added on 16 October 2022

The reason I always add the comments field is that many API keys have a limited lifetime (6 months or a year) for security reasons. Often it is useful to know when the client secret was created.

Click 'Create' to save the changes

Next, create a REST data source for the FWP REST API

Navigate to 'App Builder' and click 'Shared Components'.

In the bottom left section, click 'REST Data Sources'.

Click 'Create'

Select 'From scratch' for 'Create REST Data Source' and click 'Next'

Leave the default of 'Simple HTTP' for the value of 'REST Data Source Type'

Enter 'FWP-Competitions' for the 'Name'.

Enter 'https://football-web-pages1.p.rapidapi.com/competitions.json' for 'URL Endpoint'

Leave the optional parameter 'HTTPS Host Name' blank.

Click 'Next'

Leave 'Create New' for the 'Remote Server' parameter

Accept the values helpfully supplied by APEX for 'Base URL' and 'Service URL Path'.

Click 'Next'

Accept the default of 'No Pagination' for 'Pagination Type'.

Click 'Next'

Ensure 'Authentication Required' is checked and select 'Football Web Pages' from the drop-down menu for Credentials.

Click 'Discover'.

APEX has helpfully sent this API request to the FWP server using the Web credentials and provided us with a preview of the data set returned so we can check it looks correct.

Wizards might want to click 'More Detail' but this looks good enough for us to just click 'Create REST Data Source'.

Now we have defined Web credentials and created a REST data source, let's finally create an APEX page displaying the Competitions.

Navigate back to 'App Builder' and select the 'Football Web Pages' application.

Click 'Create Page' and 'Interactive Report' from the Page Wizard.

Click 'Next'

Enter 'Competitions' for the name of the new page.

Under 'Data Source', select 'REST Data Source' and select 'FWP Competitions' from the drop-down menu.

Click 'Create Page'

Run the 'Competitions' page

Summary

That took a while but we have created an APEX application that fetches data from a REST Data Source that requires authentication.

These are valuable building blocks to refine and extend this APEX application when we explore a range of different API's.

Background

REST APIs are a popular means of manipulating data. REST APIs use a client-server model. The server is a web server and the client is a Web application or a Python, Perl, Java, .NET, Node.js or COBOL program.

REST is an abbreviation for 'Representational State transfer' while API is another abbreviation for 'Application Program Interface'.

This all sounds complicated and almost intimidating but it's not. Database developers have been manipulating data using a client (SQL*Plus) from a server (Oracle database) for many years.

A REST API call is simply an HTTP request. This HTTP request normally reads data but can also insert, update and delete data.

REST request

A REST request has four parts

Method¶

The HTTP method is required and indicates the operation to be performed. The most common types of an HTTP request are:

• GET (Read)
• POST (Write)
• PUT/PATCH (Update)
• DELETE (Delete)

Endpoint¶

The endpoint is the full URL sent to the Web server. For example, this endpoint lists all the feeds available on GitHub.

https://api.github.com/feeds

Click on the link. Data is displayed. It's not like a conventional Web site but it's data.

The endpoint is made up of four elements:

• root-endpoint - https://api.github.com/
• path - feeds
• variables - optional variables introduced with a colon (e.g. ':user')
• query parameters - optional set of name=value pairs separated with a ampersand (e.g. '?query1=value1&query2=value2'). Query parameters are used to filter the data returned by limiting the number of records and/or using search criteria.

Headers¶

Headers are optional and used to supply additional information (credentials, parameters).

Body¶

Data in the body section is also optional and normally used for a POST, PUT or DELETE requests to insert, update or delete data.

REST response

A REST request will receive a response. This is the data requested or an error code.

The response is usually a data set in JSON format (but could be XML or even an image).

The elements of the response are:

Response Code¶

The HTTP status code indicates the success or failure of the request. A non-exhaustive list of common status codes are:

• HTTP/1.1 200-299 OK - success
  • 200 OK
  • 201 Created
  • 202 Accepted
• HTTP/1.1 300-399 OK - redirect
  • 301 Moved permanently
  • 302 Moved temporarily
  • 304 Not Modified
• HTTP/1.1 400-499 client error
  • 400 Bad Request
  • 401 Unauthorized
  • 403 Forbidden
  • 404 Not Found
  • 405 Method Not allowed
• HTTP/1.1 500-599 server error
  • 500 Internal Server Error
  • 501 Not Implemented
  • 502 Bad Gateway

Content-Type¶

Indicates the format of the returned data. For example, for JSON data, this is set to

content-type: application/json; charset=utf-8

Content Length¶

Indicates the size (in bytes) of the returned data.

where to find REST API's

Many sites offers API's to access data. Twitter, GitHub, YouTube, IMDb, Apple Music are popular examples.

These sites often require developers to register and obtain an API key (password) in order to access the service. This is normally free and allows the provider to police the service and guard against denial of service attacks (flooding the server with requests in an effort to bring it down).

Governments and public sector bodies often provide excellent sources of public, freely accessible REST API's. Some of the subject matter may be a little dry; statistics about Coronavirus, trade quotas, water quality, crime.

Programmable Web includes a searchable directory of API's.

RapidAPI also offers pointers to several API's for UK data.

Hello world

The first program I create when learning a new language is always 'Hello World!'. This is useful as it teaches you the basic syntax, how to compile or deploy the program and it is easy to check it works successfully.

The National Health Service (UK) helpfully makes a REST API available that returns 'Hello World!'

This endpoint does not require authentication and simply returns 'Hello World!'.

https://sandbox.api.service.nhs.uk/hello-world/hello/world

There are many REST API clients available and we will use a GUI REST client called Insomnia for this tutorial.

Download and install Insomnia which is available for Linux, Windows and MacOS.

Enter the endpoint into the field labelled 'GET' in the middle panel and click 'Send'.

The right hand panel contains the response from this request. The data returned is in JSON format and contains a single entry containing 'Hello World!'

The other important fields are the status code of '200 OK' which is displayed in Green and represents a successful call for the GET request.

The elapsed time for this request is quick (164 ms).

Finally, the size of the returned data is '26B' (26 bytes).

Summary

Congratulations ! You have submitted your first REST API request. It's time to put the kettle on and update your CV.

In the next article, we will build on this knowledge by using REST to create a more useful, real-life scenario.

a brief history of blogging

I have maintained a blog, on and off, for a long time (since 2005). During that time I have used a wide variety of blogging platforms (Blogger, WordPress, Typepad, Drupal, Tumblr, Django, Posterous, Jekyll, Ghost, Nikola, Hugo)

My blog was a personal blog. Looking back, some posts were essentially micro-blogging (trite one-liners), link blogging (interesting, amusing BBC news stories), endless analysis of Manchester United together with some longer form articles.

Hardly any of my content was technical despite the fact I was an IT consultant. With hindsight, there were a few reasons for this:

• I worked all day staring at a screen working on technical issues. I also travelled a lot in the UK and Europe so when I finally arrived home, my immediate thought wasn't always 'I really need get my laptop out and blog about that database performance issue'.

• 'Imposter syndrome' - whatever topic I thought of, someone, somewhere at some time would have already blogged about the same topic (normally Tim Hall) - better and more intelligently than I ever could. So who would ever read my post and, moreover, what was the point ?

• 'Laziness' - I am inherently lazy. I freely admit it. It's not necessarily a bad thing. In fact, I think most developers should be lazy (think scripting, think VM's). A technical post takes time because, to have any value, it has to be accurate. Also, it probably needs to include screenshots. Taking a series of screenshots and posting images to a blog is a very time consuming and tiresome exercise. It will also pose a significant issue for the imminent migration to the next blogging platform. A technical post requires much more time and effort than posting about Manchester United's latest victory or the film I saw at the weekend.

• Work. To be honest, this was a minor issue but often I had a nagging doubt that if I had some useful technical knowledge to share, then perhaps I should share this with my colleagues who worked on similar technical issues rather than chasing page views on the Internet. There was also the issue of anonymity; obviously I could mask my identity, my employer's name and the customer name but was this ethical ? Did it breach the corporate social media policy' ?

• Separation - I did occasionally conquer these various self-imposed mental barriers and post a vaguely technical post about Siebel CRM or Oracle. However, while the minuscule element of my readership who were interested might have commented 'At last, a technical post !', I imagined the wider audience (the other six readers) scratching their heads saying 'Well, where's the joke here ? What did he have for breakfast ? Screw that, I'm unsubscribing'.

Hashnode

Clearly, I could have overcome all of these issues by maintaining two separate blogs; a personal blog and a technical blog but, like I said, I'm lazy.

I have a love hate relationship with Twitter. I really dislike the adverts and inserted content and for me, it represents a very dangerous distraction and potential time-sink.

However, a lot of the wonderful APEX community use this platform so I was forced to sign up for the 17th time purely to follow the APEX folks who post valuable content (tagged #orclapex) and freely share their knowledge and expertise.

One of my favourite APEX bloggers, Jon Dixon, indirectly drew my attention to Hashnode.

My private email message to Jon (reproduced here without permission) sums up my initial thoughts on Hashnode

Thanks for inadvertently pointing me at hashnode. I was completely unaware of this platform but I like it as it's Markdown, hooks into GitHub and has a community.

This was useful as I always wanted a technical blog that was separate from my inane stream of consciousness that I post elsewhere.

However, a technical post does take a good deal of time (checking, testing, screenshots, iterating) but ultimately is satisfying I think.

I decided to dip my toe in the water with a post I originally posted on an internal Confluence Wiki.

This was an interesting exercise in itself. My original article was rather rushed and missed out crucial steps (grants on a package).

As I was forced to revisit the original post and review each step from scratch in a clean, vanilla environment, guess what, things didn't work as I described.

Another benefit was that Jon posted a comment saying 'Thanks' which was appreciated and also privately emailed me with a couple of suggestions for minor improvements.

That, to me, is the whole point of a development blog - to hopefully share useful technical knowledge with others but also to have a peer review and learn something new yourself.