The Pre-Production Release of APEX 20.2 is now available on apex.oracle.com!   There are lots of new and useful features that we can't wait to start using.   In this article, we'll have a look at the new REST Data Source Synchronization feature will be super handy when using APEX to extend ERP Cloud.   We gave this a runout on apex.oracle.com, using it to pull down ERP Cloud Organizations into an APEX table on a set schedule.

We configured the REST Data Source as below, using APEX Web Credentials to connect to the Vision Demo ZRNS environment, configuring the organizations HCM rest service to return the following fields for all active departments:  OrganizationId, Name and Active .

Query String variables are used to configure the selection criteria and data returned.

Before attempting synchronization, we test the web service call to make sure it is functioning correctly.

We can check this for accuracy against the source data in ERP Cloud.   

Note, Vision is shown  below with the Redwood theme; we're looking forward to trying out the new APEX 20.2 Redwood Light Theme for our ERP Cloud Integrations.

An important step is to identify the Primary Key of the returned data; this is used if you're Merging the data every time it is synchronized.

We clicked on the new Manage Synchronization link to start the configuration process.

Here we choose a target table.   

In the next step, we have to choose the Synchronization Type:
​

• Append - rows are added to the end of the table.
• Merge - rows are merged using the primary key defined earlier.
• Replace - the table is truncated (or deleted) before rows are inserted.

The Synchronization Schedule is expressed using same syntax as DBMS_SCHEDULER, so tables can be synchronized every 15 mins, 2 hours, day etc.  

All done - now we can save the Schedule and we're told when this will run.   

​We can also run it On Demand using the Save and Run button.

Looking at the new APEX_REST_SOURCE_SYNC package, we should  also be able to start synchronization programmatically using a PL/SQL procedure call.

Once Synchronization has run, the log entries give us rowcounts, message sizes, and execution times.

Data is now available in our custom table.   We used the APEX 20.2 Declarative Cards layout to show the data.

And just for fun, this is what it looks like with the Redwood theme:

We can also use the new APEX 20.2 Excel downloads to get our newly synchronized data:

Behind the scenes, the JMJ_ERP_CLOUD_DEPARTMENTS table contains the columns we requested, along with a couple that APEX uses to keep track of synchronization.

This was our first look at this new feature in APEX 20.2 and in about 15 minutes we were able to create a new workspace, ERP Cloud data source, and synchronize data into a custom table.   

​We didn't have to write a single line of code to call web services, map returned data into tables, or record web service calls and responses.  

This is fantastic functionality, and we look forward it using it in our work using APEX to extend and integrate with ERP Cloud.    We hope ERP Cloud follows the industry and continues to migrate old SOAP services to REST.  This will make integrations like this much simpler and help customers gain the maximum benefit from their ERP Cloud Investment.

This is yet another example of expectations being exceeded by the APEX Development Team!   Thanks to all those who worked hard behind the scenes to make this happen.

​With the evolution of ERP Cloud and Oracle Cloud Infrastructure (OCI), we are starting to see ERP Cloud utilize OCI features. This is a positive move and it will allow ERP Cloud customers to gain more value from their ERP Cloud investment. One area where we have seen this partnership blossom is with ERP Cloud BI Cloud Connector (BICC) and OCI Object Storage. In this post I will cover how you can leverage more from this integration using OCI Events and Notifications.

​Oracle BI Cloud Connector (BICC) provides a robust mechanism for performing bulk exports of data from ERP Cloud. Exported CSV files can be directed to Universal Content Management (UCM) (contained within the ERP Cloud environment), or to an OCI Object Storage Bucket. BICC exports are scheduled through the ERP Cloud Enterprise Scheduler Service (ESS). For those of you familiar with e-Business Suite, this is like the Concurrent Manager. So, how do we know when a scheduled job has finished so we can go fetch the file? This is the problem we will address in this post.
 
Although the example below describes a scenario using ERP Cloud, this pattern can be re-used for any scenario where you need to perform an action when a file is created in an OCI bucket, but you don’t know when the file will be created.

Oracle Cloud Infrastructure includes two integration tools that can help us with this problem. These are Events and Notifications.
 
Events Service
Events fire when certain things happen on your OCI infrastructure. In our case we want to launch an event whenever a file is created in a specific OCI Object Storage bucket. An event on its own, however, is not very useful. We need the Event to do something and this is where Notifications come in.
 
Notifications Service
A Notification can perform one of the following operations:

• Email
• Function (Cloud Based Function e.g. Java)
• HTTPS (Custom URL)
• PagerDuty
• Slack

 
The key notification for us is ‘HTTPS (Custom URL)’. This notification type is further described in section 3 below. When this type of Notification fires, it posts a JSON payload to the HTTPS Endpoint that you define. The content of the payload is specific to the OCI service that causes the Event to fire. In our case, it will contain information about the file that was created.

1 - File from ERP Cloud
In our example this file is coming from BI Cloud Connector (BICC) on ERP Cloud. As I mentioned earlier, BICC allows you to perform high volume extracts of data from ERP Cloud to either local Universal Content Manager (UCM) or an OCI Object Store Bucket. We will be using OCI Object store as this gives us more options and easier access to the files once created. BICC allows you to either call a web service to launch an extract job or to use ERP Cloud to schedule the extract. In view of this, files could appear in our OCI Object Store Bucket at any time. We need to know when the file is created so we can go and fetch it.
 
Object Store Bucket
You will need to create an Object Store bucket to receive your file. When creating your Object Store bucket, be sure the ‘Emit Object Events’ option is checked. This is what allows us to subscribe to events on the bucket.

2 - Event
As soon as the file is created in our object store bucket, the event fires. Events consist of an Event Type (this describes the action that should trigger the event) and an Event Action (what you want to happen when the event occurs). In our case the event type is 'Object - Create' and the Event Action is to call a Notification.

3 - Notification
​The Notification is the operation we want to perform when our event fires. We are using the ‘HTTPS (Custom URL)’ operation. Notifications consist of Topics which allow us to categorize our notifications. For each Topic, we can create multiple Subscriptions. This allows us to let multiple systems know when an Event occurs. In this example, we are referencing our ORDS Endpoint.

If your ORDS end point is unavailable when it is called by the Notification, OCI will re-try to post to your service for a period of up to 2 hours.

​The Endpoint service needs to have a POST handler to accept the payload from the Notification. In our case, our Endpoint is an ORDS REST service running on an APEX environment (see step 4 for details).

When you first create a notification, it will go into a status of ‘Pending confirmation’.

At the time you create the Notification, OCI sends a payload to your Endpoint containing a confirmation URL. The confirmation step ensures that you have control over the Endpoint that you define in your Notification.

4 – ORDS Service
The ORDS web service definition is pretty simple. Here is a screen shot (from SQLDeveloper) of the Module, Template and Handler for a service that can receive a payload from an OCI Notification.

The PL/SQL code you add to the handler can do anything you want. In our case it submits a Database Scheduler Job to fetch and process the file asynchronously.
 
5 – Process File
In our DB Scheduler Job, we use the OCI Object Store REST services to fetch the file from the OCI Object Store bucket. This connectivity between APEX and OCI was made much easier in APEX 19.2 when Web Credentials were enhanced to include the ‘Oracle Cloud Infrastructure (OCI)’ Authentication Type. This allows us to configure a native Web Credential for connecting to OCI Web Services. I recommend this Oracle A-Team Blog which describes in detail the OCI and APEX setups that are required to get this integration working.
 
After you have fetched the file (using APEX_WEB_SERVICE), you can then parse it (using APEX_DATA_PARSER) and process it as necessary, making the data available for your APEX Applications.

​I have been using APEX and ORDS on Amazon Web Services (AWS) for a number of years and until recently, I was not sure Oracle was ever going to compete. Oracle Cloud Infrastructure has come a long way in the past 12 months. While it still falls far short of AWS in terms of functionality, the functionality that is there is intuitive and works well. If you add great initiatives like ‘Always Free’ ATP, ORDS and APEX then you have a platform that is worth getting to know..
 
As APEX/ORDS developers and architects I encourage you to explore OCI. I also think we should start  to think differently about how we build our APEX solutions to utilize OCI where it makes sense. 

​By now, most people have heard about Robotic Process Automation (RPA) tools. While these tools come with the glow of the next new thing in technology, in my experience, ERP integration and automated scenarios are complex, and I am not yet convinced RPAs can handle them as well as the latest hype may lead one to think. I am convinced that APEX can and has handled many real-world automation and integration scenarios in the ERP world. In this post I will describe a real-world example where APEX automates and orchestrates a repetitive e-Business Suite process. ERP Cloud examples will be featured in future posts.

My whole career has been focused on building software to automate manual processes. Most of this automation work has been focused on Oracle’s e-Business Suite (EBS) and ERP Cloud. ERP software has a huge amount of functionality, but it is necessarily geared toward the most common business processes. In view of this, common processes like creating items must be broken down into their component steps, e.g., Create Item > Assign Categories to the Item > Choose Catalog Values for the Item > Create Item Cross References. If you are creating hundreds of items a week, this can be a very time-consuming (and error-prone) process. Of course, item creation is just an example. Depending on your business, there may be many opportunities for automating costly manual processes that occur in your ERP software. Given the hype around RPAs, one obvious solution is to deploy an RPA to solve your automation challenges. I am not so sure….

If you have ever used automated testing tools like Selenium or Cypress, then you have a head start in understanding RPA tools. RPAs automate steps that an end user performs via the UI. It repeats steps like button clicks, data entry into fields and page navigation in a web UI.
 
For example, if you need to onboard one new user, you might perform the following steps:

1. Login to the Application
2. Navigate to the User Entry page
3. Enter information about the new user
4. Click the Create User button
5. Navigate back to the User Entry page

 
To onboard 100 users you would repeat steps three to five 100 times while reading data from a file. This is where RPA steps in and performs this series of repetitive steps for you. Being able to perform automations at the UI level makes adoption relatively easy. There is no database to connect to, no code to write or APIs to learn. Essentially you don’t really need to involve IT!

In my opinion, RPAs fall short in two important areas when compared to an Low Code API solution like APEX.
​ 
Changing UI
RPAs rely heavily on a consistent UI. They depend on certain HTML tags being present or fields having specific IDs, Labels, etc.
 
The problem is, modern ERP solutions like ERP Cloud release new functionality on a quarterly basis. Releases often comes with changes to the UI which can break your RPA processes. A low code solution will typically call APIs (e.g. REST services) to perform automations. While APIs are not completely immune to change, there is the assumption of a contract between an API developer and its consumers. Developers break this contract at their peril.
 
Performance and Reliability
Because RPAs interact with the UI they rely on the stability and speed of the UI in order to work effectively. What if there is a 5 second delay between clicking the Save button and the next page appearing? While you can accommodate for the delay in an RPA script, it can have a huge impact on how long it takes to load high volumes of data. In addition, if a page fails to load completely, then your RPA script will fail. You are left figuring out where it failed and having to re-start it from the right spot.
 
APIs are typically more stable and always more performant than the UI they serve. When you call an API to create a record, you get a specific response from the API and can handle what happens next directly in your code. In addition, there are often API options to perform bulk loads of data where you can send thousands of records in one post.

License and Resource Costs
You should also consider that while it may not always require the specialized skills of a Developer, RPAs do still require skilled functional resources to keep them running smoothly. When comparing to APEX, you also have to consider the additional license costs that come with an RPA solution.

There are a few things you should consider before diving into any automated solution. The first is cost. You should weight the cost of developing an automated solution vs the cost of not doing it at all. You should also consider your ability to maintain the solution once it has been implemented. If you build automations based on complex integration software that requires highly specialized skills to maintain, it may end up costing you more (in consulting fees and monthly subscriptions) than the manual processes did.

APEX is a Low Code development platform that, assuming you have an Oracle database, requires no additional software licenses. If you are running EBS you can install APEX on your EBS database. If you are running ERP Cloud, you can stand up a PaaS AWS RDS Oracle instance with APEX for as little as $100/month.
 
We have already blogged that APEX is an ideal partner for both EBS and ERP Cloud. On EBS, APEX lives in the same database as EBS and can easily call EBS public APIs to automate EBS activities. In the cloud, APEX comes complete with numerous utilities to make calling web services easier (web source modules, credential stores, APEX_WEB_SERVICE APIs with automated token renewal, etc.). Used in conjunction with Oracle REST Data Services (ORDS), you can also create Webhooks to handle events from your ERP systems.

Now that we have decided on APEX to build your automation vs using an RPA, what might an APEX centric-solution look like? Let’s use the item creation example I mentioned at the start of this post to illustrate what is possible. In this use case, the goal is for a user to be able to create inventory items in EBS with as few manual steps as possible.
 
The manual steps required by a user to create an item include:

1. Login to EBS
2. Launch EBS java forms
3. Choose an Inventory Organization
4. Navigate to the Item Entry form
5. Find the item you want to copy from
6. Create a new Item as a copy of the item
7. Update the item number and description of the new item
8. Navigate to the item category form
9. Assign the correct item category to the item
10. Navigate to the Item Catalog form
11. Enter up to 5 catalog name catalog value combinations
12. Navigate to the Item Cross references form
13. Create an item cross reference for the new item
14. Repeat steps 4-13 X times

 
Here is what it looks like when we merge steps 4-13 into one APEX page.

APEX Automation Use Case
When users first navigate to the APEX page, they choose an inventory organization to copy items from and optionally enter an item prefix to filter source items they can choose from.
​

The user then adds source items to an interactive grid. Source items are items that already exist in EBS and used as the basis for creating new items in this example.

The target values are handled as follows:
-Item Number - Users can manually enter a target item number or click ‘Generate Item Numbers’ to automatically number the new items using an EBS Document Sequence.
-Item Template – Next the user chooses and item template. Items in EBS have hundreds of attributes. An item template allows you to use the template attributes to set the attributes of newly created items.
-Item Long Description – The user optionally enters a description of the item.
-Catalog Values – Catalogs allow you to assign sets of attributes to an item. The catalog attributes of the source item are brought in automatically and the just must update values for the new item.
-Cross Reference – Cross references allow you to cross reference your items to other entities e.g. Supplier Item Numbers. Finally, the user selects a cross reference type and enters a value.
 
Note: While this looks like a lot, all but two of the above fields are optional. There is even an option for users to upload items Via Excel to perform bulk loads.
 
An APEX collection is used to store the user entered data until the user is ready to initiate the copy process by clicking ‘Copy Items’.
 
The actual process automation i.e. copying the items, category assignments, catalog elements and cross references is performed by a PL/SQL package called from APEX. The PL/SQL package calls the related EBS Public APIs in sequence utilizing the data from the APEX collection for the required parameters.
 
EBS Public APIs used by this process:

• Item (ego_item_pub.process_items)
• Category (inv_item_category_pub.create_category_assignment)
• Catalogs (inv_item_catalog_elem_pub.process_item_descr_elements)
• Cross References (mtl_cross_references_pub.process_xref)

 
When the copy is complete the user is presented with the results of the copy. An added benefit of using APEX is that users can easily see which steps completed successfully and which ones failed.

There is a place for RPAs in performing repetitive tasks on moderate amounts of data on simple UIs that do not regularly change. ERPs rarely fit this mold. Data volumes are usually too high to justify a RPA versus an API due to an RPA’s slow performance compared to an API.  Additionally, change is built-in and assumed in the cloud. Inconsistent performance is not an option.
 
APEX answers many questions in the ERP world. In this post we have learned in this post that APEX can also be called on to build complex orchestrated automations both inside and outside of your firewall.

Watch this space for future posts geared towards automating manual processes in ERP Cloud using APEX.