More Adventures in Pre-Filtering Customer Lookup Fields (Dynamics CRM/Dynamics 365 for Enterprise)

When working with form-level JScript functionality on Dynamics CRM/Dynamics 365 for Enterprise (D365E), you often uncover some interesting pieces of exposed functionality that can be utilised neatly for a specific business scenario. I did a blog post last year on arguably one of the best of these functions when working with Lookup field controls - the Xrm.Page.getControl().addPreSearch method. Similar to other methods exposed via the SDK, its prudent and effective implementation can greatly reduce the amount of steps/clicks that are involved when populating Entity forms.

I’ve already covered as part of last years post just what this method does, its sister method, addCustomFilter, and also some of the interesting problems that are encountered when working with the Customer lookup field type; a special, recently introduced field type that allows you to create a multi-entity lookup/relationship onto the Account and Contact entities on one field. I was doing some work again recently using these method(s) in the exact same conditions, and again came across some interesting quirks that are useful to know when determining whether the utilisation of these SDK methods is a journey worth starting in the first place. Without much further ado, here are two¬†additional scenarios that involve utilising these methods and the “lessons learned” from each:

Pre-Filtering the Customer Lookup to return Account or Contact Records Only

Now, your first assumption with this may be that, if you wanted your lookup control to only return one of the above entity types, then surely it would be more straightforward to just setup a dedicated 1:N relationship between your corresponding entity types to achieve this? The benefits of this seem to be pretty clear - this is a no-code solution that, with a bit of ingenious use of Business Rules/Workflows, could be implemented in a way that the user never even suspects what is taking place (e.g. Business Rule to hide the corresponding Account/Contact lookup field if the other one contains a value). However, assume one (or all) of the following:

  • You are working with an existing System entity (e.g. Quote, Opportunity) that already has the Customer lookup field defined. This would, therefore, mean you would have to implement duplicate schema changes to your Entity to accommodate your scenario, a potential no-no from a best practice point of view.
  • Your entity in question already has a significant amount of custom fields, totalling more than 200-300 in total. Additional performance overheads may occur if you were to then choose to create two separate lookup fields as opposed to one.
  • The entity you are customising already has a Customer lookup field built in, which is populated with data across hundreds, maybe thousands, of records within the application. Attempting to implement two separate lookups and then going through the exercise of updating¬†every record to populate the correct lookup field could take many hours to complete and also have unexpected knock-on effects across the application.

In these instances, it may make more practical sense to implement a small JScript function to conditionally alter how the Customer Lookup field allows the user to populate records when working on the form. The benefit of this being is that you can take advantage of the multi-entity capabilities that this field type was designed for, and also enforce the integrity of your business logic/requirements on the applications form layer.

To that end, what you can look at doing is applying a custom FetchXML snippet that prevents either Account or Contact records from returning when a user clicks on the control. Paradoxically, this is not done by, as I first assumed, using the following snippet:

[snippet id=“276”]

This will lead to no records returning on your lookup control. Rather, you will need to filter the opposite way - only return Contact records where the contactid equals Null i.e. the record does not exist:

[snippet id=“277”]

Don’t Try and Pass Parameters to your addCustomFilter Function (CRM 2016 Update 1)

If your organisation is currently on Dynamics CRM 2016 Update 1, then you may encounter a strange - and from what I can gather, unresolvable - issue if you are working with multiple, parameterised functions in this scenario. To explain further, let’s assume you have a Customer Lookup and a Contact Lookup field on your form. You want to filter the Contact Lookup field to only return Contacts that are associated with the Account populated on the Customer Lookup. Assume that there is already a mechanism in place to ensure that the Customer lookup will always have an Account record populated within it, and your functions to use in this specific scenario may look something like this:

[snippet id=“278”]

The above example is a perfectly sensible means of implementing this. Because, surely, it make more practical sense to only obtain the ID of our Customer Lookup field in one place and then pass this along to any subsequent functions? The problem is that CRM 2016 Update 1 throws some rather cryptic errors in the developer console when attempting to execute the code, and does nothing on the form itself:

Yet, when we re-write our functions as follows, explicitly obtaining our Customer ID on two occasions, this runs as we’d expect with no error:

[snippet id=“279”]

I’ve been scratching my head at why this doesn’t work, and the only thing I can think of is that the first function - main - would be executed as part of the forms OnLoad event, whereas the filterContactNameLookup is only triggered at the point in which the lookup control is selected. It’s therefore highly possible that the first instance of the customerID is unobtainable by the platform at this stage, meaning that you have to get the value again each time the lookup control is interacted with. If anyone else can figure out what’s going on here or confirm whether this is a bug or not with Dynamics CRM 2016 Update 1, then do please let me know in the comments below.

Conclusions or Wot I Think

It could be argued quite strongly that the examples shown here in this article have little or no use practical use if you are approaching your CRM/D365E implementation from a purely functional point of view. Going back to my earlier example, it is surely a lot less hassle and error-prone to implement a solution using a mix of out of the box functionality within the application. The problem that you eventually may find with this is that the solution becomes so cumbersome and, frankly, undecipherable when someone is coming into your system cold. With anything, there always a balance should be striven for on all occasions and, with a bit of practical knowledge of how to write JScript functionality (something that any would-be CRM expert should have stored in their arsenal), you can put together a solution that is relatively clean from a coding point of view, but also benefits from utilising some great functionality built-in to the application.

comments powered by Disqus