In today’s beginner series article, we will look at how to create a basic entity such as Item, Customer, Vendor etc.
Creating basic entities is one of the core things anyone has to know. It is also necessary to know which properties must be defined (LookupPageID, CardPageID), which type of PageTypes should be used etc. After this video, you will be able to create the fundamental entities in your projects correctly and without any hesitations.
One of the news in AL Language that is coming with the upcoming major version (Business Central 2021 W1) in April is a new trigger AnAfterLookup. The new trigger is described on Microsoft Docs (however, the description is still a bit confusing…).
As the Lookups are often used, I already tried what’s new and how we can improve our projects with these changes. Unfortunately, there is no documentation available at the time of this article, so I am unsure what is intended and what is currently just a bug (as this is still only a part of the preview version).
Although I could use the triggers in some small examples, I still can not find any scenario where the OnAfterLookup trigger can be useful in the real world. Let’s look at some examples.
OnLookup & OnAfterLookup
The OnLookup trigger is known from the beginning of C/AL Language. It is used to build more complex Lookups that can not be created using TableRelation (or any similar property). It replaces standard lookup features (runs in a place of regular lookup feature).
Until the Roletailer Client (NAV2009RT or rather NAV 2013), there was no visual/functional difference between lookups created through any property or OnLookup trigger. However, with Roletailored client (and Web Client), there is a vast difference – a dropdown list available only for relations made through TableRelation property. The dropdown list is beneficial for users, and developers must have it in mind when creating relations.
Examples
OnAfterLookup trigger does not override the standard trigger from the TableRelation. It is triggered once the record is selected from the Dropdown box; however, if the user decides to use a standard lookup list using “Select from a full list”, the trigger is not fired.
With this in mind, it is unsurprising that both triggers written for the same field are useless (OnAfterLookup is not used as OnLookup does not use dropdowns). In the first example below, the OnAfterLookup trigger is never called.
If we only use the OnAfterLookup trigger, the trigger is fired once the user chooses the DropDown list value. As a developer, you get a RecordRef variable that contains the selected record.
The second example shows how you can print information about the selected record. However, I am still unsure how to use it in the real world – we can just get information from the selected record (and even only for those made through DropDown List, not the full list).
field("TKA Tariff No."; Rec."TKA Tariff No.")
{
trigger OnAfterLookup(Selected: RecordRef)
var
TariffNumber: Record "Tariff Number";
begin
Selected.SetTable(TariffNumber);
Message('Selected value: ' + TariffNumber."No.");
end;
}
Do you have any ideas on how to use this functionality within real projects? Have you ever used the Text parameter in the OnLookup trigger? Please share your thoughts in the comment section.
In today’s beginner series article, we will look at Snippets in AL Language. What are snippets? How can we use them? What is the reason to know how to use them/when to use them? All these questions are answered in the video below!
The first article from my new youtube channel is here (if you want to find out why I have decided to start my own youtube channel, see here)! This is the very first video for beginners – how to create a new AL project.
In this video, we will look at how to create a new AL project in Visual Studio Code, the core setting files in every project (app.json, launch.json) and what needs to be set up able to publish an extension to your own Business Central instance.
AL ID Range Manager is VS Code & Microsoft Dynamics 365 Business Central extensions made for developers who need to manage their object/field IDs across multiple projects (or in the multi-developers environment).
This is another article about Partial Records (as we already discussed earlier in the article about Partial Records and later in the first part of articles about partial record’s performance). In this article, we will explore the ration between performance, partial records and system with more tableextensions.
Let’s start with the example of our AL Code which we use for performance testing. This code is very similar to the previous article’s code; we only removed parts that were linked to the JIT loading example.
report 90001 "TKA Partial Records 2"
{
Caption = 'Partial Records 2';
UsageCategory = Administration;
ProcessingOnly = true;
ApplicationArea = All;
trigger OnPostReport()
var
GLEntry: Record "G/L Entry";
begin
ForEachGLEntries(GLEntry);
Clear(GLEntry);
GLEntry.SetLoadFields(Amount);
ForEachGLEntries(GLEntry);
end;
procedure ForEachGLEntries(var GLEntry: Record "G/L Entry"): Decimal;
var
SumTotal: Decimal;
Counter: Integer;
TempName: Text;
begin
if GLEntry.FindFirst() then
repeat
SumTotal += GLEntry.Amount;
Counter += 1;
until GLEntry.Next() < 1;
Message(TempName);
exit(SumTotal / Counter);
end;
}
The first example is trivial. We just run our AL example without any change on the empty Cronus database. Unsurprisingly, if we loop through all G/L Entry table records, generated SQL Command load all fields.
SELECT
"17"."timestamp",
"17"."Entry No_",
"17"."G_L Account No_",
... All Table Fields (56 fields hidden) ...
"17"."Last Modified DateTime"
"17"."$systemId",
"17"."$systemCreatedAt",
"17"."$systemCreatedBy",
"17"."$systemModifiedAt",
"17"."$systemModifiedBy"
FROM "CRONUS".dbo."CRONUS International Ltd_$G_L Entry$437dbf0e-84ff-417a-965d-ed2bb9650972" "17" WITH(READUNCOMMITTED)
ORDER BY "Entry No_" ASC OPTION(OPTIMIZE FOR UNKNOWN, FAST 50)
If we change our code to use SetLoadField method and specifies all fields that we need, the generated SQL Command is optimized. It contains fewer fields (= less data that needs to be transferred from a database server to an application server).
SELECT
"17"."timestamp",
"17"."Entry No_",
"17"."G_L Account No_",
"17"."Amount",
"17"."Bal_ Account Type",
"17"."Source Type",
"17"."FA Entry Type",
"17"."$systemId",
"17"."$systemCreatedAt",
"17"."$systemCreatedBy",
"17"."$systemModifiedAt",
"17"."$systemModifiedBy"
FROM "CRONUS".dbo."CRONUS International Ltd_$G_L Entry$437dbf0e-84ff-417a-965d-ed2bb9650972" "17" WITH(READUNCOMMITTED)
ORDER BY "Entry No_" ASC OPTION(OPTIMIZE FOR UNKNOWN, FAST 50)
Although it could look like a significant difference, from database server performance point of view, both queries are almost the same (the first one needs a bit more performance to load & process additional unnecessary fields).
HOWEVER, the biggest difference comes with the environment, which uses table extensions!
Table Extension impacts on performance.
Let’s imagine that we have an environment that has our extensions (that has one table extension on G/L Entry table with two Text[2048] fields) and also another 3 AL Extensions each with G/L Entry table extension each with two Text[2048] fields. For example, the table extension can look like:
We have one table from the Base App, one Table from our Extension and another 3 Tables from other extensions.
Once we rerun the example, the generated SQL Command is much complicated, contains INNER JOIN for all tables and also in addition to standard table fields, also all fields from all tables extensions (and we have absolutely no control about their design and performance-requirements).
SELECT
"17"."timestamp",
"17"."Entry No_",
"17"."G_L Account No_",
... All Table Fields (56 fields hidden) ...
"17"."Last Modified DateTime",
"17_e1"."TKA Veeery Long Description",
"17_e1"."TKA Veeery Long Description 2",
"17_e3"."TKA Veeery Long Description 5",
"17_e3"."TKA Veeery Long Description 6",
"17_e4"."TKA Veeery Long Description 7",
"17_e4"."TKA Veeery Long Description 8",
"17_e5"."TKA Veeery Long Description 3",
"17_e5"."TKA Veeery Long Description 4",
"17"."$systemId",
"17"."$systemCreatedAt",
"17"."$systemCreatedBy",
"17"."$systemModifiedAt",
"17"."$systemModifiedBy"
FROM "CRONUS".dbo."CRONUS International Ltd_$G_L Entry$437dbf0e-84ff-417a-965d-ed2bb9650972" "17" WITH(READUNCOMMITTED)
JOIN "CRONUS".dbo."CRONUS International Ltd_$G_L Entry$425df973-0cda-4173-9e7c-e91ae677bae1" "17_e1" WITH(READUNCOMMITTED) ON ("17"."Entry No_" = "17_e1"."Entry No_")
JOIN "CRONUS".dbo."CRONUS International Ltd_$G_L Entry$fa14d844-0ad9-4210-9171-9a86c033de53" "17_e3" WITH(READUNCOMMITTED) ON ("17"."Entry No_" = "17_e3"."Entry No_")
JOIN "CRONUS".dbo."CRONUS International Ltd_$G_L Entry$fa14d844-0ad9-4210-9271-9a86c033de53" "17_e4" WITH(READUNCOMMITTED) ON ("17"."Entry No_" = "17_e4"."Entry No_")
JOIN "CRONUS".dbo."CRONUS International Ltd_$G_L Entry$fa14d944-0ad9-4210-9171-9a86c033de53" "17_e5" WITH(READUNCOMMITTED) ON ("17"."Entry No_" = "17_e5"."Entry No_")
ORDER BY "Entry No_" ASC OPTION(OPTIMIZE FOR UNKNOWN, FAST 50)
Nevertheless, the second command generated absolutely same SQL Command as earlier when we had no additional tables installed. The performance difference is obvious now.
SELECT
"17"."timestamp",
"17"."Entry No_",
"17"."G_L Account No_",
"17"."Amount",
"17"."Bal_ Account Type",
"17"."Source Type",
"17"."FA Entry Type",
"17"."$systemId",
"17"."$systemCreatedAt",
"17"."$systemCreatedBy",
"17"."$systemModifiedAt",
"17"."$systemModifiedBy"
FROM "CRONUS".dbo."CRONUS International Ltd_$G_L Entry$437dbf0e-84ff-417a-965d-ed2bb9650972" "17" WITH(READUNCOMMITTED)
ORDER BY "Entry No_" ASC OPTION(OPTIMIZE FOR UNKNOWN, FAST 50)
SQL stats
What is behind these SQL Commands? In the previous article, we saw that the environment’s execution plan without any extension is not really interesting… It is just one Index Scan from the table itself.
However, the same Plan is used for environments with table extensions too, if we use fields specified using SetLoadFields! Furthermore, if we use the “old” approach (= without using SetLoadFields), SQL Server must join all table extension using INNER JOIN command that results in much complicated (and much performance-consuming)! It can be easily seen in the picture below; for example, the relative query cost of the command generated using SetLoadField is 4% compared to 93%! That is a huge difference and could result in many troubles and performance issues.
Conclusion
To summarize, it is essential to start using Partial Records for any logic with bigger-than-small complexity. Furthermore, it is crucial to minimize the number of extensions on tables, which can have many rows and/or are used within standard processes (which unfortunately does not use Partial Records yet).