Things to avoid in APEX (Part-2)

Incase you haven't seen the Part-1 of Things to Avoid in Apex, below is the link

https://adityanthiruchuri.blogspot.com/2021/08/things-to-avoid-in-apex.html

Let us understand few more tips / tricks in this post.

1. Use '==' over 'equals' whenever required

We all know that APEX is not case-sensitive.  But we need to understand that there are few methods in apex, which takes the case-type (lower/upper) into consideration.   When you use the method 'equals' it will return TRUE only if the LHS and RHS are of same value and same case-type. Look at the below example

If you still want to make use of "equals" without taking case-types into consideration you have to either use method "equalsIgnoreCase()" (or) convert both the LHS and RHS into either lower / upper case and then do the comparison, but remember that these methods wont work for "null" values.   Whereas, operator '==' does the job without any such case conversions.  Use method 'equals' to do a comparison only when you want the case-match to be taken into consideration.

2. Concatenation of Strings

If given a requirement to concatenate two strings, we all tend to use the operator '+', but what if I tell you that the method .join() consumes way lesser time than the operator '+'.. Let me elaborate by taking an example.  I am trying to prepare a String by adding all the names of Opportunities separated by a comma.  There are 8500 records present while executing the below logic.

Above code consumed CPU Time of 1049ms.  Now let me execute the same logic by getting rid off the '+' operator by using "join" instead.

Above code consumed CPU Time of just 586ms which is almost 50% lesser than the time taken by the "+" operator.  Also as you are creating a new collection variable, ensure to take the heap size to take into consideration.  You can do the above implementation in your own orgs to get the results.

3. Avoid un-necessary creation of collection variables while preparing Map

Let us take a scenario of preparing Map<Id, list<Opportunity>> where key is AccountId and value will be its associated Opportunities.  Most of us will be writing code like below

In the above logic we have created an extra list at line#6, populated it with the current iterating record and then added it to the map, so that override will happen.  We need to understand that maps are mutable and that gives us the flexibility to modify an element of a map directly without having to add it again.  Consider the logic below

By using the above approach, we completely eliminated the need of creating an extra list and using the same to override the existing value.  

4. Do not use String datatype for holding Id's

I have seen many instances where developers use data-type String to hold value of type "Id". There are many drawbacks of using that approach.  We all know that Id's can be of length 15 or 18 based on the case-sensitivity.  Let us take a simple example to understand the consequence.  I am just hardcoding for our scenario, but keep in thoughts that this is never encouraged as a practice.

Reference : https://salesforce.stackexchange.com/questions/348609/why-this-map-return-null-value/348616

We need to have our code always dynamic in nature.  We can understand from the above example that if we store 15 digit Id in a string, it will fail when used against 18 digit Id and vice-versa.  But if we use the use "Id" data-type rather than string for the above, logic works just smooth.

Hope you find this interesting.  There are still many points to add and few more will be posted soon.   Let me know in the comment section if you have any such tips.  Feedback is much appreciated and thanks for being an awesome reader.  

..Bazingaa..

Enhanced Loop(For Each) Vs Traditional Loop(For)

If you have written a trigger, you must be familiar with the enhanced for-loop.  We all must have used the enhanced for-loops in our apex code at-least once.  But are they really efficient against the traditional For-loops ? In this blog, let us dig deep into the efficiency measures of the for-loops.

Let us consider an example of iterating through list of Opportunity records and manipulating a field-value for our experiment.  Look at the below snippet where we are trying to iterate a collection in traditional and enhanced way.

From the above snippet it is so clear that enhanced for-loop is so efficient over traditional for-loop because of the reasons :

1. Un-necessary variables are not required to be created in enhanced for-loop
2. We don't need to calculate the size of the collection for iteration
3. Index is not required to access an element inside collection

These reasons are enough to conclude that enhanced for-loop has won the war against traditional for-loop.   But let me tell you that traditional for-loop is efficient than enhanced for-loop in few scenarios and one such is in CPU Time Consumption.  Yes, traditional for-loop consumes lesser time than enhanced for-loop and CPU Time out exception is a nightmare for most of the developers.    Can't believe me, let me use the above example to prove my point.  Before I proceed, let me make you aware of the below system method : 

Limits.getCPUTime()  - Returns the CPU time (in milliseconds) that has been used in the current transaction

Source:  Apex Developer Guide

I'll be using the above system method to calculate the time that a block of code consumes. 

I will be running both the loops running against a collection size of 5000.  Let me first start with enhanced for-loop to see the time consumption.  

List<Opportunity> lstOpp = [SELECT Id, Name, StageName FROM Opportunity LIMIT 5000];

Integer strtTime = Limits.getCPUTime(); //Fetch the time consumed so far
//Iterate through all the opportunities
//In an enhanced for-loop by iterating through collection oppList
for(Opportunity opp : oppList){
   opp.stageName = strStgName;
}
Integer endTime = Limits.getCPUTime(); //Get the time consumed so far
//Print the time consumed by the for-loop
system.debug('Time consumed by for-loop is '+(endTime-strtTime) +'ms');

Above code resulted with the below output 

16:49:58:065 USER_DEBUG [8]|DEBUG|Time consumed by for-loop is 178 ms

Now, let us run the same code by using traditional for-loop :

List<Opportunity> lstOpp = [SELECT Id, Name, StageName FROM Opportunity LIMIT 5000];

Integer strtTime = Limits.getCPUTime(); //Fetch the time consumed so far
//Iterate through all the opportunities
//In an enhanced for-loop by iterating through collection oppList
for(Integer i=0; i<lstOpp.size(); i++){
   lstOpp[i].stageName = strStgName;
}
Integer endTime = Limits.getCPUTime(); //Get the time consumed so far
//Print the time consumed by the for-loop
system.debug('Time consumed by for-loop is '+(endTime-strtTime) +'ms');

Above code resulted with the below output 

16:52:30:323 USER_DEBUG [11]|DEBUG|Time consumed by for-loop is 107ms

It is now clear that traditional for-loop took less time than the enhanced for-loop.  Now let us make few tweaks to the traditional for-loop.  Rather than having lstOpp.size(); at the condition level, let's create a new variable for holding the same, so that calculation of collection size can be avoided at each iteration.  As .size() is present at the condition place and condition check happens for each iteration .size() calculation happens for each iteration.

List<Opportunity> lstOpp = [SELECT Id, Name, StageName FROM Opportunity LIMIT 5000];

Integer strtTime = Limits.getCPUTime(); //Fetch the time consumed so far
//Iterate through all the opportunities
//In an enhanced for-loop by iterating through collection oppList
//Created a new variable "j" to hold the size(size will be calculated only once)
for(Integer i=0, j=lstOpp.size(); i<j; i++){
   lstOpp[i].stageName = strStgName;
}
Integer endTime = Limits.getCPUTime(); //Get the time consumed so far
//Print the time consumed by the for-loop
system.debug('Time consumed by for-loop is '+(endTime-strtTime) +'ms');

Change made : Created a new variable "j" to hold the size and the size will be calculated only once unlike the prior implementation, where the size calculation happens for each iteration.

16:52:30:323 USER_DEBUG [11]|DEBUG|Time consumed by for-loop is 97ms

PS : You may not get the same exact output when you execute the above snippet in your ORG, values will always be approximately derived.

From the above execution, we understood that traditional for-loop took just half the amount of time Enhanced for-loop consumes.  This reason is enough to conclude that traditional for-loop is efficient over Enhanced For-Loop.  If you have to iterate child records which comes as part of inner SOQL Query, traditional for-loop will take just 1/3rd time of enhanced for-loop.

Then why do we have Enhanced For-Loop ?

Everyone would have started their coding exercise only through traditional for-loop. If traditional for-loop is that great, then why do we have enhanced for-loop.  ? Do we have to replace all our enhanced for-loops with traditional for.  ? 

Answer is BIG "NO".  There are instances where enhanced for-loop is much efficient than the traditional one and one good example is HEAP size consumption.  Enhanced for-loop consumes very less heap size when compared with traditional for-loop.  I recommend you to use methods of "Limits" class to check the heap size, time consumption on your loops and then take a call on the kind of for-loop that better suits your purpose.  Hope you find this interesting. 

Thanks for being an awesome reader.   Feedback is much appreciated.

..Bazingaa..

Query for all the fields of an object in SOQL

We all might have come across scenarios where we want to query for all the fields of an object.  In SQL we use the symbol * to fetch all the fields, but unfortunately we do not have any such in SOQL.  SF recently came up with amazing new feature in Spring'21 where all the fields belonging to a single sObject can be queried.

We can now use the below keywords in our SOQL.

• FIELDS(ALL) - Fetches all(including standard and custom) the fields of object.
• FIELDS(STANDARD) - This fetches only the standard fields of the object.
• FIELDS(Custom) - This fetches only the custom fields of the object.

For using the above Keywords, the pre-requisite is "LIMIT" keyword should be used with utmost value of "200".  This means that not more than 200 records can be fetched if any of the above mentioned keywords are present in the SOQL.  And also is not supported with an unbounded set of fields in this API

//Below query will fetch all the fields of the Account Object

SELECT FIELDS(ALL) FROM Account LIMIT 200

//Below query will fetch only the Standard fields of the Account Object
SELECT FIELDS(STANDARD) FROM Account LIMIT 200

//Below query will fetch only the custom fields of the Account Object
SELECT FIELDS(CUSTOM) FROM Account LIMIT 200

 PS: Above can be executed in "Query Editor" present inside Developer Console to check the results.  Only FIELDS(STANDARD) is made available to be used inside APEX.

Source : https://developer.salesforce.com/blogs/2021/01/new-soql-fields-function-is-ga

It is undoubtedly a wonderful feature but with the limitation of not fetching more than 200 records and not making it completely available in apex, it could not solve our purpose in all the scenarios.

To overcome this, let us make use of Dynamic SOQL to get all the fields of the record.  To make this generic enough, let's create a new static method which accepts object Name as String.   With the received objectName it will return a SOQL Query(String) that includes all the fields

public class Utils {
    public static String prepareSOQL(String sobjectName){
	String strSOQL = 'SELECT ';
	Map<String, Schema.SObjectField> fMap = Schema.getGlobalDescribe().get(sobjectName.toLowerCase()).getDescribe().Fields.getMap();
    
	for (Schema.SObjectField ft : fMap.values()){ 
		Schema.DescribeFieldResult fd = ft.getDescribe();
		if (fd.isUpdateable()){ // add field only if it is accessable
			strSOQL += fd.getName()+',';
		}
	}
	strSOQL = strSOQL.removeEnd(',');
	strSOQL += ' FROM '+sobjectName;
	return strSOQL;
    }
}

The above method accepts a string (objectName) and returns a SOQL Query of type string which includes all the fields and this can be invoked from any place when needed

//lstAcc will hold all the records with all the fields
String strSOQL = Utils.prepareSOQL('Account');

List<Account> lstAcc = Database.query(strSOQL);

If we use the approach of dynamic SOQL, we don't have to worry about the LIMIT clause, but remember to use this approach of querying for all the fields only when needed and also keep in thoughts that the SOQL Query length should not exceed more than 10K.

Tip : Usage of .keyset() is not supported in Dynamic SOQL.

Yes, you read it right.  We all tend to use Maps and have them part of our SOQL query.  But map.keyset() in a dynamic SOQL Query will not fetch the desired result.  You need to assign the map.keyset() to another set variable and use set in your SOQL.  It is because variables / collections can be used in dynamic SOQL, but not the methods.  

Hope you find it interesting.  Thank you for visiting and feedback is much appreciated!!! 

"....Bazingaa..."