Writing efficient code

Question

I would like to understand the difference between snippet 1 and snippet 2. In snippet 1 I get a DML rows limit exception whereas in snippet 2, I don't get any exception. What makes snippet 2 not throw any error over snippet 1? Snippet 1 uses two for loops (newCase and eMLlist) and both the for loops uses maps to fetch the values. Snippet 2 on the other end uses a list casetoInsert and a map closedCasesM.get(caseLoop.ParentId).Equipment_Maintenance_Items__r. I am puzzled as to what makes snippet 2 efficient. Appreciate if you could highlight the difference.

Snippet 1:

 for(Case cd : newCase.values()){
            system.debug('cd.Id'+cd.Id);
            for(Equipment_Maintenance_Item__c e : eMLlist.values()){
                Equipment_Maintenance_Item__c emi = new Equipment_Maintenance_Item__c();
                
                emi.Quantity__c = e.Quantity__c;
                emi.Equipment__c = e.Equipment__c;
                emi.Maintenance_Request__c = cd.Id;
                emiList.add(emi);
            }}
        
        insert emiList;

Snippet 2:

 Map<Id,Case> closedCasesM = new Map<Id,Case>([SELECT Id, (SELECT Id, Equipment__c, Quantity__c FROM Equipment_Maintenance_Items__r) FROM Case WHERE Id IN :caseMap]);
        
        system.debug('newcasethatisinserted'+ closedCasesM);
        for(Case caseLoop : casetoInsert){
        for(Equipment_Maintenance_Item__c e : closedCasesM.get(caseLoop.ParentId).Equipment_Maintenance_Items__r){
            Equipment_Maintenance_Item__c emi = new Equipment_Maintenance_Item__c();
            emi.Quantity__c = e.Quantity__c;
            emi.Equipment__c = e.Equipment__c;
            emi.Maintenance_Request__c = caseLoop.Id;
            
            emiList.add(emi);
        }
        }
        insert emiList;

Answer 1

While neither code snippet is a complete example, the key distinguishing factor appears to be these lines:

for(Equipment_Maintenance_Item__c e : eMLlist.values()){
        

versus

for(Equipment_Maintenance_Item__c e : closedCasesM.get(caseLoop.ParentId).Equipment_Maintenance_Items__r){

In the first version, your inner for loop iterates over all Equipment_Maintenance_Item__c records for every Case (the outer loop), whether the Equipment_Maintenance_Item__c record is related to the current Case or not. This results in multiplicative creation of irrelevant records at high volume - count(Case) * count(Equipment_Maintenance_Item__c).

In the second version, you appear to be correctly filtering to iterate over, in the inner loop, only the Equipment_Maintenance_Item__c records related to the current Case, which results, apparently, in creating only the desired records - only count(Equipment_Maintenance_Item__c).

So in fact, it's not an efficiency concern here so much as a behavior problem - the code as implemented is not behaving correctly - that ultimately manifests a limits issue.