Trigger creating duplicate records

Question

I have written some code which create tasks whenever I create a new staff record based on some conditions.

Trigger:

trigger triggerTaskImplemetationonJob on Staff__c (before insert, before update) {

    List<Task> taskList = new List<Task>();


    List<Staff__c> jobList = trigger.New;
    String[] groupNames = jobList[0].Group__c != null
      ? jobList[0].Group__c.split(';')
      : new List<String>();
    system.debug('groupNames :::' + groupNames);

    list<GroupMember> gm = [select group.name,group.DeveloperName,UserOrGroupId from GroupMember where group.DeveloperName =: groupNames]; 
    List<Task> tasks = new List<Task>();

    system.debug('developerName ::' + group.DeveloperName); 
    system.debug('*****group****'+gm);

    for(Staff__c jobObj : jobList){
        if(gm != null && gm.size() > 0){
            for(GroupMember g : gm){
                taskList.add(new Task(
                    ActivityDate = Date.today(),
                    WhatID = jobObj.Id,
                    ownerID =  g.UserOrGroupId,
                    Status = 'Open',
                    type = 'Other',
                    Priority = 'Normal'
            }
        }
    }

    if(taskList.size() > 0)
        insert taskList;


}

The group__c is a multi picklist field and which is allows me to select multiple public groups and create tasks to the users in those public groups

Ex: I have three group . A, B, C group A holds X,Y as users group B holds Z as users group C holds P,Q as users

Whenver I create a new staff recording by selecting the group names from mutliple picklist field ( ex: I seleted A, B for the first time) and it created 3 tasks for me . Now I added one more picklist value ( I added group C also) and saved the record. In this time, it is suppose to create only 2 more tasks, but its created 5 new tasks in those 3 are already exists.

I don't want to create duplicate tasks in this case. I know that I am not checking some condition and which is causing the duplicate record creation. Please help me how to fix it

Answer 1

You're creating new records each time around, on every insert and update. Just clicking Edit and Save will cause more records to appear. Instead, you can either query for existing records, or to save a query, you can just check for differences.

Map<Id, Set<String>> jobs = new Map<Id, Set<String>>();
Set<String> allJobNames = new Set<String>();

for(Integer index = 0, size = Trigger.new.size(); index < size; index++) {
    Staff__c
        oldRec = Trigger.isInsert? new Staff__c(): Trigger.old[index],
        newRec = Trigger.new[index];
    Set<String> 
        oldJobs = new Set<String>((oldRec.Groups__c==null? '': oldRec.Groups__c).split(';')),
        newJobs = new Set<String>((newRec.Groups__c==null? '': newRec.Groups__c).split(';'));
    newJobs.removeAll(oldJobs);
    allJobNames.addAll(newJobs);
    jobs.put(Trigger.new[index].Id, newJobs);
}

At this point, you now have a list of jobs for each staff entry that are new since the last time your code ran. Note that this code is bulkified, so it can support any number of records per transaction (currently limited to 200).

Next, you need to get all of the groups you're interested in:

Map<String, List<GroupMember>> membersByGroupName = new Map<String, List<GroupMember>>();
for(String jobName: allJobNames) {
    membersByGroupName.put(jobName, new GroupMember[0]);
}
for(GroupMember record: [SELECT Group.DeveloperName, UserOrGroupId FROM GroupMember WHERE Group.DeveloperName = :allJobNames]) {
    membersByGroupName.get(record.Group.DeveloperName).add(record);
}

Finally, you just need to build your tasks:

Task[] newTasks = new Task[0];
for(Id staffRecordId: jobs.keySet()) {
    Set<Id> userIds = new Set<Id>();
    for(String jobName: jobs.get(staffRecordId)) {
        for(GroupMember member: membersByGroupName.get(jobName)) {
            if(!userIds.contains(member.UserOrGroupId)) {
                // Create tasks here
            }
        }
    }
}
insert newTasks;

Other notes:

You never need to check if the results of a query are null.

Account[] a = [SELECT Id FROM Account LIMIT 5];

a is guaranteed to be a list, not a null value.

Iterating over an empty loop is harmless.

String[] strings = new String[0];
for(String value: strings) {
    System.assert(false);
}

This code will run without error, because a list with zero elements results in a for-loop that runs zero times.

Empty DML operations don't count against you.

Account[] a = new Account[0];
for(Integer i = 0; i < 200; i++) {
    insert a;
}

This code runs without error in a fairly trivial amount of time, because no records are being inserted, so no governor limits are broken. More to the point, it's okay to insert a list without checking if it's empty.