How to reduce database time used to query millions of records in a batch class?

Question

I am running a batch class that queries Line_Item__c in our production org. The query is executed in start method of Batch class. Since we have a total of more than 12 million records in our database for Line_Item__c, the execution is being terminated without any error. As a hotfix, I indexed query using specific accounts, which I received by generating report on Line_Item__c, and it ran within seconds. How can I apply a permanent fix for this? (Batch size is gives as 200 in execute batch from where the class is called) Here goes my code:

global class Batchclass implements Database.Batchable<SObject>{

    public Date invDate;

    public Batchclass(Date batInvDate){
        invDate = batInvDate;
    }

    global Database.QueryLocator start(Database.BatchableContext BC){
        String query = 'SELECT {around 50 fields} FROM Line_Item__c WHERE Package_Type__c != NULL AND Post_Status__c = \'unposted\' AND Invoice_Date__c = :invDate';
        return Database.getQueryLocator(query);
    }

    global void execute(Database.BatchableContext BC, List<Line_Item__c> LIList){

        if (LIList.size() > 0) {
            Apexclassprocess.dataProcess(LIList);
            
        }
    }

    global void finish(Database.BatchableContext BC) {
        // execute any post-processing operations
  }
}

Answer 1

If you need to segment your data 10 times, you could use 10 queuable methods, each method will call a batch from a segment of data. You could segment your data using the soql LIMIT keyword for instance.