There are  1000 of resources online   to get overwhelmed by them to find a particular solution is expected 

To avoid this what we can do is follow these steps

1)Understand the problem statement and try to find out a solution for that specific problem

let's say you have to parse a unstructured xml and store it in hive table.

So rather than randomly looking out you can come with a easiest way out to solve this by asking intelligent questions to google and using your brain

It is never a bad idea to use a existing implementing functionality or api if it covers almost evything of a use case

2)Come out for a discussion

Now since you have finally made it its time to discuss this out with mentors and team mates to look at complexity of the problem and how can we improve the solution 

This is the best way to improve specially when you are using batch data source for populating streaming data and your mentor can come to save you in this scenario

3)Final solution 

Now its time to let the solution be on to the quality amalysis team for regression and other testing and its time to solve the bugs again

Happy learning and building