We sample because we want to get an overview of the population. It's usually impractical to study the entire population.
For example, say you want to know what percent of people who go to a certain grocery store have a dog. It wouldn't make sense to sample every single person who goes to that grocery store; there's just too many people. But if you took a sample of say, 100 people, and asked them if they had a dog, you could get a rough estimate of all the people.