Presto supports full scan right now, but pushes down key value equality pairs to the connectors for additional connector side pruning (e.g. partitions in Hive). The engine itself will also do some additional partition pruning at planning time with the partitions returned by the connector.
We currently have a pull request open right now that will allow range predicates to be be pushed into the connectors. This will allow connectors the ability to implement range/skip scans.
The core presto engine does not take advantage of indexes right now.
Presto is designed for OLAP type workloads and is thus primarily used heavily internally at Facebook for data analysis and reporting. It won't be replacing traditional databases (like mysql) that are typically used for user-facing features, but may be suitable for serving user-facing traffic doing OLAP type queries such as generating dashboards. We are actually looking into doing something like that right now for Facebook.
We currently have a pull request open right now that will allow range predicates to be be pushed into the connectors. This will allow connectors the ability to implement range/skip scans.
The core presto engine does not take advantage of indexes right now.