Query vs. Scan

There are two possible ways to retrieve items from a DynamoDB table: query and scan. Scan works on any table, no matter what is the structure of its keys, and goes through all items filtering out what's not relevant. Performance will seriously suffer if the table is big, but for small tables such operation is acceptable:

public class MyData {
  public Collection<Item> items() {
    return this.region.table()
      .frame()
      .where("name", Conditions.equalTo("John Doe"))
      .through(new ScanValve().withLimit(50));
  }
}

By providing an instance of ScanValve to the Frame we instruct it which strategy to use for fetching items from the DynamoDB table.

Query operation requires a table to have hash/range key schema and demands at least hash to be mentioned in the query conditions. For example:

public class MyData {
  public Collection<Item> items() {
    return this.region.table()
      .frame()
      .where("hash", Conditions.equalTo(83788322))
      .through(new QueryValve().withScanIndexForward(false));
  }
}

This example will fetch items from the table in a backward order.