Top rows

Supported in: Batch

Picks the top rows in each sorted partition.

Transform categories: Aggregate

Declared arguments

• Dataset - Input dataset.
  Table
• Partition by columns - Set of columns to determine each partition from.
  Set<Column<AnyType>>
• Sort specification - Specification for how to sort each partition. At least one specification is required.
  List<Tuple<Column<AnyType>, Enum<Ascending, Descending>>>
• optional Number of rows - Number of rows to select, defaults to 1.
  Literal<Integer>

Examples

Example 1: Base case

Argument values:

• Dataset: ri.foundry.main.dataset.a
• Partition by columns: {airline}
• Sort specification: [(airport, DESCENDING), (miles, ASCENDING)]
• Number of rows: null

Input:

Output:

---

Example 2: Base case

Argument values:

• Dataset: ri.foundry.main.dataset.a
• Partition by columns: {}
• Sort specification: [(airline, DESCENDING), (airport, DESCENDING), (miles, ASCENDING)]
• Number of rows: 2

Input:

Output:

---

Example 3: Base case

Argument values:

• Dataset: ri.foundry.main.dataset.a
• Partition by columns: {}
• Sort specification: []
• Number of rows: 1

Input:

Output:

---