Files
spark/src/DataStreamSuite/DataStream/Primitives/Distinct/DistinctRowEvaluator.cs
T
John Colvin 9311c27f47 feat: Resurrect DataStream primitives
They were originally commented out because the expression parsing utils
weren't ready (now they are).
2023-09-29 13:10:50 -04:00

37 lines
1.3 KiB
C#

using System;
using System.Collections;
using BCG.Utilities;
namespace BCG
{
public class DistinctRowEvaluator : GroupByEvaluator
{
private int myIdGenerator = 0;
public DistinctRowEvaluator (string[] columnNames) : base (null, columnNames) {}
public DistinctRowEvaluator(string colMapExpression, string[] columnNames)
: base (colMapExpression, null, columnNames) {}
public override IBlueRow evaluate (DataPacket packet)
{
myOutputTable = myOutputTable == null ? new GroupByOutputTable (packet.schema) : myOutputTable;
string hashString = GroupByStream.getRepString (packet, myGroupByCols);
if (packet.newRows.Count == 0)
//if there are no rows for this group we remove the group from the output table
myOutputTable.removeRow (hashString);
else if (myOutputTable.getRow (hashString) == null)
//a row is added for the first time (or first time after the removal of the group)
myOutputTable.addRow (
BlueDataUtility.getNewReplacedRow (
BlueDataUtility.extractRow (packet.newRows, 0),
myIdGenerator++, "id"),
hashString);
return myProject (myOutputTable.getRow (hashString));
}
}
}