Module com.darkcollective.relix.semantic

Package com.darkcollective.relix.semantic.internal

Class DownsampleColumns

java.lang.Object

com.darkcollective.relix.semantic.internal.DownsampleColumns

public final class DownsampleColumns
extends Object

Which columns a DOWNSAMPLE consolidates, and what each one is called.

Two phases need this answer and they must give the same one: schema inference builds the operator's output heading from it, and the executor fills that heading in. When the two decided separately, a column either phase admitted and the other did not became a row of the wrong width. So the rule is stated once, here, and both read it.

The rule

• The grouping keys and the bucket timestamp come first and are not consolidations — they identify the bucket rather than summarising it.
• ConsolidationFunction.COUNT summarises the rows, so it yields a single count column reducing no input column at all.
• Every other function yields one <fn>_<column> column per eligible input column: the numeric ones for AVG/SUM, whose arithmetic is defined over nothing else, and every scalar one for MIN/MAX, which compare rather than compute and rank a date or a name as readily as a number. A nested column is never consolidated — an array has no position in the engine's value order.

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
static final record	DownsampleColumns.Consolidation	One consolidated output column, and the input column reduced into it.

Field Summary

Fields

Modifier and Type	Field	Description
static final String	BUCKET_COLUMN	The output column a bucket's timestamp lands in.
static final String	COUNT_COLUMN	The output column ConsolidationFunction.COUNT produces.

Method Summary

Modifier and Type	Method	Description
static List<DownsampleColumns.Consolidation>	of(ConsolidationFunction function, List<String> groupingKeys, String timestampColumn, Schema input)	The consolidated columns a DOWNSAMPLE emits, in output order.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Details

BUCKET_COLUMN

public static final String BUCKET_COLUMN

The output column a bucket's timestamp lands in.

See Also:

• Constant Field Values

COUNT_COLUMN

public static final String COUNT_COLUMN

The output column ConsolidationFunction.COUNT produces.

See Also:

• Constant Field Values

Method Details

of

public static List<DownsampleColumns.Consolidation> of(ConsolidationFunction function,
 List<String> groupingKeys,
 String timestampColumn,
 Schema input)

The consolidated columns a DOWNSAMPLE emits, in output order.

These follow the grouping keys and the bucket column, which together make up the whole output heading.

Taken as loose parts rather than as a DownsampleNode because the executor asks the same question of a physical plan node, which carries these fields and not the logical node they came from.

Parameters:

function - the consolidation; must not be null

groupingKeys - the bucket's grouping keys, which are not consolidated

timestampColumn - the column bucketed on, which is not consolidated either

input - the schema of the operator's input; must not be null

Returns:

the consolidations, in output order; never null, possibly empty when no input column is eligible