AggregationDataStore: Static Attribute Aggregation

[aklish]

The AggregationDataStore needs a schema for each entity that:

• Specifies that an attribute is a metric
• Specifies that an attribute or relationship is a dimension
• Specifies that an attribute is a Data/Time column and supports grain.
• Specifies that an Attribute is a friendlyName or human displayable column for that dimension.
• Specifies whether a dimension has small, medium, or large cardinality.

Expected Behavior

New AggregationDataStore annotations should be created:

1. MetricAggregation
2. Cardinality
3. ComputedMetric
4. FriendlyName

Examples

@MetricAggregation(default = Sum.class, aggregations = {Max.class, Median.class})

@ComputedMetric(expression = '(fieldA * fieldB) / 100')

@Cardinality(size = MEDIUM)

@FriendlyName

Current Behavior

N/A

Possible Solution

Elide calls populateEntityDictionary on each data store.

The AggregationDataStore should only bind entities that have at least one MetricAggregation annotation. For each entity bound, it should use the EntityDictionary to lookup other data store specific annotations and construct a schema object for that entity. The schema will be used as foundational metadata for the store to perform its functions.

[QubitPi]

@aklish Is @Cardinality is same thing as the cardinality in JPQL database? Do we simply trust user who will set whatever value it will be for size?

[aklish]

@Cardinality is a new annotation that will have a value of SMALL, MEDIUM, or LARGE. We trust the developer to set it. The values should be an enum. Also, let's rename @ComputedMetric to @MetricComputation.

[QubitPi]

Ok. Yeah enum is my choice, too