CREATE AGGREGATING INDEX

Different syntax is used when creating an aggregating index on an empty table or a table populated with data. After an aggregating index is created, Firebolt automatically updates the index as new data is ingested. For more information, see Using aggregating indexes.

Syntax–aggregating index on an empty table

CREATE AGGREGATING INDEX <agg_index_name> ON <fact_table_name> (
  <key_column>[,...<key_columnN>],
  <aggregation>[,...<key_columnN>]
);

The index is populated automatically as data is loaded into the table.

Syntax–aggregating index on a populated table

CREATE AGGREGATING INDEX [IF NOT EXISTS] <agg_index_name> ON <fact_table_name> (
  <key_column>[,...<key_columnN>],
  <aggregation>[,...<aggregationN>]
);

Parameters

Parameter	Description
<index>	Specifies a unique name for the index
<table>	Specifies the name of the fact table referenced by this index
<column>	Specifies column name from the <table> used for the index
<aggregation>	Specifies one or more aggregation functions to be applied on a <key_column>, such as SUM, COUNT, AVG, and more.

Example–create an aggregating index

In the following example, we create an aggregating index on the fact table my_fact, to be used in the following query:

SELECT
  product_name,
  count(DISTINCT source),
  sum(amount)
FROM
  my_fact
GROUP BY
  product_name;

The aggregating index is created with the statement below.

CREATE AGGREGATING INDEX my_fact_agg_idx ON my_fact (
  product_name,
  count(distinct source),
  sum(amount)
);

To benefit from the performance boost provided by the index, include in the index definition all columns and measurements that the query uses.