Aggregation analysis rule

In AWS Clean Rooms, an aggregation analysis rule generates aggregate statistics using COUNT, SUM, and/or AVG functions along optional dimensions. When the aggregation analysis rule is added to a configured table, it enables the member who can query to run queries on the configured table.

The aggregation analysis rule supports uses cases such as campaign planning, media reach, frequency measurement, and attribution.

The supported query structure and syntax are defined in Aggregation query structure and syntax.

The parameters of the analysis rule, defined in Aggregation analysis rule - query controls, include query controls and query results controls. Its query controls include the ability to require that a configured table is joined to at least one configured table owned by the member who can query, either directly or transitively. This requirement allows you to ensure that the query is run on the intersection (INNER JOIN) of your table and theirs.

Aggregation query structure and syntax

Queries on tables that have an aggregation analysis rule must adhere to the following syntax.

    --select_aggregate_function_expression
SELECT 
aggregation_function(column_name) [[AS] column_alias ] [, ...]

    --select_grouping_column_expression                        
  [, {column_name|scalar_function(arguments)} [[AS] column_alias ]][, ...]   

--table_expression
FROM table_name [[AS] table_alias ]
  [[INNER] JOIN table_name [[AS] table_alias] ON join_condition] [...]

--where_expression
[WHERE where_condition]          

--group_by_expression                          
[GROUP BY {column_name|scalar_function(arguments)}, ...]]                  

--having_expression
[HAVING having_condition]                               

--order_by_expression    
[ORDER BY {column_name|scalar_function(arguments)} [{ASC|DESC}]] [,...]]

The following table explains each expression listed in the preceding syntax.

FROM consumer_table 
INNER JOIN provider_table
ON
consumer_table.identifier1 = provider_table.identifier1
AND
consumer_table.identifier2 = provider_table.identifier2

For aggregation query structure and syntax, be aware of the following:

Aggregation analysis rule - query controls

With aggregation query controls, you can control how the columns in your table are used to query the table. For example, you can control which column is used for joining, which column can be counted, or which column can be used in WHERE statements.

The following sections explain each control.

Aggregation controls

By using aggregation controls, you can define which aggregation functions to allow, and what columns they must to be applied to. Aggregation functions can be used in the SELECT, HAVING, and ORDER BY expressions.

Join controls

A JOIN clause is used to combine rows from two or more tables, based on a related column between them.

You can use Join controls to control how your table can be joined to other tables in the table_expression. AWS Clean Rooms only supports INNER JOIN. INNER JOIN statements can only use columns that have been explicitly categorized as a joinColumn in your analysis rule, subject to the controls that you define.

The INNER JOIN must operate on a joinColumn from your configured table and a joinColumn from another configured table in the collaboration. You decide which columns from your table can be used as joinColumn.

Each match condition within the ON clause is required to use the equality comparison condition (=) between two columns.

Multiple match conditions within an ON clauses can be:

The following is an example of a query with an AND logical operator.

SELECT some_col, other_col 
FROM table1 
    JOIN table2 
    ON table1.id = table2.id AND table1.name = table2.name

The following is an example of a query with an OR logical operator.

SELECT some_col, other_col 
FROM table1 
    JOIN table2 
    ON table1.id = table2.id OR table1.name = table2.name

Following is an example of transitivity.

ON 
my_table.identifer = third_party_table.identifier
....
ON
third_party_table.identifier = member_who_can_query_table.id

Dimension controls

Dimension controls control the column along which the aggregation columns can be filtered, grouped, or aggregated.

Scalar functions

Scalar functions control which scalar functions can be used on dimension columns.

The following scalar functions are supported:

For more details, see the AWS Clean Rooms SQL Reference.

Aggregation analysis rule - query results controls

With the aggregation query results controls, you can control which results are returned by specifying one or more conditions that each output row must meet for it to be returned. AWS Clean Rooms supports aggregation constraints in the form of COUNT (DISTINCT column) >= X. This form requires that each row aggregates at least X distinct values of a choice from your configured table (for example, a minimum number of distinct user_id values). This minimum threshold is automatically enforced, even if the submitted query itself does not use the specified column. They are enforced collectively across each configured table in the query from the configured tables from each member in the collaboration.

Each configured table must have at least one aggregation constraint in their analysis rule. Configured table owners can add multiple columnName and associated minimum and they are enforced collectively.

Aggregation constraints

Aggregation constraints control which rows in the query results are returned. To be returned, a row must meet the specified minimum number of distinct values in each column specified in the aggregation constraint. This requirement applies even if the column isn't explicitly mentioned in the query or in other parts of the analysis rule.

Aggregation analysis rule structure

The following example shows a predefined structure for an aggregation analysis rule.

In the following example, MyTable refers to your data table. You can replace each user input placeholder with your own information.

{
  "aggregateColumns": [
    {
      "columnNames": [MyTable column names], "function": [Allowed Agg Functions]
    },
  ],
  "joinRequired": ["QUERY_RUNNER"],  
  "joinColumns": [MyTable column names],
  "dimensionColumns": [MyTable column names],
  "scalarFunctions": [Allowed Scalar functions],
  "outputConstraints": [
    {
      "columnName": [MyTable column names], "minimum": [Numeric value] 
    },
  ]
}

Aggregation analysis rule - example

The following example demonstrates how two companies can collaborate in AWS Clean Rooms using aggregation analysis.

Company A has customer and sales data. Company A is interested in understanding product return activity. Company B is one of Company A's retailers and has returns data. Company B also has segment attributes on customers that are useful to Company A (for example, purchased related products, uses customer service from the retailer). Company B doesn't want to provide row-level customer return data and attribute information. Company B only wants to enable a set of queries for Company A to obtain aggregate statistics on overlapping customers at a minimum aggregation threshold.

Company A and Company B decide to collaborate so that Company A can understand product return activity and deliver better products at Company B and other channels.

To create the collaboration and run an aggregation analysis, the companies do the following:

Troubleshooting aggregation analysis rule issues

Use the information here to help you diagnose and fix common issues when you work with aggregation analysis rules.

My query didn’t return any results

This can happen when there are no matching results or when the matching results don’t meet one or more minimum aggregation thresholds.

For more information about minimum aggregation thresholds, see Aggregation analysis rule - example.