Questions about Aggregation Provider and Aggregation Function. Topic, #1000

RogeGS · 2025-06-05T16:15:53Z

RogeGS
Jun 5, 2025

Good afternoon. I've been reviewing the official Aggregate Provider documentation, which explains how to customize it to improve queries in the cube, and I have some questions about this topic to understand some concepts and how the application works

1. I understand that the application creates some Aggregate Providers by default. 
What criteria does the application use to create them by default?. That is, is it based on the hierarchies defined in the cube,
     on the primary key fields defined in each store in the cube, or on a combination of both?
   
    2. What is the difference between the Aggregate Provider and Partial Aggregation? 
           Could you explain this with an example, if possible?

    3. I've read the recommended method for Atoti-Server to obtain aggregate provider recommendations based on query history 
        and using JConsole.
          Could you tell me if there is a similar method for recommendations using the Python Atoti API?
           If not, what method do you recommend for studying Atoti-API queries with Python and analyzing which aggregate 
                providers need to be customized to improve the application?

 4. Using the cache for Provided aggregates
	When do you recommend using this configuration option?
	What values are recommended to set upfront?

On the other hand, while reading the aggregation function customizations, I saw the example of aggregating two vectors
for their first two components.
https://docs.activeviam.com/products/atoti/server/latest/docs/cube/aggregation-functions/#aggregating-vectors
When it is customized, SubVectorSumBinding and SumVectorAggreg 10000 ationBinding are explained.
But the code capture for the SumVectorAggregationBinding implementation is not included. However, there is a
sentence that states the following: "The bindings of the SumFunction can be reused to sum the vectors of the two columns."
I'm not entirely clear on the meaning of this sentence. That is, does it refer to the fact that Sum's native aggregation function for
vectors can be used to implement SubVectorSumBinding? Can you clarify this for me?

Answered by tibdex

Jun 24, 2025

What criteria does the application use to create them by default

There are no default aggregate providers. When Cube.aggregate_providers is empty, all the aggregates are computed "just in time" by issuing queries to the database.

Aggregate Provider and Partial Aggregation

In Atoti Python SDK there is always a JIT aggregate provider fallback, all the items in Cube.aggregate_providers are partial providers.

Could you tell me if there is a similar method for recommendations using the Python Atoti API

Sadly no.

what method do you recommend for studying Atoti-API queries with Python and analyzing which aggregate providers need to be customized to improve the application

We don't have d…

View full answer

tibdex · 2025-06-24T20:53:52Z

tibdex
Jun 24, 2025
Maintainer

What criteria does the application use to create them by default

There are no default aggregate providers. When Cube.aggregate_providers is empty, all the aggregates are computed "just in time" by issuing queries to the database.

Aggregate Provider and Partial Aggregation

In Atoti Python SDK there is always a JIT aggregate provider fallback, all the items in Cube.aggregate_providers are partial providers.

Could you tell me if there is a similar method for recommendations using the Python Atoti API

Sadly no.

what method do you recommend for studying Atoti-API queries with Python and analyzing which aggregate providers need to be customized to improve the application

We don't have docs for this yet. You can measure the impact of an aggregate provider configuration change through query plans returned by passing explain=True to Session.query_mdx() or Cube.query(). If you need more help, please use the chat button on the lower right corner of https://www.activeviam.com.

Using the cache for Provided aggregates

What do you mean?

5 replies

RogeGS Jul 1, 2025
Author

first at all, thanks for your answer. they have hepled me to clarify the concepts.

in the question 4 im referending on one of the option when you configurate the aggregator providers.
in this case the cache memory option.
about this topic i asked when is it recomende using cache ? and What amount of memory or percentage of cache is recommended to start with?.

thanks so much.

tibdex Jul 1, 2025
Maintainer

Nothing is related to cache here so I don't understand what you're referring to.

RogeGS Jul 1, 2025
Author

in this page https://docs.activeviam.com/products/atoti/server/latest/docs/cube/configuration/

we have this example of configuration of the cube

StartBuilding.cube()
    .withName("MyCube")
    .withMeasures(measures)
    .withDimensions(dimensions)
    .withAggregateProvider()
    .jit()
    .withProperty("rangeSharing", "true")
    .withPartialProvider()
    .bitmap()
    .excludingHierarchy("Trades", "Trades")
    .end()
    .includingOnlyMeasures("pnlDelta.SUM", "pnl.SUM")
    .withPartialProvider()
    .bitmap()
    .includingOnlyHierarchy("Underlyings", "Underlyings")
    .and("HistoricalDates", "HistoricalDates")
    .end()
    .withPartialProvider()
    .bitmap()
    .includingOnlyMeasures("pnlDelta.SUM", "pnl.SUM")
    .filteredOn(Map.of(LevelIdentifier.simple("Date"), new Date(2022, Calendar.JUNE, 14)))
    .withPartialProvider()
    .excludingMeasures("pnlDelta.SUM", "pnl.SUM")
    .withDrillthroughExecutor()
    .withKey("TIME_BUCKET_DT_EXECUTOR")
    .withProperty("bucketHierarchy", "TimeBucketDynamic")
    .withProperty("bucketedLevel", "Value Date")
    .end()
    ****.withAggregatesCache()
    .withSize(10_000)****
    .cachingOnlyMeasures("contributors.COUNT", "pnl.SUM")
    .withSharedContextValue(
        new DrillthroughProperties(
            new CustomComparator<>(List.of("Desk", "Currency"), null),
            List.of("BookId", "City"),
            List.of(
                new CalculatedDrillthroughColumnDescription(
                    "DoubleAdder", "delta + gamma", "delta,gamma")),
            List.of(
                new CalculatedDrillthroughColumnSetDescription(
                    "PnlCurrencyColumnSet",
                    Immutable.map("prefix", "pnl in", "currencies", "EUR,USD,GBP,JPY,CHF,ZAR")
                        .toProperties()))))
    .withSharedContextValue(new MdxContext())
    .build();

in this code we have this

****.withAggregatesCache()
    .withSize(10_000)****

that indiques to me we can config cache for this aggregate provider.

tibdex Jul 1, 2025
Maintainer

This GitHub repository is for Atoti Python SDK Community Edition, if you need support for Atoti Server, please use the chat button on the lower right corner of https://www.activeviam.com/.

RogeGS Jul 2, 2025
Author

ah ok!!, sorry.
Thanks a lot for your answers

Questions about Aggregation Provider and Aggregation Function. Topic, #1000

Uh oh!

Uh oh!

RogeGS Jun 5, 2025

Replies: 1 comment · 5 replies

Uh oh!

tibdex Jun 24, 2025 Maintainer

Uh oh!

Uh oh!

RogeGS Jul 1, 2025 Author

Uh oh!

tibdex Jul 1, 2025 Maintainer

Uh oh!

Uh oh!

RogeGS Jul 1, 2025 Author

Uh oh!

tibdex Jul 1, 2025 Maintainer

Uh oh!

RogeGS Jul 2, 2025 Author

RogeGS
Jun 5, 2025

Replies: 1 comment 5 replies

tibdex
Jun 24, 2025
Maintainer

RogeGS Jul 1, 2025
Author

tibdex Jul 1, 2025
Maintainer

RogeGS Jul 1, 2025
Author

tibdex Jul 1, 2025
Maintainer

RogeGS Jul 2, 2025
Author