[teiid-issues] [JBoss JIRA] (TEIID-4880) Join Elimination in Star Schema

[Madou Coulibaly (JIRA)]

     [ https://issues.jboss.org/browse/TEIID-4880?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Madou Coulibaly updated TEIID-4880:
-----------------------------------
    Description: 
Suppose we have the following star schema :
* Fact_Sales (Fact table)
* Dim_Date, Dim_Store and Dim_Product (Dimension tables)

Each table has a PK (id field)
Fact_Sales has a foreign key for each dimension (date_id, store_id, product_id) in order to link them.

Then, suppose we create an aggregate view of the fact table and its dimensions:

SELECT
\*
FROM Fact_Sales F
INNER JOIN Dim_Date D    ON (F.Date_Id = D.Id)
INNER JOIN Dim_Store S   ON (F.Store_Id = S.Id)
INNER JOIN Dim_Product P ON (F.Product_Id = P.Id)

Suppose now we only project fields from the fact table (Fact_Sales) and just one of its dimension (Dim_Product) in a query:

SELECT
Fact_Sales.<fields>, Dim_Product.<fields>
FROM aggregate_view

During the Query Optimization, Teiid should be able to use the PK_FK constraints which link the fact table with its dimension in order to deduct that no fields from the other dimensions (Dim_Date, Dim_Store) are projected so the joins with these tables can be removed for the Query Plan.


  was:
Suppose we have the following star schema :
* Fact_Sales (Fact table)
* Dim_Date, Dim_Store and Dim_Product (Dimension tables)

Each table has a PK (id field)
Fact_Sales has a foreign key for each dimension (date_id, store_id, product_id) in order to link them.

Then, suppose we create an aggregate view of the fact table and its dimensions:
_SELECT
*
FROM Fact_Sales F
INNER JOIN Dim_Date D    ON (F.Date_Id = D.Id)
INNER JOIN Dim_Store S   ON (F.Store_Id = S.Id)
INNER JOIN Dim_Product P ON (F.Product_Id = P.Id)_

Suppose now we only project fields from the fact table (Fact_Sales) and just one of its dimension (Dim_Product) in a query:
_SELECT
Fact_Sales.<fields>, Dim_Product.<fields>
FROM aggregate_view_

During the Query Optimization, Teiid should be able to use the PK_FK constraints which link the fact table with its dimension in order to deduct that no fields from the other dimensions (Dim_Date, Dim_Store) are projected so the joins with these tables can be removed for the Query Plan.




> Join Elimination in Star Schema
> -------------------------------
>
>                 Key: TEIID-4880
>                 URL: https://issues.jboss.org/browse/TEIID-4880
>             Project: Teiid
>          Issue Type: Feature Request
>          Components: Query Engine
>    Affects Versions: 8.7.11.6_2
>            Reporter: Madou Coulibaly
>            Assignee: Steven Hawkins
>
> Suppose we have the following star schema :
> * Fact_Sales (Fact table)
> * Dim_Date, Dim_Store and Dim_Product (Dimension tables)
> Each table has a PK (id field)
> Fact_Sales has a foreign key for each dimension (date_id, store_id, product_id) in order to link them.
> Then, suppose we create an aggregate view of the fact table and its dimensions:
> SELECT
> \*
> FROM Fact_Sales F
> INNER JOIN Dim_Date D    ON (F.Date_Id = D.Id)
> INNER JOIN Dim_Store S   ON (F.Store_Id = S.Id)
> INNER JOIN Dim_Product P ON (F.Product_Id = P.Id)
> Suppose now we only project fields from the fact table (Fact_Sales) and just one of its dimension (Dim_Product) in a query:
> SELECT
> Fact_Sales.<fields>, Dim_Product.<fields>
> FROM aggregate_view
> During the Query Optimization, Teiid should be able to use the PK_FK constraints which link the fact table with its dimension in order to deduct that no fields from the other dimensions (Dim_Date, Dim_Store) are projected so the joins with these tables can be removed for the Query Plan.



--
This message was sent by Atlassian JIRA
(v7.2.3#72005)