17.10 - Reasonable Indexed Join Plan With Star Join Optimization and a Fact Table NUSI - Advanced SQL Engine - Teradata Database

Teradata Vantageā„¢ - SQL Request and Transaction Processing

Product
Advanced SQL Engine
Teradata Database
Release Number
17.10
Release Date
July 2021
Content Type
Programming Reference
User Guide
Publication ID
B035-1142-171K
Language
English (United States)

With join optimization, the following join plan is generated when the collection of join columns (color, size, and options) makes up a nonunique secondary index of the large table:

Operation Joined Tables Total Processing Time (seconds)
Spool 3: Product Join duplicated color, direct options                             0.31
Spool 5: Product Join duplicated size, direct 3                             4.43
Spool 1: Nested Join duplicated 5, index widgets                           22.73

Completion Time

The total estimated execution time is 27.26 seconds.

The estimated performance improvement factor is 246.

EXPLAIN Output for Optimized Join Plan

Part of the EXPLAIN output for this optimized join plan is shown below.

2) Next, we do an all-AMPs RETRIEVE step from TEST.Color by way of an
   all-rows scan with no residual conditions into Spool 2, which is
   duplicated on all AMPs. The size of Spool 2 is estimated to be 4
   rows. The estimated time for this step is 0.08 seconds.
3) We execute the following steps in parallel.
   a) We do an all-AMPs JOIN step from Spool 2 (Last Use) by way of an
      all-rows scan, which is joined to TEST.Options. Spool 2 and
      TEST.Options are joined using a product join. The result goes
      into Spool 3, which is built locally on the AMPs. The size of
      Spool 3 is estimated to be 20 rows. The estimated time for this
      step is 0.23 seconds.
   b) We do an all-AMPs RETRIEVE step from TEST.Size by way of an
      all-rows scan with no residual conditions into Spool 4, which is
      duplicated on all AMPs. The size of Spool 4 is estimated to be 20
      rows. The estimated time for this step is 0.24 seconds.
4) We do an all-AMPs JOIN step from Spool 3 (Last Use) by way of an
   all-rows scan, which is joined to Spool 4 (Last Use). Spool 3 and
   Spool 4 are joined using a product join. The result goes into Spool
   5, which is duplicated on all AMPs. The size of Spool 5 is estimated
   to be 400 rows. The estimated time for this step is 4.19 seconds.
5) We do an all-AMPs JOIN step from Spool 5 (Last Use) by way of an
   all-rows scan, which is joined to TEST.Widgets by way of index # 4.
   Spool 5 and TEST.Widgets are joined using a nested join. The result
   goes into Spool 1, which is built locally on the AMPs. The size of
   Spool 1 is estimated to be 200 rows. The estimated time for this
   step is 22.73 seconds.