Re: Implementing Incremental View Maintenance - Mailing list pgsql-hackers

From Tatsuo Ishii
Subject Re: Implementing Incremental View Maintenance
Date
Msg-id 20200508.155253.1094555961446663417.t-ishii@sraoss.co.jp
Whole thread Raw
In response to Re: Implementing Incremental View Maintenance  (Andy Fan <zhihui.fan1213@gmail.com>)
List pgsql-hackers
>> +1,  This is a smart idea.   How did you test it?  AFAIK, we can test it
> with:
> 
> 1.  For any query like SELECT xxx,  we create view like CREATE MATERIAL VIEW
> mv_name as SELECT xxx;  to test if the features in the query are supported.

No I didn't test the correctness of IVM with TPC-DS data for now.
TPC-DS comes with a data generator and we can test IVM something like:

SELECT * FROM IVM_vew EXCEPT SELECT ... (TPC-DS original query);

If this produces 0 row, then the IVM is correct for the initial data.
(of course actually we need to add appropreate ORDER BY and LIMIT
clause to the SELECT statement for IVM if neccessary).

> 2.  Update the data and then compare the result with SELECT XXX with SELECT
> * from  mv_name to test if the data is correctly sync.

I wanted to test the data updating but I am still struggling how to
extract correct updating data from TPC-DS data set.

Best regards,
--
Tatsuo Ishii
SRA OSS, Inc. Japan
English: http://www.sraoss.co.jp/index_en.php
Japanese:http://www.sraoss.co.jp



pgsql-hackers by date:

Previous
From: Masahiko Sawada
Date:
Subject: Re: Fix pg_buffercache document
Next
From: Fabien COELHO
Date:
Subject: Re: PG 13 release notes, first draft