Re: Star Schema from 'Flattened' CSV
From: Jerry Stuckle <jstucklex_at_attglobal.net>
Date: Fri, 3 Nov 2017 08:50:11 -0400
Message-ID: <othom2$gdl$1_at_jstuckle.eternal-september.org>
>
> No.
>
> dim_product
> - prod_id
> - prod_desc
>
> dim_vendor
> - vend_id
> - vend_name
>
> factless_quote
> - quote_id
> - prod_id
> - vend_id
> - price
> - <date data>
>
> fact_performance
> - quote_id
> - prod_id
> - vend_id
> - price
> - <date data>
> - <volume data>
>
> Jerry: Star-schema models do not usually fully comply with Codd-style
> normalisation.
>
> Geoff M
>
Date: Fri, 3 Nov 2017 08:50:11 -0400
Message-ID: <othom2$gdl$1_at_jstuckle.eternal-september.org>
On 11/3/2017 12:34 AM, Geoff Muldoon wrote:
> In article <be6b1334-18bd-4639-ad6b-01ae14732861_at_googlegroups.com>,
> transreductionist_at_gmail.com says...
>> >> Not sure what you are suggesting. Is this a correct assessment? I can go one of 2 ways: >> >> 1. Use 2 fact tables with 1 dimension: >> One fact table for vendor 1 >> One fact table for vendor 2
>
> No.
>
> dim_product
> - prod_id
> - prod_desc
>
> dim_vendor
> - vend_id
> - vend_name
>
> factless_quote
> - quote_id
> - prod_id
> - vend_id
> - price
> - <date data>
>
> fact_performance
> - quote_id
> - prod_id
> - vend_id
> - price
> - <date data>
> - <volume data>
>
> Jerry: Star-schema models do not usually fully comply with Codd-style
> normalisation.
>
> Geoff M
>
Geoff,
That depends on the schema. It can conform to a normalized database, but usually does not. This can easily lead to data integrity problems - which won't happen in a normalized database.
You can have both - but it takes more work.
-- ================== Remove the "x" from my email address Jerry Stuckle jstucklex_at_attglobal.net ==================Received on Fri Nov 03 2017 - 13:50:11 CET