Re: What is integrated?
Date: Sun, 10 Nov 2002 02:00:56 +0000
Message-ID: <tversusqgv74c4oa219emcvsi1n1orilt4_at_4ax.com>
Paul Vernon wrote:
>"Scotty" <invallid_at_invalid.spam> wrote in message
>news:0ucnsu4p1a30feev13v556269hjhfbfmh3_at_4ax.com...
>> Hi all, I'm just writing a paper on the Operational Data Store's role
>> and am a little unsure as to what is being referred to as 'integrated'
>> As in;
>> Subject oriented
>> Integrated
>> Volatile
>> etc..
>>
>> Would I be right in assuming that integrated would mean that two data
>> types are merged to one? As in a date format medium to long or a byte
>> to a Varchar?
>
>I think you will find that 'Integrated means the you need to integrate
>multiple *databases*, which means more things that 'integrating data types
>(although that is usually one of the tasks).
Here's what Ive written so far:
Integrated
Integration of the data is one of the most important aspects of the operational data stores role. The process of integration aims to homogenise the several heterogeneous data sources throughout an organisations IS infrastructure. This will include such tasks as; integrating similar data from different sources with dissimilar naming conventions, this is often needed as various applications throughout a business will use diverse attribute naming for similar data, i.e. Customer_Id and Cust_No. It may also be the case that the data is represented in different domains, for instance; date as 1-jan-2002 or 01-01-02 and data types represented as byte, int, long or varchar(20) etc.. These data sets cannot be held within a single data store until they are transformed into one type, the task of this is often carried out with
Extract Transform and Load (ETL) tools, which have the complex job of extracting data from legacy data stores, transforming it into one type and then load into the ODS. These ETL tools are often responsible for the success of an ODS project and as such I will be discussing them in further detail within this paper.
- :: -
Christ it sounds like a load of gibberish when I re-read it.
>Also you do not show a good understanding of what a data type (aka domain)
>actually is. The things you mentioned are physical storage concepts not data
>types. Here are some data types
>
>Boolean
>TimePoint
>UnicodeCharacterString
>Mass
>Length
>Dollars
>Euros
>TimeDuration
>Rational
>Integer
>PersonId
>ISO3166Country
Thanks Paul. I seem to be making a habit of getting just about everything wrong around here. Received on Sun Nov 10 2002 - 03:00:56 CET