data vault rmoug training days 2006 colorado convention center denver, colorado february 15-16
TRANSCRIPT
Data Vault
RMOUG Training Days
2006Colorado Convention Center
Denver, Colorado February 15-16
Data Vault; What’s The Combination?
Jeff MeyerEnterprise Data Integration – Oracle DBA
Department of Technology Services
Denver Public Schools
Data Vault
Who are we? DBAs Managers Analysts
Enterprise Data Warehouse Projects Currently in process Planned
Data Marts
Data Vault
Brief History and Revisit Some Definitions
Three Basic Building Blocks of the Data Vault
Advanced Features Questions
Data Vault
Brief History and Revisit Some Definitions
Three Basic Building Blocks of the Data Vault
Advanced Features Questions
Data Vault – Brief History and Revisit Some Definitions 1970 – Dr. E.F. Codd of IBM 1979 – First Working Relational
Database by Relational Software Incorporated
Oracle v2 1991 – William H. Inmon published
‘Building the Data Warehouse’
Data Vault – Brief History and Revisit Some Definitions Legacy System –
‘… any system that has been put into production.’
(para-phrased W.H. Inmon)
Operational Data Store – ‘… a subject-oriented, integrated, volatile, current
or near current collection of operational data.’
W.H. Inmon
Data Vault – Brief History and Revisit Some Definitions Data Warehouse –
‘… a subject-oriented, integrated, time-variant, non-volatile collection of data designed for support of business decisions’
W.H. Inmon
Data Vault – ‘… a detail-oriented, historical tracking and uniquely linked
set of normalized tables that support one or more functional areas of business.’
Dan Linstedt
Data Vault – Brief History and Revisit Some Definitions
Data Mart – ‘… a subset of a data warehouse, for use by a single
department or function.’
www.e-formation.co.nz/glossary.asp
Corporate Information Factory – ‘… the framework that exists that surrounds the data
warehouse; typically contains an ODS, a data warehouse, data marts, DSS applications, exploration warehouses, and so forth.’
W.H. Inmon
Data Vault – Brief History and Revisit Some Definitions
* Source: Bill Inmon and Claudia Imhoff
Data Vault – Why?
Why do we need it? We finally have a Data Model that will work for small,
medium, or large business Anyone building a Data Warehouse can use these
techniques.
We’ve got issues in constructing the data warehouse from 3rd normal form, or star schema form. There are inherent road blocks to each method that we must
solve technically through our Data Model.
Data Vault
Brief History and Revisit Some Definitions
Three Basic Building Blocks of the Data Vault
Advanced Features Questions
Data Vault – Three Basic Building Blocks
Hub – stand alone table; list of unique business keys; used for business identification
Satellite – descriptive data; historical data; used for descriptive information for the HUB or LINK
Link – associative table; list of unique relationships between keys; used for relationships between HUBs and LINKs
Data Vault – Three Basic Building Blocks
Preview
Hub Employees
Hub Schools
ELAName
EEOCDates
Hub Students
EEOCName
ShotsAddrs
Assign Enrollments
Data Vault – Three Basic Building Blocks
HUB
Primary Key
<Business Key>
Load DTS
Record Source
Primary Key
<Business Key>
Load DTS
Record Source
Sample Data Set “CUSTOMER”
CONTRACTS2-2-200093KFLLA10
CONTRACTS2-2-2000929ABC29859
CONTRACTS2-2-2000PAFJG28958
FINANCE2-2-2000PPRU_32597
SALES8-3-2001HUJI_BFIOQ6
SALES6-4-2001LLOA_82J5J5
CONTRACTS3-7-2000KKO92854_dd4
CONTRACTS1-25-2000DKEF3
CONTRACTS10-2-2000ABC925_24FN2
RCRD SRCLOAD DTSCUSTOMER #ID
MANUFACT10-12-2000ABC1234561
CONTRACTS2-2-200093KFLLA10
CONTRACTS2-2-2000929ABC29859
CONTRACTS2-2-2000PAFJG28958
FINANCE2-2-2000PPRU_32597
SALES8-3-2001HUJI_BFIOQ6
SALES6-4-2001LLOA_82J5J5
CONTRACTS3-7-2000KKO92854_dd4
CONTRACTS1-25-2000DKEF3
CONTRACTS10-2-2000ABC925_24FN2
RCRD SRCLOAD DTSCUSTOMER #ID
MANUFACT10-12-2000ABC1234561
A Hub is a list of unique business keys.
Data Vault – Three Basic Building Blocks
SATELLITE
Primary KeyLoad DTS
DetailBusiness Data
Aggregation Data
{Update User}{Update DTS}
Record Source
Primary KeyLoad DTS
DetailBusiness Data
Aggregation Data
{Update User}{Update DTS}
Record Source
CONTRACTS10-2-2000ABC925_24FN2
MANUFACT10-12-2000ABC1234561
RCRD SRCLOAD DTSCUSTOMER #ID
CONTRACTS10-2-2000ABC925_24FN2
MANUFACT10-12-2000ABC1234561
RCRD SRCLOAD DTSCUSTOMER #ID
CONTRACTSWorldwide Suppliers Inc10-14-20002
CONTRACTSWorldPart10-2-20002
CONTRACTSABC DEF Incorporated12-2-20001
MANUFACTABC Worldwide Suppliers, Inc10-31-20001
MANUFACTABC Suppliers, Inc10-14-20001
MANUFACTABC Suppliers10-12-20001
RCRD SRCNAMELOAD DTSCSID
CONTRACTSWorldwide Suppliers Inc10-14-20002
CONTRACTSWorldPart10-2-20002
CONTRACTSABC DEF Incorporated12-2-20001
MANUFACTABC Worldwide Suppliers, Inc10-31-20001
MANUFACTABC Suppliers, Inc10-14-20001
MANUFACTABC Suppliers10-12-20001
RCRD SRCNAMELOAD DTSCSID
CUSTOMER NAME SATELLITE
A Satellite is a time-dimensional table housing detailed information about the hub’s business keys.
Data Vault – Three Basic Building Blocks
Hub Employees
ELAName
EEOCDates
Employees HUB and some of its Satellites
Data Vault – Three Basic Building Blocks
LINK
Primary Key
Load DTS
Record Source
CONTRACTS10-2-2000ABC925_24FN2
MANUFACT10-12-2000ABC1234561
RCRD SRCLOAD DTSCUSTOMER #ID
CONTRACTS10-2-2000ABC925_24FN2
MANUFACT10-12-2000ABC1234561
RCRD SRCLOAD DTSCUSTOMER #ID
FINANCE10-14-2000CONT259101
FINANCE10-14-2000CONT212100
RCRD SRCLOAD DTSCONTACT #ID
FINANCE10-14-2000CONT259101
FINANCE10-14-2000CONT212100
RCRD SRCLOAD DTSCONTACT #ID
FINANCE10-14-20001012
FINANCE10-14-20001001
RCRD SRCLOAD DTSCONTACT IDCSID
FINANCE10-14-20001012
FINANCE10-14-20001001
RCRD SRCLOAD DTSCONTACT IDCSID
A Link is an associative or intersection table, representing theconnection between information between business elements.
Link Table
Data Vault – Three Basic Building Blocks
Hub Employees
ELAName
EEOCDates
Hub Schools
Geo CdAddr
FloorBldg
Assign
Sat
Hub and SatellitesHub and Satellites
Link and Satellites
Data Vault
Brief History and Revisit Some Definitions
Three Basic Building Blocks of the Data Vault
Advanced Features Questions
Data Vault – Advanced Features Point-In-Time –
A structure which sustains integrity of joins across time to all the SATELLITES that are connected to the HUB or LINK.
Bridge – A single row table that contains the latest Load Date Time
Stamp (DTS). Similar to Point-In-Time except it spans a subject-area or a schema.
User Grouping Link – The information provides the user with a customized view
from a reporting standpoint and does not affect the underlying information.
Data Vault – Advanced FeaturesPoint-In-Time (PIT)
MANUFACT10-12-2000ABC1234561
RCRD SRCLOAD DTSCUSTOMER #ID
MANUFACT10-12-2000ABC1234561
RCRD SRCLOAD DTSCUSTOMER #ID
ABC DEF Incorporated12-2-20001
ABC Worldwide Suppliers, Inc10-31-20001
NAMELOAD DTSCSID
ABC DEF Incorporated12-2-20001
ABC Worldwide Suppliers, Inc10-31-20001
NAMELOAD DTSCSID
123 World Drive12-5-20001
123 World Dr10-14-20001
ADDRESSLOAD DTSCSID
123 World Drive12-5-20001
123 World Dr10-14-20001
ADDRESSLOAD DTSCSID
12-5-200012-2-200012-5-20001
10-14-200012-2-200012-2-20001
10-31-2000
10-14-2000
NAME_LOAD_DTS
10-14-2000
10-14-2000
ADDRESS_LOAD_DTS
10-31-20001
10-14-20001
LOAD DTSCSID
12-5-200012-2-200012-5-20001
10-14-200012-2-200012-2-20001
10-31-2000
10-14-2000
NAME_LOAD_DTS
10-14-2000
10-14-2000
ADDRESS_LOAD_DTS
10-31-20001
10-14-20001
LOAD DTSCSID
A structure which sustains integrity of joins across time to all the satellites that are connected to the hub.
Customer Name Satellite Customer Address Satellite
Hub Key
Load Date
{Sat Load DTS}
{Sat Load DTS}
{Rec Source}
Hub Key
Load Date
{Sat Load DTS}
{Sat Load DTS}
{Rec Source}
Data Vault – Advanced FeaturesBridge
A single row table that contains the latest Load DTS with multiple columns. A Bridge is not a helper table.
Similar to a PIT Table except it spans or applies to a subject-area or schema. A PIT Table is HUB (LINK) and SATELLITE specific.
Data Vault – Advanced FeaturesUser Grouping Link
Primary Key
Load DTS
Record Source
Primary Key
Load DTS
Record Source
EXCEL10-2-2000Small Customers2
EXCEL10-12-2000Big Customers1
RCRD SRCLOAD DTSGrouping LabelID
EXCEL10-2-2000Small Customers2
EXCEL10-12-2000Big Customers1
RCRD SRCLOAD DTSGrouping LabelID
FINANCE10-14-2000ABC-1101
FINANCE10-14-2000ABC295882100
RCRD SRCLOAD DTSCustomer #ID
FINANCE10-14-2000ABC-1101
FINANCE10-14-2000ABC295882100
RCRD SRCLOAD DTSCustomer #ID
EXCEL10-14-20001011
EXCEL10-14-20001001
RCRD SRCLOAD DTSCustomer #Grp#
EXCEL10-14-20001011
EXCEL10-14-20001001
RCRD SRCLOAD DTSCustomer #Grp#
The User Grouping Link, allows users to “state” how they want roll-ups to occur – in situations where source data doesn’t exist.
BASE TABLE:
Data Vault – How is DPS using DV Hub_Students
Student_IDSIS_CodeLoad_DTSRec_SRC
Hub_Students
Student_IDSIS_CodeLoad_DTSRec_SRC
Hub_Schools
School_IDSchool_NumberLoad_DTSRec_SRC
Hub_Schools
School_IDSchool_NumberLoad_DTSRec_SRC
Hub_Employees
Employee_IDHR_Emp_IDDPSIDLoad_DTSRec_SRC
Hub_Employees
Employee_IDHR_Emp_IDDPSIDLoad_DTSRec_SRC
Lnk_School_Enrollments
Sch_Enr_IDSchool_IDStudent_IDGrade_NameLoad_DTSRec_SRC
Lnk_School_Enrollments
Sch_Enr_IDSchool_IDStudent_IDGrade_NameLoad_DTSRec_SRC
Lnk_Teacher_Schools
Teacher_School_IDSchool_IDEmployee_IDLoad_DTSRec_SRC
Lnk_Teacher_Schools
Teacher_School_IDSchool_IDEmployee_IDLoad_DTSRec_SRC
The direction of the arrows equate to crow’s feet.
Data Vault – Why is DPS using DV
Storage considerations. Vertical partitioning of data (rate of
change). All the FACTS all the TIME. Scalability and Extensibility.
Data Vault – What was not covered. How to apply Data Vault Modeling. Best practices. Lessons Learned. Dan Linstedt’s use of DECODE in
determining changed data capture. Who’s data is it? SLAs? The new regulations / compliance that will
affect all of us.
Data Vault – Questions?
Data Vault - References DATA VAULT OVERVIEW: THE NEXT EVOLUTION IN DATA MODELING
Dan Linstedt - Core Integration Partners, Inc.http://www.tdan.com/i021hy01.htm
DATA VAULT™ OVERVIEW THE NEXT EVOLUTION IN DATA MODELING SERIES 2Dan Linstedt - Core Integration Partners, Inc.http://www.tdan.com/i023hy02.htm
DATA VAULT - SERIES 3 END-DATES AND BASIC JOINS Dan Linstedt - Core Integration Partnershttp://www.tdan.com/i024hy02.htm
DATA VAULT - SERIES 4 LINK TABLES Dan Linstedt - Core Integration Partnershttp://www.tdan.com/i027ht04.htm
DATA VAULTTM OVERVIEW THE NEXT EVOLUTION IN DATA MODELING SERIES 5 – LOADING TABLES Dan Linstedt - Core Integration Partnershttp://www.tdan.com/i027ht04.htm
Data Vault Modeling – Class Materials and Notes; copyright 2002-2003Dan Linstedt – Core Integration Partnershttp://www.coreintegration.com
Home of the Data Vault; www.danlinsedt.com Audit the Data – or Else. Un-audited Data Access Puts Business at High Risk; Bloor, Robin
and Baroudi, Carol; Lumigent, Inc.; copyright 2004
Data Vault – Contact Information
JEFFREY [email protected]
Data Vault
RMOUG Training Days
2006Colorado Convention Center
Denver, Colorado February 15-16