BI 4 ALL Demonstration 02 Data Models www

Скачать презентацию BI 4 ALL Demonstration 02 Data Models www Скачать презентацию BI 4 ALL Demonstration 02 Data Models www

41768b0b635894fe0dea1068429ab4ab.ppt

  • Количество слайдов: 43

BI 4 ALL Demonstration 02 Data Models www. Instant. BI. com 1/7/2012 BI 4 ALL Demonstration 02 Data Models www. Instant. BI. com 1/7/2012

Agenda ® Some names you should know ® Key issues in modeling ® Traditional Agenda ® Some names you should know ® Key issues in modeling ® Traditional DW Design mechanisms ² Star Schemas ² Time Variance + Stability Analysis ® Business Intelligence Modeling Best Practice ® Terminology ® Naming Standards ® Each Type of Table ® Some details on BI 4 ALL ® Live demonstration of data models ® Next Steps Public Information Copyright 2012 – Instant Business Intelligence. 2

Some Names You Should Know ® Bill Inmon ² ‘Father of Data Warehousing’ ² Some Names You Should Know ® Bill Inmon ² ‘Father of Data Warehousing’ ² Prolific writer ² Was the ‘public pioneer’ of data warehousing ³ Pioneers have arrows in their backs!!! ² Originator of time variant + stability analysis data models ² Data modeling background ® Ralph Kimball ² Co-inventor of the Xerox Star workstation ³ This is a whole other story ² Co-Founder ³ Worlds of Metaphor Computer Systems best DSS Software (until IBM killed it) ² Solved the problem of DSS on motorola 68020 with 16 MB memory and 320 MB disk drive ³ By implementing dimensional models on an RDBMS ² Engineering Public Information Copyright 2012 – Instant Business Intelligence. background 3

Key Issues in Modelling ® Ease of use for the end user ® Speed Key Issues in Modelling ® Ease of use for the end user ® Speed of response for end user ® Ease of extensions/expansions for IT ® Ease of understanding for IT ® Ability to store large amounts of history ® Lowest cost development/maintenance of ETL to populate the model ® Could list many others… Public Information Copyright 2012 – Instant Business Intelligence. 4

BI Design Methods ® Three different methods widely used ® Choice of use depends BI Design Methods ® Three different methods widely used ® Choice of use depends on many factors ® Star Schema ² Transaction ³ Relational and ROLAP Online Analytical Processing ® Time Variance + Stability Analysis ² Changes to non-txn data eg. Customer ® Third Normal Form ² Volumes ² Until of changes are low recently all ODSs have been implemented as 3 NF Public Information Copyright 2012 – Instant Business Intelligence. 5

Why Use Star Schemas? ® Business people understand them ² Matches the business model Why Use Star Schemas? ® Business people understand them ² Matches the business model ² Often intuitively obvious ® Easy to query ® Supports complex questions easily ® No wrong answers due to join problems ® Excellent performance on star schema aware databases ² Oracle, Informix, DB 2 Public Information Copyright 2012 – Instant Business Intelligence. 6

What Does a Star Schema Look Like? Public Information Copyright 2012 – Instant Business What Does a Star Schema Look Like? Public Information Copyright 2012 – Instant Business Intelligence. 7

Why Use Integer Keys? ® Performance ² Integers is the fastest data type to Why Use Integer Keys? ® Performance ² Integers is the fastest data type to operate on ® Space and throughput ² Integers are shorter than account numbers ² Disk savings in tables and indexes ² Speed from disk to processor ® Flexibility ² Allows multi-level summary tables ² No IT involvement to create new summaries!! ® Large Stars!! Public Information Copyright 2012 – Instant Business Intelligence. 8

Why Use TV + SA? ® “Be able to show me the value of Why Use TV + SA? ® “Be able to show me the value of any field at any time in the past” ® Archival databases ® When people really do not know what they want ® Not easy to query ² Cannot give this to business people Public Information Copyright 2012 – Instant Business Intelligence. 9

What Does TV + SA Look Like? Each 3 NF entity passed to the What Does TV + SA Look Like? Each 3 NF entity passed to the DW is split into 3 (or more) entities based on stability analysis and access analysis. An element of time is added to the key. Public Information Copyright 2012 – Instant Business Intelligence. 10

Best Practice Since 1996 Business Users Desktops Highly Summarised (in star schemas or cubes) Best Practice Since 1996 Business Users Desktops Highly Summarised (in star schemas or cubes) Structured for Department Meta Data Everywhere 3 -5 years Lightly Summarised (in star schemas) Structured for Enterprise ODS integration/ transformation Operational Systems (System of Record) Public Information Copyright 2012 – Instant Business Intelligence. 1 -2 years 3 NF 5 -10 years Current Detail in archival data store (3 NF time variant model) Older detail (3 NF time variant model) 11

New BI 4 ALL Proposition Business Users Desktops Do not have to build data New BI 4 ALL Proposition Business Users Desktops Do not have to build data marts Highly Summarised (in star schemas or cubes) Views for Department Meta Data Everywhere 3 -5 years Virtual Lightly Summarised (star schemas) Structured for Enterprise ODS star integration/ schema transformation Operational Systems (System of Record) Public Information Copyright 2012 – Instant Business Intelligence. 3 -5 years 5 -10 years Current Detail in archival data store (star schema) Older detail (star schema) 12

BI 4 ALL Proposition ® Increased query performance ² Attributes ‘broken out’ to type BI 4 ALL Proposition ® Increased query performance ² Attributes ‘broken out’ to type 1 dims, directly joined to fact tables ² All joins through integer keys ² Almost no type 2 dimensions ² Same model can support ODS/DW query patterns ® You do not have to build data marts ² Can answer all your queries directly from the ODS/DW ² Approx 40% reduction in ETL Coding/Maintenance ² HW/SW cost avoidance for data marts ® All data readily available to end user tools ² Oracle Discoverer, Business Objects etc. ² All star schema ³ Even the archival data store Public Information Copyright 2012 – Instant Business Intelligence. 13

Enough of the Background Show me something new!!! Enough of the Background Show me something new!!!

Terminology ® Fact Table ²A table that stores facts ® Fact ²A number about Terminology ® Fact Table ²A table that stores facts ® Fact ²A number about the business ³ Eg. Revenue, profit, costs, price, quantity, interest, txn amount ® Dimension Table (Major Dimension) ²A table containing dimension fields about a specific major dimension of the business eg. Time, Product, Customer ® Dimension Field (Minor Dimension) ² Generally a descriptive character field describing some aspect of some ‘object’ that is involved in the business. (Time, Product, customer) ³ Eg. Time - day, week, month, quarter year ³ Eg. Product - name, colour, size, weight, height, group, category ³ Eg. Customer – name, birthdate, age, sex, marital status, life stage Public Information Copyright 2012 – Instant Business Intelligence. 15

Terminology ® Association Table 1 ²A table that associates two or more tables to Terminology ® Association Table 1 ²A table that associates two or more tables to each other. ³Eg. Customer 1 is wife of customer 2 ® Association Table 2 ²A table that provides time variance support within a star schema ® Type 1 Dimension Table ² Replace the record when some fields change ® Type 2 Dimension Table ² Close the previous record and open a new record when some specific fields on the dimension table change. Public Information Copyright 2012 – Instant Business Intelligence. 16

Terminology Integer Keys Fact Minor Dimension Public Information Copyright 2012 – Instant Business Intelligence. Terminology Integer Keys Fact Minor Dimension Public Information Copyright 2012 – Instant Business Intelligence. Fact Table Major Dimension 17

BI 4 ALL Naming Standards ® TD_ : Table - Dimension ® TF_ : BI 4 ALL Naming Standards ® TD_ : Table - Dimension ® TF_ : Table - Fact ® VM_ : View – di. Mension ® VF_ : View - Fact Public Information Copyright 2012 – Instant Business Intelligence. 18

Month – A Type 1 Dimension create view dbo. vm_month as Select a. pk_TD Month – A Type 1 Dimension create view dbo. vm_month as Select a. pk_TD 0005 pk_vm_month , day_date first_day_of_month , a. month_name_sdesc , a. month_in_year , a. calendar_qtr , a. month_in_qtr , a. financial_qtr , a. month_in_fncl_qtr , a. year_name , a. year_num , a. level_col from dbo. TD 0005 a where (a. level_col = 'level 2') or a. pk_TD 0005 = 0 go Public Information Copyright 2012 – Instant Business Intelligence. ® All Table access is via views ® Each dimension view has: ²A pk_ ² Defines a level for the table ² Exposes a zero row ² Opportunity to rename fields 19

Demographics – A Type 1 Dimension create view dbo. vm_all_demographic as select a. pk_TD Demographics – A Type 1 Dimension create view dbo. vm_all_demographic as select a. pk_TD 0001 pk_vm_all_demographic , a. integer_01 demo_type , a. varchar 255_01 demo_type_sdesc , a. varchar 255_02 demo_type_ldesc , a. money_01 income_low , a. money_02 income_high , a. varchar 255_03 income_band , a. integer_02 age_low , a. integer_03 age_high , a. char 1_01 gender , a. varchar 255_04 age_band , a. varchar 255_05 education_level , a. varchar 255_06 marital_status , a. varchar 255_07 life_stage , a. char 1_02 dependents_flag , a. integer_04 number_dependents , a. varchar 255_12 org_size_band , a. integer_05 number_employees_low , a. integer_06 number_employees_high , a. varchar 255_09 number_employees_band , a. money_01 revenue_low , a. money_02 revenue_high , a. varchar 255_03 revenue_band , a. varchar 255_10 industry_classification , a. integer_07 number_in_household , a. varchar 255_13 head_of_hh_name , a. level_col , a. dim_table_number from dbo. TD 0001 a where ( a. level_col = 'detail' and a. dim_table_number = 101) or a. pk_TD 0001 = 0 Public Information Copyright 2012 – Instant Business Intelligence. ® Note more complex dimension: ² Exposes all demographics for all types of parties ² Contains ‘bands’ as well as high low values ² Shows the dimension table number ² Shows the renaming of meaningless fields to meaningful fields Dimension table number 20

Product – A Type 2 Dimension create view dbo. vm_product as select a. pk_TD Product – A Type 2 Dimension create view dbo. vm_product as select a. pk_TD 0200 pk_vm_product , a. varchar 255_01 sku_number , a. varchar 255_02 barcode_num , a. varchar 255_03 prod_name , a. varchar 255_04 prod_sub_group , a. varchar 255_05 prod_group , a. varchar 255_06 prod_category , a. varchar 255_07 prod_class , a. varchar 255_08 prod_line , a. varchar 255_09 prod_size , a. varchar 255_10 prod_colour , a. varchar 255_11 prod_bundle_size , a. integer_01 shelf_life_days , a. money_01 unit_alloc_cost , a. money_02 unit_abc_cost , a. money_03 list_price_per_unit , a. varchar 255_22 unit_of_measure_code , a. varchar 255_23 unit_of_measure_sdesc , a. char 1_01 discount_allowed , a. decimal_nn discount_rate_1 -5 , a. varchar 255_nn discount_sdesc_1 -5 , a. varchar 255_nn discount_ldesc_1 -5 , a. char 1_02 returnable_flag , a. varchar 255_26 return_inspection_req_ldesc , a. char 1_03 sellable_flag , a. char 1_04 component_flag , a. char 1_05 requires_shipping_flag , a. char 1_06 taxable_flag , a. varchar 255_24 tax_type_1_sdesc , a. varchar 255_25 tax_type_2_sdesc , a. date_from , a. date_to , a. current_flag , a. level_col , a. dim_table_number from dbo. TD 0200 a where a. level_col = 'detail' and a. dim_table_number = 146 ® Type 2 dimensions have: ² Date from/to ² Current flag ® Note product dimension has a Public Information Copyright 2012 – Instant Business Intelligence. hierarchy for products ² From SKU -> prod line ² Actually not an ‘enforced’ hierarchy Date From Date To Current Flag 21

Reporting Structure – Type 2 Dimension ® Company structures change all create view dbo. Reporting Structure – Type 2 Dimension ® Company structures change all create view dbo. vm_comp_reporting_struct as select a. pk_TD 0004 pk_vm_comp_reporting_struct , a. date_from , a. date_to , a. current_flag , a. varchar 255_01 parent_company , a. varchar 255_02 org_name , a. varchar 255_03 region_name , a. varchar 255_04 channel_type , a. varchar 255_05 channel_category , a. varchar 255_06 channel_group , a. varchar 255_07 channel_name , a. varchar 255_08 team_type , a. varchar 255_09 team_category , a. varchar 255_10 team_group , a. varchar 255_11 team_name , a. varchar 255_12 person_type , a. varchar 255_13 person_category , a. varchar 255_14 person_group , a. varchar 255_15 person_name , a. level_col , a. dim_table_number from dbo. TD 0004 a where ( a. level_col = 'detail' And a. dim_table_number = 115) Or a. pk_TD 0004 = 0 the time ® A type 2 dimension is used to maintain reporting versions of company structure ® A separate recursive hierarchy is kept to maintain actual hierarchies in case of ‘missing’ levels of hierarchies Public Information Copyright 2012 – Instant Business Intelligence. 22

Partys , org_name , number_employees create view dbo. vm_party as select , gross_revenue pk_vm_party Partys , org_name , number_employees create view dbo. vm_party as select , gross_revenue pk_vm_party , gross_profit , prefix , net_profit , first_name , ss_org_key /* source system organisation key */ , second_name , phone_switch_num , third_name , start_up_date , surname , years_since_startup , suffix , blacklisted_ind /* 1 = Yes, 2 = No */ , concat_name , vio_flag /* very important organisation */ , address_to_name , ss_household_key /* source system household key */ , formal_salutation , vih_flag /* very important household */ , informal_salutation , cust_type_sdesc , written_salutation , cust_type_ldesc , email_address , web_address , tax_number , govt_identifier_type , govt_identifier , ss_person_key /* source system person key */ , marital_status , phone_home_num , phone_work_num , phone_mobile_num , phone_fax_num , birthdate , current_age , vip_flag , system_assigned_rating ² , person_assigned_rating , customer_flag , employee_flag , reseller_flag , sales_rep_flag , pmail_allowed_flag /* is paper mail allowed ? */ , email_allowed_flag /* is e-mail allowed ? */ , gender ® Partys can be people, companies, households or anything else you want to keep as a ‘party’ ® Party is a type 1 dimension Though we may implement as type 2 at #### Public Information Copyright 2012 – Instant Business Intelligence. 23

Types of Partys ® In the ‘as is’ model a Party can be: ² Types of Partys ® In the ‘as is’ model a Party can be: ² vm_person ² vm_household ² vm_organisation ² vm_sales_rep ² vm_all_cust ² vm_person_cust ² vm_org_cust ² vm_household_cust ² vm_vendor Public Information Copyright 2012 – Instant Business Intelligence. 24

Account/Address Indicator Tables create view dbo. vm_account_inds As select pk_vm_account_inds , credit_allowed_flag , payment_15_days_flag Account/Address Indicator Tables create view dbo. vm_account_inds As select pk_vm_account_inds , credit_allowed_flag , payment_15_days_flag , payment_30_days_flag , payment_45_days_flag , payment_60_days_flag , payment_90_days_flag , interest_penalty_flag , monthly_fee_flag , annual_fee_flag , level_col , dim_table_number create view dbo. vm_address_inds as select pk_vm_address_inds , weather_hazard_flag , flood_hazard_flag , storm_hazard_flag , cyclone_hazard_flag , hurricane_hazard_flag , drought_hazard_flag , earthquake_hazard_flag , level_col , dim_table_number Public Information Copyright 2012 – Instant Business Intelligence. ® For accounts we might record such things as credit allowed, payment terms and whether fees are levied. ® For addresses we might want a short cut to get to addresses that are subject to certain severe weather or other natural dangers. 25

Product Cross Holding Inds create view dbo. vm_prod_x_hold_inds as select pk_vm_prod_x_hold_inds , product_01 , Product Cross Holding Inds create view dbo. vm_prod_x_hold_inds as select pk_vm_prod_x_hold_inds , product_01 , product_02 , product_03 , product_04 , product_05 , product_06 , product_07 , product_08 , product_09 , product_10 , product_11 , product_12 , product_13 , product_14 , product_15 , product_16 , product_17 , product_18 , product_19 , product_20 , a. level_col , a. dim_table_number Public Information Copyright 2012 – Instant Business Intelligence. ® Sometimes we want to know who owns a combination of products. ® And sometimes we want to know who owns a combination of products but not some other specific product. ® This is much easier to do with product cross holding indicators. 26

Associating Partys ® Partys can be associated for: create view dbo. vf_party_asoc as select Associating Partys ® Partys can be associated for: create view dbo. vf_party_asoc as select pk_vf_party_asoc , asoc_key_vf_party_asoc , dk_vm_party_1 , dk_vm_party_2 , dk_vm_asoc_code , dk_vm_asoc_role , dk_vm_date_from , dk_vm_date_to , date_from , date_to , current_flag , asoc_code , asoc_role , asoc_amount 1 , asoc_amount 2 , party_1_type , party_2_type , a. asoc_table_number from dbo. TF 0002 a ² ² A reason (code) In a role For a period Specified by an amount ® Remember a party can be: ² ² ² ² ² vm_person vm_household vm_organisation vm_sales_rep vm_all_cust vm_person_cust vm_org_cust vm_household_cust vm_vendor ® Many of these associations are exposed separately through their own views Public Information Copyright 2012 – Instant Business Intelligence. 27

Partys and Addresses ® Partys are associated to addresses for: create view dbo. vf_party_address_asoc Partys and Addresses ® Partys are associated to addresses for: create view dbo. vf_party_address_asoc as select pk_vf_party_address_asoc , asoc_key_vf_party_address_asoc , dk_vm_party , dk_vm_address , dk_vm_asoc_code , dk_vm_asoc_role , dk_vm_date_from , dk_vm_date_to , date_from , date_to , current_flag , asoc_code , asoc_role , asoc_amount 1 , asoc_amount 2 , party_type , a. asoc_table_number from dbo. TF 0002 a Public Information Copyright 2012 – Instant Business Intelligence. ²A reason (code) ² In a role ² For a period ² Specified by an amount ® Addresses are considered independent of Partys ² That is, if no-one lives in a house the house still has an address ® Many of these associations are exposed separately through their own views 28

Some Simple Fact Tables create view dbo. vf_account_close_txn as select pk_vf_account_close_txn , dk_vm_account_txn_date , Some Simple Fact Tables create view dbo. vf_account_close_txn as select pk_vf_account_close_txn , dk_vm_account_txn_date , dk_vm_account_txn_minute , dk_vm_account , dk_vm_sub_campaign , dk_vm_contact_method , dk_vm_product , dk_vm_customer_demographic , dk_vm_currency , dk_vm_unit_of_measure , dk_vm_geography , dk_vm_geo_code , dk_vm_map_reference , dk_vm_card_type , dk_vm_card , dk_vm_txn_type , dk_vm_sales_rep , dk_vm_account_close_rsn_code , dk_vm_comp_reporting_struct , dk_vf_account_dims_asoc , dk_vf_customer_dims_asoc , dk_vf_campaign_dims_asoc , txn_amount , txn_units , txn_tstamp , appsys_reference_number_str , txn_sdesc , txn_ldesc , a. fact_table_number from dbo. TF 0101 a Public Information Copyright 2012 – Instant Business Intelligence. ® Lets look closely at the fact table ² ² ² ² Note a pk_ We put pks on fact tables now Used to lookup fact rows for updates Used to link asoc tables to fact tables Note the set of dimension tables that are linked Because it is an account close we would want to analyse this by many dimensions and it should run as fast as possible Note company reporting structure Note three keys ³ dk_vf_account_dims_asoc ³ dk_vf_customer_dims_asoc ³ dk_vf_campaign_dims_asoc ³ More ² on these later Then note there are some facts about the closed account 29

Invoice Lines create view dbo. vf_invoice_line as pk_vf_invoice_line , dk_vm_invoice_date , dk_vm_account , dk_vf_account_dims_asoc Invoice Lines create view dbo. vf_invoice_line as pk_vf_invoice_line , dk_vm_invoice_date , dk_vm_account , dk_vf_account_dims_asoc , dk_vm_sub_campaign , dk_vm_product , dk_vm_customer , dk_vf_customer_dims_asoc , dk_vm_customer_demographic , dk_vm_currency , dk_vm_unit_of_measure , dk_vm_geography , dk_vm_geo_code , dk_vm_map_reference , dk_vm_order_date , dk_vm_comp_reporting_struct , dk_vm_account_status , dk_vm_account_type , dk_vf_campaign_dims_asoc , appsys_reference_number_str , invoice_number , line_number_units , unit_base_amount , unit_extra_amount_01 , unit_tax_amount_02 , ext_base_amount , ext_extra_amount_01 , ext_tax_amount_02 , line_sdesc , line_ldesc , line_comments 1 , line_comments 2 , order_date , scheduled_delivery_date , scheduled_payment_date , number_lines , dk_vf_invoice_header , a. fact_table_number Public Information Copyright 2012 – Instant Business Intelligence. select ® Another very simple fact table ® Invoice lines ² Note link to invoice header fact row ² This is a fact<->fact table link ² Note again three columns ³ dk_vf_account_dims_asoc ³ dk_vf_customer_dims_asoc ³ dk_vf_campaign_dims_asoc ³ More on these later 30

Demonstration of Data Models Demonstration of Data Models

The Process of BI 4 ALL Modeling The Process of BI 4 ALL Modeling

BI 4 ALL Modeling ® Associations dramatically change the process of modeling ® Makes BI 4 ALL Modeling ® Associations dramatically change the process of modeling ® Makes the modelers life MUCH easier ® Eliminates some significant problems by design ® Dramatically enhances productivity ® Dramatically reduces development times Public Information Copyright 2012 – Instant Business Intelligence. 33

BI 4 ALL Dimension Modeling ® You identify the dimensional data ® Define and BI 4 ALL Dimension Modeling ® You identify the dimensional data ® Define and design dimension tables ² Do not worry about what fact tables it will be linked to ² Attempt to define unique discrete values only ² If there are large combinations of values break the input data up into more smaller dimension tables to keep the number of rows in the dimension tables low ® Decide which dimension tables need to be type 2 ² There are usually very few, and they usually have very few rows in them, product, geography are two examples ® Determine ‘real key’ in the data ® Map to a BI 4 ALL table or define a new table Public Information Copyright 2012 – Instant Business Intelligence. 34

BI 4 ALL Fact Modeling ® You identify the transaction/key measures data ® Define BI 4 ALL Fact Modeling ® You identify the transaction/key measures data ® Define and design fact tables ² Initially, only worry about linking dimensions that are unique to the fact row and dimensions obviously required for performance ² Do not worry about linking ‘low use’ dimensions ² Always put relevant association keys onto fact tables (party/account) ² Assume that ‘if it should be linked’ it will be linked ² Add extra keys for performance as the reports are designed and as analysis requirements become more clear ® Map to a BI 4 ALL table or define a new table Public Information Copyright 2012 – Instant Business Intelligence. 35

BI 4 ALL History Modeling ® Identify areas where you want to retain history BI 4 ALL History Modeling ® Identify areas where you want to retain history ² Entities/accounts ² You already defined might add ³ End of month outstanding amounts ³ Daily outstanding balances ® Decide how you are going to retain the history ² Type 2 dimension table? ² Develop a new association? ² Develop an event fact table with ‘current/previous’ amounts ³Eg. Banking transactions with before/after withdrawal balance Public Information Copyright 2012 – Instant Business Intelligence. 36

BI 4 ALL Modeling ® BI 4 ALL Modeling is far faster and easier BI 4 ALL Modeling ® BI 4 ALL Modeling is far faster and easier than other current methods of designing models ® It is also produces a faster/cheaper EDW ® No loss of functionality Public Information Copyright 2012 – Instant Business Intelligence. 37

Let’s Just Re-Summarise Because this is important Let’s Just Re-Summarise Because this is important

Best Practice Since 1996 Business Users Desktops Highly Summarised (in star schemas or cubes) Best Practice Since 1996 Business Users Desktops Highly Summarised (in star schemas or cubes) Structured for Department Meta Data Everywhere 3 -5 years Lightly Summarised (in star schemas) Structured for Enterprise ODS integration/ transformation Operational Systems (System of Record) Public Information Copyright 2012 – Instant Business Intelligence. 1 -2 years 3 NF 5 -10 years Current Detail in archival data store (3 NF time variant model) Older detail (3 NF time variant model) 39

New BI 4 ALL Proposition Business Users Desktops Do not have to build data New BI 4 ALL Proposition Business Users Desktops Do not have to build data marts Highly Summarised (in star schemas or cubes) Views for Department Meta Data Everywhere 3 -5 years Virtual Lightly Summarised (star schemas) Structured for Enterprise ODS star integration/ schema transformation Operational Systems (System of Record) Public Information Copyright 2012 – Instant Business Intelligence. 3 -5 years 5 -10 years Current Detail in archival data store (star schema) Older detail (star schema) 40

Next Steps Next Steps

Next Steps ® View a detailed live demonstration of the BI 4 ALL models Next Steps ® View a detailed live demonstration of the BI 4 ALL models that is confidential ® If interested perform a more detailed review of the models to determine fitness for purpose in your organisation Public Information Copyright 2012 – Instant Business Intelligence. 42

Summary ® Some names you should know ® Key issues in modeling ® Traditional Summary ® Some names you should know ® Key issues in modeling ® Traditional DW Design mechanisms ² Star Schemas ² Time Variance + Stability Analysis ® Business Intelligence Modeling Best Practice ® Terminology ® Naming Standards ® Each Type of Table ® Some details on BI 4 ALL ® Live demonstration of data models ® Next Steps Public Information Copyright 2012 – Instant Business Intelligence. 43




  • Мы удаляем страницу по первому запросу с достаточным набором данных, указывающих на ваше авторство. Мы также можем оставить страницу, явно указав ваше авторство (страницы полезны всем пользователям рунета и не несут цели нарушения авторских прав). Если такой вариант возможен, пожалуйста, укажите об этом.