Se ha denunciado esta presentación.
Utilizamos tu perfil de LinkedIn y tus datos de actividad para personalizar los anuncios y mostrarte publicidad más relevante. Puedes cambiar tus preferencias de publicidad en cualquier momento.

Lightning talk at PG Conf UK 2018

403 visualizaciones

Publicado el

Here's the presentation I gave at PGConf UK yesterday, with additional slides that I had to take ouHere's the presentation I gave at PGConf UK yesterday, with additional slides that I had to take out in order to fit within the 5-minute time limit. Slide 10 shows the result of running a transformation I developed, that adds audit columns to every table. This is a simple demonstration of the tool's ability to assist you in managing your databases - the transformation can also be used if you compare a data model to a database - even if the model doesn't actually include the audit columns, you can use the transformation to ensure that the comparison assumes they are there in the model, thus finding fewer differences between the model and the database.

Publicado en: Tecnología
  • Inicia sesión para ver los comentarios

Lightning talk at PG Conf UK 2018

  1. 1. Data Modelling It’s a lot more than drawing diagrams George McGeachie Metadata Matters Limited
  2. 2. My lightning talk at PG day in 2014 (I liked that venue – Horwood House) My entry in the 2nd Quadrant blog last week – used example of automating Data Vault design and creation https://blog.2ndquadrant.com/data-modelling-lot-just-diagram/ 2 This is my favourite theme
  3. 3. The right tool can give you a lot more than just this messy Diagram – would you want to work with this diagram? 3 This is my favourite theme
  4. 4. A data model is a lot more than just a Diagram 4 This is my favourite theme
  5. 5.  Check against your design standards  The tedious stuff, like making sure all your tables have the standard audit columns  Do you need JSON?  How much will this DB grow?  Managing (and comparing) schema, table & column versions  Building Data Vaults – see 2nd Quadrant blog 5 Automate tasks – before you build the database
  6. 6. Available automation A Contextual menu is one way of accessing automation – check the model, export JSON to a file, apply Naming Standards, adding audit columns
  7. 7. Add your own model checks, along with automatic fixing for those problems if possible (e.g. adding surrogate key) 7 Check your design meets your design standards
  8. 8. 8 Make sure all your tables have the standard audit columns Don’t blink or you’ll miss it
  9. 9. 9 The PDM – without audit columns
  10. 10. 10 Less than 2 seconds later …
  11. 11. 11 Do you need JSON? { "Name" : "Departments", "Code" : "Departments", "Fully Qualified Name" : "Group0.Departments", "Fully Qualified Code" : "Group0.Departments", "Owner" : "Group0", "Object Type" : "Table", "id" : "8731F3EE-8E53-46C6-A873-81C522F51717", "description" : "contains the names and heads of th "note" : "<NONE>", "Columns" : [ { "Name" : "DepartmentID", "Code" : "DepartmentID", "Fully Qualified Name" : "Departments "Fully Qualified Code" : "Group0.Depa "Object Type" : "Column", "id" : "D23F6064-87A8-4D1D-92D0-70F35 "description" : "short one", "note" : "<NONE>", "Data Type" : "INT4", "Length" : "4", "Precision" : "0", "Primary?" : "TRUE", "FK?" : "FALSE", "Mandatory?" : "TRUE",
  12. 12.  Extract current statistics, define growth rates 12 How much will this database grow?
  13. 13. Estimate of the size of the Database "PhysicalDataModel_1"... Number Estimated size Object ------------------------- ----------------------- ---------------------------------------- ------------ 1,556 312 KB Table "Contacts" 17,370 3,475 KB Table "Customers" 917 KB Index "IX_customer_name" 130 7 KB Table "Departments" 1,945 390 KB Table "Employees" 182 13 KB Table "FinancialCodes" 2,179 39 KB Table "FinancialData" 259 130 KB Table "MarketingInformation" 259 KB Long data types 274 KB Index "MarketingTextIndex" 1,379 9,651 KB Table "Products" 9,650 KB Long data types 34 KB Index "IX_product_name" 60 KB Index "IX_product_description" 41 KB Index "IX_product_size" 41 KB Index "IX_product_color" 28,453 619 KB Table "SalesOrderItems" 382,637,520 11,595,077 KB Table "SalesOrders" 3,268 467 KB Table "SpatialContacts" 467 25 KB Table "SpatialShapes" ------------------------- ----------------------- ---------------------------------------- ------------ 11,621,481 KB Total estimated space The data will be distributed on the following tablespaces: Estimated size Tablespace ----------------------- ---------------------------------------------- 1,367 KB system 13 Estimate Database size
  14. 14.  Write your own estimation script 14 If you don’t like the way it’s done
  15. 15.  Branching  Comparing versions 15 Versioning
  16. 16.  Check models into the repository, but don’t update the mainline until they’ve been approved 16 Check in model for peer review
  17. 17. 17 Integrate the 2nd Branch back into the 1st Branch Models updated with selected changes Still able to access version 1
  18. 18. 18 Simon and Hannu say … Page 53 • Understand Database Dependencies ◦ e.g. Table  View  Procedure I only have the first edition of this excellent book
  19. 19.  ETL Jobs  Forms and Reports  Applications  XML Message Schemas  Regulatory Requirements  Business Processes  Use Cases  JIRA tickets etc. 19 Databases have connections
  20. 20. 20 Choose your tools carefully What Tools are there? The big 3 ERwin, ER/Studio, PowerDesigner Others Dezign Sparx EA ModelRight Silverrun IBM Infosphere Data Architect Toad Data Modeller might not all support PG
  21. 21. George McGeachie Co-author of “Data Modeling Made Simple with PowerDesigner”, data modeller and strategist, SAP PowerDesigner trainer, and data modelling tool junkie. @metadatajunkie Blog – metadatajunkie.wordpress.com https://www.linkedin.com/in/georgemcgeachie/ George.McGeachie@MetadataMatters.com Mobile: +44 (0) 794 293 0648

×