Temporal database
1 like402 views
A temporal database allows data to be tracked over time by attaching time periods to records. There are two approaches - valid time tracks when data is true in the real world, while transaction time tracks when it is stored in the database. The presentation discusses these approaches, provides examples of inserting, deleting, and updating temporal data records, and explains how temporal data can be useful for applications that involve historical or future data.
Temporal database
- 1. Temporal Database Linh Chau Software Engineer November 8th, 2017
- 2. 22 Agenda ▪ Introduction ▪ Bi-temporality ▪ Example ▪ Q&A
- 3. 33 What is temporal database?
- 4. 44 Traditional way ID Name Price (x1000 VND) 1 Coconut Oil 650 2 Honey 220 3 Medical Oil 120
- 5. 55 Traditional way ▪ Store only current state of data. ▪ Modify data will overlap the existing data.
- 6. 66 Questions ▪ How to retrieve history data? ▪ What happen if you delete data by accident? ▪ What if I need to retrieve the data in the past, or insert a reserved data for future use? → We need something like time period to attach to the data, this is temporal data. A temporal database is a database with built-in support for handling data involving time.
- 7. 77 Temporal database ▪ Time period is attached to the data (time varying data). ▪ Express when the data is valid or stored in the database. ▪ Divided into two approaches: ▪ Time period when the data is true in the real world (valid time). ▪ Time period when data is stored in database (transaction time).
- 8. 88 Valid time approach ID Name Price (x1000 VND) 1 Coconut Oil 650 2 Honey 220 3 Medical Oil 120
- 9. 99 Valid time approach ID Name Price Valid From Valid Till 2 Honey 250 01/01/2016 01/09/2017 2 Honey 220 01/09/2017 02/02/2020 2 Honey 300 03/03/2025 04/04/2050 2 Honey 500 06/09/3000 09/06/3005
- 10. 1010 Valid time approach ▪ Records the time when a fact is true in the real world. ▪ Represents data in the past, current for future ▪ Enables queries that show the state of the fact at a given time.
- 11. 1111 Transaction time approach ID Name Price Valid From Valid Till Known From Known Till 2 Honey 250 01/01/2016 01/09/2017 08/08/2017 31/12/9999 2 Honey 220 01/09/2017 02/02/2020 08/08/2017 31/12/9999 2 Honey 300 03/03/2025 04/04/2050 08/08/2017 31/12/9999 2 Honey 500 06/09/3000 09/06/3005 08/08/2017 08/11/2017 2 Honey 5000 06/09/3000 09/06/3005 08/11/2017 31/12/9999
- 12. 1212 Transaction time approach ▪ Records the time period during which a database entry is accepted as correct. ▪ Enables queries that show the state of the database at a given time. ▪ Only occur in the past or up to the current time. ▪ In a transaction time table, records are never deleted.
- 14. 1414 Bi-Temporality 07.01 13.02 27.05 25.07 10.01 15.02 07.06 19.06 Validity Knowing China Switzerland Japan Vietnam
- 15. 1515 Bi-Temporality 07.01 13.02 27.05 25.07 10.01 15.02 07.06 19.06 Validity Knowing China Switzerland Japan Vietnam
- 16. 1616 Bi-Temporality ▪ In bi-temporality, we take care both point of view the validity and point of view of the change. ▪ Through them, we get the exact valid information of the system at specific time. ▪ No physical delete of the row in database.
- 17. 1717 Use cases ▪ You want to reprint of a customer’s invoices of July 2017. ▪ What was the profit last year? ▪ Where do you enter the new address of this customer since the first next month? ▪ The client wants to book a reservation of your hotel next year.
- 18. 1818 Application domains ▪ Financial apps: portfolio management, accounting, banking. ▪ Record-keeping: personal, medical, inventory management. ▪ Scheduling apps: airline, train, hotel reservation... ▪ CRM apps (customer history / future) ⇒ Time varying data has always been in business requirements.
- 19. 1919 Example
- 20. 2020 Insert Insert record “A”, valid from 01.01.2016.
- 24. 2424 Insert Action Data Valid From Valid Till Known From Known Till Insert A 01.01.2016 31.12.9999 sysdate 31.12.9999 ▪ 01.01.2016: when the data is valid ▪ 31.12.9999: infinity ▪ sysdate: “now”
- 28. 2828 Delete Action Data Valid From Valid Till Known From Known Till Update A - - - sysdate ▪ “-”: no change
- 29. 2929 Close Update record “A”, set valid until 06.06.2016.
- 32. 3232 Close Action Data Valid From Valid Till Known From Known Till Update A - - - sysdate Insert A 01.01.2016 06.06.2016 sysdate 31.12.9999
- 33. 3333 Update future with infinity Update record by value “B”, effect from 06.06.2016.
- 34. 3434 Update future with infinity sysdate1 Valid Know 01.01.2016 A
- 35. 3535 Update future with infinity sysdate1 Valid Know 01.01.2016 A A sysdate2 06.06.2016 B
- 36. 3636 Update future with infinity Action Data Valid From Valid Till Known From Known Till Update A - - - sysdate Insert A 01.01.2016 06.06.2016 sysdate 31.12.9999 Insert B 06.06.2016 31.12.9999 sysdate 31.12.9999
- 37. 3737 Update the past with infinity Update record by value “C”, effect from 03.03.2016.
- 38. 3838 Update the past with infinity sysdate1 Valid Know 01.01.2016 A A sysdate2 06.06.2016 B
- 39. 3939 Update the past with infinity sysdate1 Valid Know 01.01.2016 A A sysdate2 06.06.2016 B sysdate3 C 03.03.2016
- 40. 4040 Update the past with infinity Action Data Valid From Valid Till Known From Known Till Update A - - - sysdate Update B - - - sysdate Insert A 01.01.2016 03.03.2016 sysdate 31.12.9999 Insert C 03.03.2016 31.12.9999 sysdate 31.12.9999
- 41. 4141 Update the past with close date after first update Update record by value “C”, effect from 03.03.2016 to 09.09.2016.
- 42. 4242 Update the past with close date after first update sysdate1 Valid Know 01.01.2016 A A sysdate2 06.06.2016 B
- 43. 4343 Update the past with close date after first update sysdate1 Valid Know 01.01.2016 A A sysdate2 06.06.2016 B sysdate3 C 03.03.2016 09.09.2016 B
- 44. 4444 Update the past with close date after first update Action Data Valid From Valid Till Known From Known Till Update A - - - sysdate Update B - - - sysdate Insert C 03.03.2016 09.09.2016 sysdate 31.12.9999 Insert A 01.01.2016 03.03.2016 sysdate 31.12.9999 Insert B 09.09.2016 31.12.9999 sysdate 31.12.9999
- 45. 4545 Update the past with close date before first update Update record by value “C”, effect from 03.03.2016 to 04.04.2016
- 46. 4646 Update the past with close date before first update sysdate1 Valid Know 01.01.2016 A A sysdate2 06.06.2016 B
- 47. 4747 Update the past with close date before first update sysdate1 Valid Know 01.01.2016 A A sysdate2 06.06.2016 B sysdate3 C 03.03.2016 09.09.2016 A A 04.04.2016
- 48. 4848 Update the past with close date before first update Action Data Valid From Valid Till Known From Known Till Update A - - - sysdate Keep B - - - - Insert C 03.03.2016 04.04.2016 sysdate 31.12.9999 Insert A 01.01.2016 03.03.2016 sysdate 31.12.9999 Insert A 04.04.2016 06.06.2016 sysdate 31.12.9999
- 49. 4949 When to use Ask yourself two questions: ▪ Does your business need to know the situation as it was known at a particular date? ▪ Does your business use information that was effective in the past or will become effective in the future? ⇒ If you answer YES on one or both of these question then your data varies over time and you could consider adopting a temporal approach.
- 50. 5050 Q&A
- 51. 5151
- 52. 5252 References ▪ http://www.datasciencecentral.com/profiles/blogs/temporal-databases-why- you-should-care-and-how-to-get-started (3 parts) ▪ https://www.linkedin.com/pulse/emerging-data-design-bitemporal-mike- lapenna (2 parts) ▪ Kudos to my ex-colleagues.
- 53. 5353 IT SecurityNEVISIT Consulting Application Management Software Solutions Linh Chau, Software Engineer linh.chau@adnovum.ch www.adnovum.ch