language integrated provenance
play

Language-integrated Provenance Stefan Fehrenbach James Cheney - PowerPoint PPT Presentation

Language-integrated Provenance Stefan Fehrenbach James Cheney PPDP 2016 A database Agencies oid name based_in phone 1 EdinTours Edinburgh 8740 2489 123 2 Burnss Glasgow 9307 2394 104 ExternalTours oid name destination type


  1. Language-integrated Provenance Stefan Fehrenbach James Cheney PPDP 2016

  2. A database Agencies oid name based_in phone 1 EdinTours Edinburgh 8740 2489 123 2 Burns’s Glasgow 9307 2394 104 ExternalTours oid name destination type price in £ 3 EdinTours Edinburgh bus 20 4 EdinTours Loch Ness bus 50 5 EdinTours Loch Ness boat 200 6 EdinTours Firth of Forth boat 50 7 Burns’s Islay boat 100 8 Burns’s Mallaig train 40 2

  3. Language-integrated query Agencies query { oid name based_in phone for (a <-- agencies) 1 EdinTours Edinburgh 8740 2489 123 for (e <-- externalTours) where (a.name == e.name 2 Burns’s Glasgow 9307 2394 104 && e.type == “boat”) [(name = e.name, [(phone = a.phone)] ExternalTours } oid name destination type price in £ 3 EdinTours Edinburgh bus 20 4 EdinTours Loch Ness bus 50 5 EdinTours Loch Ness boat 200 6 EdinTours Firth of Forth boat 50 7 Burns’s Islay boat 100 8 Burns’s Mallaig train 40 3

  4. Language-integrated query Agencies query { oid name based_in phone for (a <-- agencies) 1 EdinTours Edinburgh 8740 2489 123 for (e <-- externalTours) where (a.name == e.name 2 Burns’s Glasgow 9307 2394 104 && e.type == “boat”) [(name = e.name, [(phone = a.phone)] ExternalTours } oid name destination type price in £ 3 EdinTours Edinburgh bus 20 name phone 4 EdinTours Loch Ness bus 50 EdinTours 8740 2489 123 5 EdinTours Loch Ness boat 200 EdinTours 8740 2489 123 6 EdinTours Firth of Forth boat 50 Burns’s 9307 2394 104 7 Burns’s Islay boat 100 8 Burns’s Mallaig train 40 4

  5. Where-provenance Agencies query { oid name based_in phone for (a <-- agencies) 1 EdinTours Edinburgh 8740 2489 123 for (e <-- externalTours) where (a.name == e.name 2 Burns’s Glasgow 9307 2394 104 && e.type == “boat”) [(name = e.name, [(phone = a.phone)] ExternalTours } oid name destination type price in £ 3 EdinTours Edinburgh bus 20 name phone 4 EdinTours Loch Ness bus 50 EdinTours 8740 2489 123 5 EdinTours Loch Ness boat 200 EdinTours 8740 2489 123 6 EdinTours Firth of Forth boat 50 Burns’s 9307 2394 104 7 Burns’s Islay boat 100 8 Burns’s Mallaig train 40 5

  6. Where-provenance Agencies query { oid name based_in phone for (a <-- agencies) 1 EdinTours Edinburgh 8740 2489 123 for (e <-- externalTours) where (a.name == e.name 2 Burns’s Glasgow 9307 2394 104 && e.type == “boat”) [(name = e.name, [(phone = a.phone)] ExternalTours } oid name destination type price in £ 3 EdinTours Edinburgh bus 20 name phone 4 EdinTours Loch Ness bus 50 EdinTours 8740 2489 123 5 EdinTours Loch Ness boat 200 EdinTours 8740 2489 123 6 EdinTours Firth of Forth boat 50 Burns’s 9307 2394 104 7 Burns’s Islay boat 100 8 Burns’s Mallaig train 40 6

  7. Where-provenance Agencies query { oid name based_in phone for (a <-- agencies) 1 EdinTours Edinburgh 8740 2489 123 for (e <-- externalTours) where (a.name == e.name 2 Burns’s Glasgow 9307 2394 104 && e.type == “boat”) [(name = e.name, [(phone = a.phone)] ExternalTours } oid name destination type price in £ 3 EdinTours Edinburgh bus 20 name phone 4 EdinTours Loch Ness bus 50 EdinTours 8740 2489 123 5 EdinTours Loch Ness boat 200 EdinTours 8740 2489 123 6 EdinTours Firth of Forth boat 50 Burns’s 9307 2394 104 7 Burns’s Islay boat 100 8 Burns’s Mallaig train 40 7

  8. Where-provenance Agencies query { oid name based_in phone for (a <-- agencies) 1 EdinTours Edinburgh 8740 2489 123 for (e <-- externalTours) where (a.name == e.name 2 Burns’s Glasgow 9307 2394 104 && e.type == “boat”) [(name = e.name, [(phone = a.phone)] ExternalTours } oid name destination type price in £ 3 EdinTours Edinburgh bus 20 name phone 4 EdinTours Loch Ness bus 50 EdinTours 8740 2489 123 5 EdinTours Loch Ness boat 200 EdinTours 8740 2489 123 6 EdinTours Firth of Forth boat 50 Burns’s 9307 2394 104 7 Burns’s Islay boat 100 8 Burns’s Mallaig train 40 8

  9. Lineage (why-provenance) Agencies query { oid name based_in phone for (a <-- agencies) 1 EdinTours Edinburgh 8740 2489 123 for (e <-- externalTours) where (a.name == e.name 2 Burns’s Glasgow 9307 2394 104 && e.type == “boat”) [(name = e.name, [(phone = a.phone)] ExternalTours } oid name destination type price in £ 3 EdinTours Edinburgh bus 20 name phone 4 EdinTours Loch Ness bus 50 EdinTours 8740 2489 123 5 EdinTours Loch Ness boat 200 EdinTours 8740 2489 123 6 EdinTours Firth of Forth boat 50 Burns’s 9307 2394 104 7 Burns’s Islay boat 100 8 Burns’s Mallaig train 40 9

  10. Lineage (why-provenance) Agencies query { oid name based_in phone for (a <-- agencies) 1 EdinTours Edinburgh 8740 2489 123 for (e <-- externalTours) where (a.name == e.name 2 Burns’s Glasgow 9307 2394 104 && e.type == “boat”) [(name = e.name, [(phone = a.phone)] ExternalTours } oid name destination type price in £ 3 EdinTours Edinburgh bus 20 name phone 4 EdinTours Loch Ness bus 50 EdinTours 8740 2489 123 5 EdinTours Loch Ness boat 200 EdinTours 8740 2489 123 6 EdinTours Firth of Forth boat 50 Burns’s 9307 2394 104 7 Burns’s Islay boat 100 8 Burns’s Mallaig train 40 10

  11. Language-integrated provenance builds on Language-integrated query Provenance in databases • LINQ: … Meijer, Beckman, Bierman . • Why and where: … Buneman, SIGMOD 2006 Khanna, Tan . ICDT 2001 • The script- writer’s dream. Cooper . • On the expressiveness of implicit DBPL 2009 provenance … Buneman, Cheney, Vansummeren . TODS 2008 • Query shredding: … Cheney, • Perm: … Glavic, Alonso . ICDE 2009 Lindley, Wadler . SIGMOD 2014 • Effective quotation: … Cheney, • Using SQL for efficient generation Lindley, Radanne, Wadler . and querying … Glavic, Miller, PEPM 2014 Alonso . Buneman Festschrift 2013 11

  12. This talk 1. Why? W 2. Language-integrated where-provenance in Links W 3. Rewriting Links to Links The paper 4. User-defined where-provenance L 5. Lineage in Links and its translation to Links 6. Performance 12

  13. Why? Easy access to data and its provenance Provenance is not data – it is metadata data without provenance is less than complete provenance on its own is quite useless data with fake provenance is an affront Calculating provenance and propagating it manually is hard or least cumbersome enough to want to automate it 13

  14. W Where-provenance in Links Mark data carrying provenance metadata with an abstract type: Prov( O ) O is a base type Γ ⊢ M : Prov( O ) Γ ⊢ M : Prov( O ) Two operations: Γ ⊢ data M : O Γ ⊢ prov M : (String, String, Int) No constructor! – only the runtime can create provenance-annotated data Print as a comment, because it cannot appear in a program anyway: “ EdinTours ” #(“Agencies”, “name”, 2) 14

  15. Language-integrated query in Links query { for (a <-- agencies) for (e <-- externalTours) where (a.name == e.name && e.type == “boat”) [(name = e.name, [(phone = a.phone)] } 15

  16. Language-integrated query in Links query { var agencies = table “Agencies” for (a <-- agencies) with (oid: Int, for (e <-- externalTours) with (name: String, where (a.name == e.name with (based_in: String, && e.type == “boat”) with (phone: String) [(name = e.name, where phone prov default , [(phone = a.phone)] where name prov default } 16

  17. Language-integrated query in Links query { var agencies = table “Agencies” for (a <-- agencies) with (oid: Int, for (e <-- externalTours) with (name: String, where (a.name == e.name with (based_in: String, && e.type == “boat”) with (phone: String) [(name = e.name, where phone prov default , [(phone = a.phone)] where name prov default } agencies : [(oid: Int, agencies : [(name: String, agencies : [(based_in: String, agencies : [(phone: String)] 17

  18. Language-integrated query in Links query { var agencies = table “Agencies” for (a <-- agencies) with (oid: Int, for (e <-- externalTours) with (name: String, where (a.name == e.name with (based_in: String, && e.type == “boat”) with (phone: String) [(name = e.name, where phone prov default , [(phone = a.phone)] where name prov default } [(name = “ EdinTours ”, #(“ ExternalTours ”, “name”, 5) agencies : [(oid: Int, [( phone = “8740 2489 123”), #(“Agencies”, “phone”, 1) agencies : [(name: String, [ (name = “ EdinTours ”, #(“ ExternalTours ”, “name”, 6) agencies : [(based_in: String, [( phone = “8740 2489 123”), #(“Agencies”, “phone”, 1) agencies : [(phone: String)] [ (name = “Burns’s”, #(“ ExternalTours ”, “name”, 7) [( phone = “9307 2394 104”)] #(“Agencies”, “phone”, 2) : [(name: String, phone: String)] 18

Download Presentation
Download Policy: The content available on the website is offered to you 'AS IS' for your personal information and use only. It cannot be commercialized, licensed, or distributed on other websites without prior consent from the author. To download a presentation, simply click this link. If you encounter any difficulties during the download process, it's possible that the publisher has removed the file from their server.

Recommend


More recommend