Polyglot Graph Databases using OCL as pivot
0 likes•520 views
This document discusses using the Object Constraint Language (OCL) as a pivot language for polyglot graph databases. OCL could serve as a common query language that is translatable to different native graph languages. This would eliminate the need for code generation between languages and allow a graph database to support multiple query languages. The document also proposes extensions to OCL, such as adding syntax for graph patterns, to make it more suitable as a graph query language.
![*Sparsity Technologies — Powering Extreme Data! sparsity–
technologies.com
º!*Sparsity Technologies — Powering Extreme
Data!
sparsity–
technologies.com!
º! Sparksee Graph Database
PGQL
SELECT pos.name
WHERE (p@1024)–[PERSON_WORKS]->w,
w <- [POSITON_WORKS] - pos,
NOT (w - [PREVIOUS_WORK] -> ())
Cypher
MATCH (p) – [:PERSON_WORKS] -> (w)<-[:POSITION_WORKS]-pos
WHERE NOT (w – [:PERVIOUS_WORK] -> ()) AND id(p) = 1024
RETURN pos.name
Conceptual
Schema!
Design! code!](https://image.slidesharecdn.com/graphtasparksee-160316123612/85/Polyglot-Graph-Databases-using-OCL-as-pivot-5-320.jpg)
Polyglot Graph Databases using OCL as pivot
- 1. *Sparsity Technologies — Powering Extreme Data! sparsity– technologies.com º! Sparksee Graph Database Polyglot graph databases using OCL as pivot March 2016. Raquel Pau!
- 2. *Sparsity Technologies — Powering Extreme Data! sparsity– technologies.com º! Sparksee Graph Database! • The development process • The languages mismatch problem • OCL as pivot graph query language Agenda!
- 3. *Sparsity Technologies — Powering Extreme Data! sparsity– technologies.com º!*Sparsity Technologies — Powering Extreme Data! sparsity– technologies.com! º! Sparksee Graph Database Position! Company! Person! Work! Startup! LargeCompany! 0..1 next! 0..1 previous! name:String! Conceptual Schema! Design! code! 0..1 ! *! *! OCL context System:query(Person p):String body:p.work->select(w| w.previous->isEmpty()).position.name !
- 4. *Sparsity Technologies — Powering Extreme Data! sparsity– technologies.com º!*Sparsity Technologies — Powering Extreme Data! sparsity– technologies.com! º! Sparksee Graph Database Position! Company! Person! Work! Startup! LargeCompany! name:String! PERSON_WORKS! POSITION_WORKS! PREVIOUS_WORK! COMPANY_WORKS! IS_A! IS_A! Normalization process due to the data model mismatch! Conceptual Schema! Design! code!
- 5. *Sparsity Technologies — Powering Extreme Data! sparsity– technologies.com º!*Sparsity Technologies — Powering Extreme Data! sparsity– technologies.com! º! Sparksee Graph Database PGQL SELECT pos.name WHERE (p@1024)–[PERSON_WORKS]->w, w <- [POSITON_WORKS] - pos, NOT (w - [PREVIOUS_WORK] -> ()) Cypher MATCH (p) – [:PERSON_WORKS] -> (w)<-[:POSITION_WORKS]-pos WHERE NOT (w – [:PERVIOUS_WORK] -> ()) AND id(p) = 1024 RETURN pos.name Conceptual Schema! Design! code!
- 6. *Sparsity Technologies — Powering Extreme Data! sparsity– technologies.com º!*Sparsity Technologies — Powering Extreme Data! sparsity– technologies.com! º! Sparksee Graph Database Conceptual schemas — Classes — N-ary relationships — Properties — Reification — Taxonomies Property graph model — Nodes — Binary Relationships — Properties Data Model Mismatches!
- 7. *Sparsity Technologies — Powering Extreme Data! sparsity– technologies.com º!*Sparsity Technologies — Powering Extreme Data! sparsity– technologies.com! º! Sparksee Graph Database DESIRED APPROACH! REAL APPROACH! 1. Specification! 2. Implementation! 3. QA Tests! 4. New changes! 5. Specification! 6. Implementation! 7. ….! 1. Specification! 2. Implementation! 3. QA Tests! 4. New changes! 5. Implementation! 6. QA Tests!
- 8. *Sparsity Technologies — Powering Extreme Data! sparsity– technologies.com º!*Sparsity Technologies — Powering Extreme Data! sparsity– technologies.com! º! Sparksee Graph Database Why not UML/OCL as a pivot language?
- 9. *Sparsity Technologies — Powering Extreme Data! sparsity– technologies.com º!*Sparsity Technologies — Powering Extreme Data! sparsity– technologies.com! º! Sparksee Graph Database UML/OCL as a pivot language 1. Code generation becomes unnecessary. 2. A database could become polyglot due to its high expressivity. 3. Executable specifications => Specifications become code. 4. The same constructions are valid for constraints and update operations.
- 10. *Sparsity Technologies — Powering Extreme Data! sparsity– technologies.com º!*Sparsity Technologies — Powering Extreme Data! sparsity– technologies.com! º! Sparksee Graph Database OCL as Graph Query Language: Desired extensions • Graph patterns: OCL has not an specific syntax to express graph patterns, but allows to define variables and their values. • Supported types: Graph is needed to support graph operations such as shortest path. • Syntactic sugar: needed to express sorting clauses and projections.
- 11. *Sparsity Technologies — Powering Extreme Data! sparsity– technologies.com º! Sparksee Graph Database! Thanks! Q&A