Schema Mediated Exchange of Temporal XML Data

1
Schema Mediated Exchange of Temporal XML Data

• Curtis Dyreson Washington State
  University
• Richard T. Snodgrass University of Arizona
• Sabah Currim University of
  Arizona
• Faiz Currim University of
  Iowa

2
Scenario

• Genomic data from NCBI
• Data collection is growing/changing
• Want data and data provenance (who, what, when, )

3
Obtaining Web Data
Overwrite D
Write D
4
Data Evolves

• Download XML formatted data (as of January 1)
• ltgene name"TRY4"gt
• ltdescgttrypsin 4lt/descgt
• ltontology ref"MGI" function"unknown"/gt
• lt/genegt
• Download again (as of March 6)
• ltgene name"TRY4"gt
• ltdescgttrypsin 4, beta-cell receptorlt/descgt
• ltontology ref"MGI" function"synthesizes
  trypsinogen"/gt
• lt/genegt

5
Refreshing the Data (using SDOs)

• What about versions between D and Dold?
• Did I download valid data?
• My DB is pretty big

6
Did I Download the Right Data?

• Validate against schema

Schema
Namespace
Validating Parser
XML Data
7
Fragment of the Genomic Schema
ltelement name"gene"gt ltcomplexTypegt
ltattribute name"name" type"text"
use"required"/gt ltsequencegt
ltelement name"desc" type"string"/gt
ltelement ref"ontology" minOccurs"0"
maxOccurs"unbounded"/gt lt/sequencegt
lt/complexTypegt lt/elementgt
8
Uses of an XML Schema

• Validation
• XML editors
• Guides query formulation
• Query optimization
• Provides a web service binding

9
A Temporal Data Collection

• Validate the delta with the temporal schema
• cost is size of change

10
Outline

• Motivation
• tXSchema
• Architecture
• Summary

11
Goals for a Temporal Schema

• Make it easy to create a schema for temporal data
• Identify which data is temporal
• Upwards compatibility
• Minimal extensions of XML Schema
• Reuse off-the-shelf parsers/tools
• Support
• Valid and transaction time
• Data (element) versioning
• Schema versioning
• Logical/physical independence
• Flexible timestamp representation and location

12
Persistent Elements

• An item is an element that persists across
  snapshots.
• Item identifier (like a temporally-invariant key)
• lttxsitemIdentifiergt
• lttxsfield path_at_name/gt
• lt/txsitemIdentifiergt

January snapshot
March snapshot
13
Extend a Snapshot Schema

• Specify which elements are temporal
• Temporal elements have
• Item identifiers
• Simple constraints (state/event,
  existence/content-varying)
• ltelement name"gene"gt
• lttxstemporalgt
• lttxsitemIdentifiergt
• lttxsfield path"_at_name"/gt
• lt/txsitemIdentifiergt
• lttxstransactionTime kind"state"
  contentVarying"true"
• existenceVarying"no gaps"/gt
• lt/txstemporalgt
• .definition of gene from the snapshot
  schema omitted for space
• lt/elementgt

14
Versions

• A version is a change in an item.
• DOM inequivalence

January snapshot
March snapshot


15
Temporal Genomic Data
ltdataTemporalgt ltdatagtltgeneTemporal
itemRef"1"/gtlt/datagt ltgeneItem itemId"1"gt
ltgeneVersiongtlttime start"2005-01-01"
end"2005-03-05"/gt ltgene name"TRY4"gt
ltdescgttrypsin 4lt/descgt
ltontologyTemporal itemRef"2"/gt lt/genegt
lt/geneVersiongt ltgeneVersiongtlttime
start"2005-03-06" end"now"/gt next
version of gene lt/geneVersiongt
lt/geneItemgt ltontologyItem itemId"2"gt ontology
item lt/dataTemporalgt
16
Outline

• Motivation
• tXSchema
• Architecture
• Summary

17
Validating Temporal Data

• Snapshot data validated with a snapshot schema
• Construct a representational schema (details in
  paper)
• Can also validate the delta

Snapshot Schema
Namespace
Validating Parser
XML Data
18
Property of a Good Construction

• Every snapshot must conform to the snapshot
  schema

Temporal Schema
(Temporal) Validating Parser
Valid
Temporal data
Valid
19
Outline

• Motivation
• tXSchema
• Architecture
• Summary

20
Related Work Temporal XML

• Change detection and management
• Nguyen, Abiteboul, Cobena, Preda, SIGMOD 2001
• Xylemes Alerter, described in Data Engineering
  Bulletin, 2001
• Dyreson, Lin, Wang WWW 2004
• Leonardi, Bhowmick, ER 2006
• Representing time-varying XML documents
  (versioning)
• Chawathe, Abiteboul, Widom, ICDE 1998
• Dyreson, Böhlen, Jensen, VLDB 1999
• Chien, Tsotras, Zaniolo, VLDB 2000
• Marian, Abiteboul, Cobena, Mignet, VLDB 2001
• Buneman, Khanna, Tajima, Tan, SIGMOD 2002, TODS
  2004
• Rosado, Marquez, Gonzalez, ECDM 2006
• XML Versioning Use Cases (W3C)

21
Related Work XML Schemas

• XML Schema languages
• Many, but XML Schema is backed by the W3C
• Incremental XML validation
• Bouchou Halfeld-Ferrari, DBPL 2003
• Papkonstantinou Vianu, ICDT 2003
• Barbosa, Mendelzon, Libkin, Mignet, Arenas, ICDE
  2004
• Temporal XML schemas
• Currim, Currim, Dyreson, Snodgrass, EDBT 2004
• Dyreson, Snodgrass, Currim, Currim, Joshi, XSDM
  2006

22
An Overarching Vision

• Aspect-oriented programming
• Cross-cutting concerns
• Augment behavior without changing the code
• Example aspects logging, garbage collection

Program .java
Aspect .java
23
Aspects for Data?

• What are cross-cutting concerns?
• Milieu of metadata
• Time is an aspect

24
Aspects in Schema Design

• Schema for aspect schema for data
• Our paper describes the plumbing for a temporal
  aspect

data (snapshot) schema
aspect schema
schema tapestry
schema weaver
25
Our Contributions

• Temporal schema specification
• What is time-varying
• Some simple constraints
• Validate temporal data
• ?Dt-now cost
• Upwards compatible with XML Schema
• Handle schema evolution (Dyreson et al., XSDM
  06)
• Suite of tools
• Reuse and extend existing tools
• www.cs.arizona.edu/tau

26
tXSchema Project Tools (Beta)

• t VALIDATOR Validating temporal XML document
  for conventional and temporal constraints
• SQUASH Generating a temporal document from a
  sequence of snapshot documents
• UNSQUASH Extracting snapshot documents from a
  temporal document
• RESQUASH Changing a document representation to
  be consistent with the new physical annotation.