slides - UCLA Computer Science

Download Report

Transcript slides - UCLA Computer Science

CS240A: Databases and Knowledge Bases
Temporal Databases
Carlo Zaniolo
Department of Computer Science
University of California, Los Angeles
The Problem is Harder than you think
No support for temporal queries in SQL,
although information in real life tend to be time-stamped and
my applications re temporal in nature
Temporal Events (date of accident), vs.
 and temporal states (the state of being an employee in a company—
duration, besides start event and end event).
Many research approaches proposed but no commercial
Extending SQL (and SQL-2 in particular) is an Herculean task
Temporal DBs come in three main types: transaction time, valid
time, and bi-temporal. Each require a specialized treatment.
These issues pose challenges at the logical level (data
representation and queries)
but significant challenges also occur at the physical level:
efficient storage, indexing techniques and query optimization
techniques are needed.
Time in History
As today, there is a poor understanding of
temporal issues and their complexities among
database specialists
This parallels the general ignorance about time
we find in modern life, where the solutions of
all temporal problems are taken for granted
But finding efficient ways to represent and
measure time have created major technical and
social challenge to civilizations for thousands of
Time Changes Everything
Temporal extensions based on TSQL2 included
in the SQL standards in 2011
Two DBMS vendors (IBM and Teradata) have
recently announced support for bi-temporal DBs
in their databases
 Transaction
 Valid
Time is called System Time
time is called Business Time
Storage and simple retrieval of bi-temporal
queries is supported, but complex temporal
queries are not supported in DB2.