Handling Bitemporal Schema Versions in Multi-temporal Environment for Data Warehouse

Abstract

Bitemporal schema versioning for DWs that maintain valid-time data or bitemporal data has already been discussed in the literature. But, the data in DWs may come from data sources having varied temporal formats, i.e., from operational sources that may or may not maintain time element or may contain only transaction-time, only valid-time or both (bitemporal time), generating multi-temporal data formats and therefore a multi-temporal environment for DW. The paper, therefore, presents a discussion on providing support for managing bitemporal schema versioning of multi-temporal data formats in DW that includes support for snapshot data and transaction data also in addition to bitemporal and valid-time data. The storage options to provide support for managing different bitemporal versions of schema and data created as a result of various revisions have not been discussed so far for bitemporal, transaction-time and snapshot data. We propose different storage solutions for maintaining various bitemporal versions of the schema with their data. The procedures for populating data in the new versions from old versions according to the storage solution adopted are also discussed for different temporal data formats. While discussing the different storage solutions, synchronous and non-synchronous mappings between data and schema versioning are also discussed. In this paper, we also propose meta-schema with bitemporal timestamps to record the changes occurring in the structure of schema for multiple temporal data formats.