Comparing Very Large Database Snapshots
نویسندگان
چکیده
Detecting and extracting modi cations from information sources is an integral part of data warehousing. For unsophisticated sources, in practice it is often necessary to infer modi cations by periodically comparing snapshots of data from the source. We call this problem the snapshot di erential problem. We show that this is closely related to outerjoins. In this paper we extend the traditional join algorithms to perform outerjoins. We then make the outerjoin algorithms more e cient by using compression techniques. We also examine how text comparison algorithms can be used to solve the snapshot di erential problem.
منابع مشابه
Flow Visualization by Conditional Sampling of a Single X-Wire Probe in a Very Long Run Experiment
Flow visualization techniques using tracer markers such as die, smoke, hydrogen bubbles, etc., have been widely used in experimental investigations of large scale structures of a variety of flow fields. They have played an important role in understanding the physics of the coherent structures formation and evolution in the transitional as well as the turbulent regions of the flow fields. Howeve...
متن کاملFlow Visualization by Conditional Sampling of a Single X-Wire Probe in a Very Long Run Experiment
Flow visualization techniques using tracer markers such as die, smoke, hydrogen bubbles, etc., have been widely used in experimental investigations of large scale structures of a variety of flow fields. They have played an important role in understanding the physics of the coherent structures' formation and evolution in the transitional as well as the turbulent regions of the flow fields. Howev...
متن کاملPerformance Analysis of Dynamic Finite Versioning Schemes: Storage Cost vs. Obsolescence
Dynamic finite versioning (DFV) schemes are an effective approach to concurrent transaction and query processing, where a finite number of consistent, but maybe slightly out-of-date, logicalsnapshots of the database can be dynamically derived for query access. In DFV, the storage overhead for keeping additional versions of changed data to support the logical snapshots and the amount of obsolesc...
متن کاملPacket Tracking for Post-Silicon Diagnosis and Debug of Networks-on-Chip
Networks-on-chips (NoCs) is the main trend and solution to provide higher communication bandwidth and more efficient transaction mechanism for larger chip multiprocessors (CMPs). Since the complexity and size of the chip is growing faster, postsilicon verification becomes a big problem. In this paper, we propose an advanced solution called packet tracking to help debugging and improve observabi...
متن کاملEecient Snapshot Diierential Algorithms for Data Warehousing
Detecting and extracting modi cations from information sources is an integral part of data warehousing. For unsophisticated sources, it is often necessary to infer modi cations by periodically comparing snapshots of data from the source. Although this snapshot di erential problem is closely related to traditional joins, there are significant di erences, which lead to simple new algorithms. In p...
متن کامل