Transaction-Based Workspace Refresh for Long Database Transactions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems for managing long database transactions face inefficiencies when refreshing a workspace view, as they often require re-baselining all changes made to the entire dataset, which can take a prohibitively long time, especially in systems with millions or hundreds of millions of records.
Innovation Solution
A method involving the creation of a data structure that logs changes made to the database during a long transaction, allowing for recursive identification and refresh of only those transactions directly or indirectly related to the initial transaction, thereby avoiding the need for a full re-baseline and speeding up the refresh process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing solutions re-baseline the workspace view to apply all changes in the parent workspace, then the workspace is fully refreshed with all changes, but the process takes a prohibitively long time for very large datasets
Solution Approach 1:
The patent segments the workspace refresh process by dividing it into two approaches: a traditional full re-baseline approach and a new incremental refresh approach. The incremental approach segments the refresh operation to only include changes from specific transactions identified through a data structure, rather than reapplying all changes to the entire dataset. This segmentation allows selective refreshing of only the necessary portions of the workspace.
Solution Approach 2:
The patent implements partial action by performing only the necessary subset of refresh operations. Instead of re-baselining the entire workspace (excessive action), the system identifies and applies only the changes from transactions that occurred after the original transaction was opened and are related to the records being committed. This partial refresh achieves the required completeness without the excessive time cost of a full re-baseline.
2Productivity
If a data structure is created to log all further transactions' changes, then the recursive identification of related transactions becomes possible, but the device complexity increases
Solution Approach 1:
The patent applies preliminary action by creating and maintaining a data structure that logs all transactions and their associated changes as they occur in the parent workspace. This data structure is built incrementally as transactions are committed, storing information about which transactions modified which records. When a refresh is needed, this pre-built data structure enables rapid identification of related transactions without requiring complex analysis at refresh time.
Solution Approach 2:
The patent introduces a data structure as an intermediary between the transaction logging system and the workspace refresh mechanism. This intermediary structure stores the mapping between transactions and the records they modified, enabling efficient recursive identification of related transactions. The data structure acts as a mediator that simplifies the refresh process by providing direct access to transaction-related information without requiring complex queries or analysis of the underlying database.
Data Source
AI summary
Aspects of the present disclosure aim to improve upon systems for managing long database transactions, and the way in which changes made to a database from within a workspace view during a long transaction are committed to the underlying parent database. The present disclosure aims to solve problems caused by the need to refresh a workspace view as part of the process of resolving conflicts between competing workspace views operating on the same portions of the database. Aspects described herein solve the problem through the creation of a separate data structure which effectively logs all of the further transactions which commit changes to parent database. The data structure can be arranged to include all changes committed after a first transaction is opened and worked upon. The process of committing the changes made in the first transaction to the workspace includes performing a recursive search using the data structure.