Information Processing Apparatus for Multi-Row Table Data Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques fail to effectively extract values corresponding to search keys from table data when the search key is described across multiple rows, as they rely on extracting values from the same row as the search key, missing values when the key is on a subsequent row.
Innovation Solution
An information processing apparatus and method that acquires lists of rows from table data based on a column value and extracts values corresponding to a search key by identifying and combining table data across multiple pages, even when the search key is not on the same row as its value.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a method of extracting a value described on the right side of the same row as the search key is applied, then the extraction process is simple, but the value cannot be extracted when the search key is described only in a subsequent row
Solution Approach 1:
The patent transitions from a one-dimensional row-based extraction approach to a two-dimensional approach that considers both row and column positions. By introducing the concept of searching upward to the first row and then extracting from the right side, the system handles multi-row key descriptions while maintaining structured extraction logic.
Solution Approach 2:
The patent performs a preliminary search to determine whether the search key is described in the first row or a subsequent row before executing the extraction. This preliminary classification enables the system to apply different extraction strategies based on the key's position, ensuring accurate value retrieval regardless of row distribution.
2Adaptability or versatility
If the search key is described across multiple rows, then the data representation is flexible, but the conventional extraction method fails to retrieve the corresponding value
Solution Approach 1:
The patent implements a dynamic extraction approach that adapts to the structural characteristics of the table data. The extraction method changes based on whether the search key spans multiple rows, allowing the system to handle both single-row and multi-row key descriptions effectively while maintaining precise value retrieval.
Solution Approach 2:
The patent introduces an intermediary search step that identifies the first row containing the search key before performing the actual value extraction. This intermediary action serves as a bridge between the flexible multi-row key representation and the structured value extraction process, ensuring accurate retrieval regardless of key distribution.
3Productivity
If only the row containing the search key is processed, then the processing scope is limited, but values from other rows cannot be retrieved
Solution Approach 1:
The patent creates a universal extraction method that functions for both single-row and multi-row key descriptions. By making the extraction process adaptable to different key distribution patterns, the system retrieves complete value information without requiring separate processing logic for different scenarios, maintaining efficiency while preventing information loss.
Solution Approach 2:
The patent performs a preliminary analysis to determine the key's row distribution pattern before executing extraction. This preliminary step enables the system to expand its processing scope appropriately - only searching upward when needed - thus maintaining processing efficiency while ensuring complete value retrieval.
Data Source
AI summary
An information processing apparatus acquires a plurality of lists including one or more rows from table data including a plurality of columns and a plurality of rows, based on a value of one column in the plurality of columns and extracts, based on a list including a predetermined search key among the acquired plurality of lists, a value corresponding to the search key. Thus, for example, even in a case where there is no character string on the right side of the search key, value information corresponding to the search key is extractable.