This use case demonstrates how M|Box's Table Indexer and Table Index Matcher nodes can be combined to efficiently index and search multi-field address data, even with faulty or incomplete inputs.
Workflow Overview:
The workflow loads
A reference table of full addresses, divided into three address fields.
A search table including misspelled search terms for these fields
The Table Indexer constructs a fast, fault-tolerant search index from all three columns of the reference table.
The Table Index Matcher lets you select a column to search the index and quickly, reliably and fault-tolerantly finds a match. The node also gives you the quality of the match and the number of the matched row in the reference table.
The search result is joined with the original table using using the row number to get the full, correct address.
Index-based matching with the M|Box Table Indexing and Matching nodes ensures robust and scalable address search, even in the presence of incorrect, fuzzy inputs or spelling errors. This makes it ideal for:
Matching user-entered or third-party address data against a standardized reference set
Searching addresses with incomplete information (e.g. only a street name)
Improving address search and matching quality without sacrificing performance or explainability