normalize

Generic Description

Normalizes Unicode text to a requested normalization form. It is documented separately because text normalization and slicing often become grouping keys, search inputs, or exported identifiers.

Simple example:

RETURN normalize('Café', 'NFC') AS value

Consumer-Level Explanation

Use it when visually similar strings must compare consistently across imported sources. Prefer this function when the text operation is part of the query contract and should be visible during review.

More Detailed Explanation

normalize runs inside the row pipeline, which makes text cleanup, extraction, and grouping keys visible to Nexyron instead of hidden in application code. That is especially important for imported document metadata, labels, and identifiers that feed later matching or indexing.

Advanced Example

This example applies normalize while shaping document metadata, so text cleanup is part of the auditable query rather than an application-side afterthought.

WITH ['Café', 'Café'] AS titles
UNWIND titles AS raw_title
RETURN raw_title, normalize(raw_title, 'NFC') AS normalized_title
ORDER BY normalized_title

Real Use Cases

normalizing document titles and source identifiers before grouping
extracting prefixes, suffixes, or tokens from imported text fields
checking payload size or Unicode normalization before ingestion cleanup

Real Limitations And Tradeoffs

query-time cleanup is explicit but can hide upstream data-quality debt if overused
string operations are not semantic search; use vector or semantic procedures for meaning-based matching
byte length and character length answer different questions and should not be substituted blindly