为何Lucene默认不区分字符串与数值范围查询均按字符串处理？

Great question—this design choice cuts to the core of Lucene's foundational philosophy. Let’s break down why queries like ['2' TO '6'] and [2 TO 6] both default to string-based range checks:

• Lucene’s Core: All Data is a String Term
  At its heart, Lucene is built on an inverted index where every piece of indexed data is stored as a string-based term. Whether you index a number like 2 or a string like "2", they become identical terms in the index (unless you use a custom analyzer to modify them). When running a range query, the default parser compares these terms lexicographically (string dictionary order), not numerically. That’s why [2 TO 6] won’t match 10—lex order puts "10" before "2" because the first character "1" is smaller than "2".

• Backward Compatibility & Legacy Roots
  Lucene started as a text search engine, where strict typing wasn’t a primary concern. Early versions had no native numeric field support—developers simply stored numbers as strings. When numeric-specific features (like IntPoint or NumericRangeQuery) were added later, the team kept the default query behavior unchanged to avoid breaking existing code. This ensured legacy apps relying on string-based range checks for numbers continued working as expected.

• Flexibility Over Forced Typing
  Lucene doesn’t enforce strict typing by design—it puts control in your hands. If you need numeric range queries, you opt in explicitly:

  1. Index numeric fields using dedicated types like IntPoint, LongPoint, or DoublePoint (these use optimized encoding for fast range lookups).
  2. Use numeric-specific queries (like NumericRangeQuery) or configure your query parser to recognize numeric fields and handle them appropriately.
     The default "string-first" behavior keeps the core system simple, letting you add complexity only when you need it.

• Avoiding Ambiguity in Queries
  Automatic type detection would introduce unpredictable behavior. For example, a field might contain both numeric strings ("2") and text strings ("two")—Lucene couldn’t guess which type you intended for a range query like [2 TO 6]. By treating everything as a string by default, it removes ambiguity and makes query results predictable.

In short, Lucene’s default approach comes from its text-search origins, a focus on backward compatibility, and a philosophy of keeping the core system flexible and unopinionated. Numeric range behavior is available—you just need to explicitly configure your fields and queries to use it.

内容的提问来源于stack exchange，提问作者alexanoid