fix(formula): break tokenizer identifiers when current word is a keyword

The tokenizer already broke multi-word identifiers when the NEXT word
was a keyword, but not when the identifier collected SO FAR was a
keyword. This meant "WHERE Region" was merged into one token when
tokenizing "SUM(Revenue WHERE Region = East)".

Now the tokenizer also checks if the identifier built up to the current
space IS a keyword (WHERE, SUM, AVG, MIN, MAX, COUNT, IF), which
correctly produces separate tokens for "Revenue", "WHERE", "Region".

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

src/formula/parser.rs

@@ -212,6 +212,14 @@ fn tokenize(s: &str) -> Result<Vec<Token>> {
                         ) {
                             break;
                         }
+                        // Break if the identifier collected so far is a keyword
+                        let trimmed = ident.trim_end().to_ascii_uppercase();
+                        if matches!(
+                            trimmed.as_str(),
+                            "WHERE" | "SUM" | "AVG" | "MIN" | "MAX" | "COUNT" | "IF"
+                        ) {
+                            break;
+                        }
                         // Also break if the next word is a keyword
                         let rest: String = chars[j..].iter().collect();
                         let next_word: String = rest
@@ -500,15 +508,6 @@ mod tests {
 
     // ── Aggregate with WHERE filter ─────────────────────────────────────
 
-    /// NOTE: WHERE inside aggregate parens is broken when the inner expression
-    /// is a bare identifier. The tokenizer treats "Revenue WHERE" as a single
-    /// multi-word identifier because it greedily consumes spaces followed by
-    /// non-operator characters. The WHERE-inside-aggregate syntax only works
-    /// if the inner expression is a number, parenthesized, or otherwise
-    /// terminated before the WHERE keyword.
-    ///
-    /// Top-level WHERE (outside parens) works fine because split_where handles
-    /// it before tokenization.
     #[test]
     fn parse_sum_with_top_level_where_works() {
         let f = parse_formula(
@@ -667,6 +666,14 @@ mod tests {
         assert!(parse_expr("").is_err());
     }
 
+    #[test]
+    fn tokenizer_breaks_at_where_keyword() {
+        use super::tokenize;
+        let tokens = tokenize("Revenue WHERE Region").unwrap();
+        // Should produce 3 tokens: Ident("Revenue"), Ident("WHERE"), Ident("Region")
+        assert_eq!(tokens.len(), 3, "Expected 3 tokens, got: {tokens:?}");
+    }
+
     // ── Multi-word identifiers ──────────────────────────────────────────
 
     #[test]