Add tests for wildcard + regexp match-all cases

Signed-off-by: Michael Froh <[email protected]>
opensearch-project · Jun 11, 2024 · aa62135 · aa62135
1 parent 3ee80ee
commit aa62135
Show file tree

Hide file tree

Showing 3 changed files with 88 additions and 13 deletions.
diff --git a/...-api-spec/src/main/resources/rest-api-spec/test/search/270_wildcard_fieldtype_queries.yml b/...-api-spec/src/main/resources/rest-api-spec/test/search/270_wildcard_fieldtype_queries.yml
@@ -50,6 +50,12 @@ setup:
         id: 5
         body:
           my_field: "AbCd"
+  - do:
+      index:
+        index: test
+        id: 6
+        body:
+          other_field: "test"
   - do:
       indices.refresh: {}
 
@@ -198,3 +204,26 @@ setup:
               my_field:
                 value: ".*06-08.*Cluster-Manager Node.*"
   - match: { hits.total.value: 0 }
+
+---
+"wildcard match-all works":
+  - do:
+      search:
+        index: test
+        body:
+          query:
+            wildcard:
+              my_field:
+                value: "*"
+  - match: { hits.total.value: 5 }
+---
+"regexp match-all works":
+  - do:
+      search:
+        index: test
+        body:
+          query:
+            regexp:
+              my_field:
+                value: ".*"
+  - match: { hits.total.value: 5 }
diff --git a/server/src/main/java/org/opensearch/index/mapper/WildcardFieldMapper.java b/server/src/main/java/org/opensearch/index/mapper/WildcardFieldMapper.java
@@ -23,6 +23,7 @@
 import org.apache.lucene.search.DocIdSetIterator;
 import org.apache.lucene.search.IndexSearcher;
 import org.apache.lucene.search.MatchAllDocsQuery;
+import org.apache.lucene.search.MatchNoDocsQuery;
 import org.apache.lucene.search.MultiTermQuery;
 import org.apache.lucene.search.Query;
 import org.apache.lucene.search.QueryVisitor;
@@ -448,14 +449,20 @@ public Query wildcardQuery(String value, MultiTermQuery.RewriteMethod method, bo
                 };
             }
 
-            return new WildcardMatchingQuery(
-                name(),
-                matchAllTermsQuery(name(), getRequiredNGrams(finalValue)),
-                matchPredicate,
-                value,
-                context,
-                this
-            );
+            Set<String> requiredNGrams = getRequiredNGrams(finalValue);
+            Query approximation;
+            if (requiredNGrams.isEmpty()) {
+                // This only happens when all characters are wildcard characters (* or ?),
+                // or it's the empty string.
+                if (value.length() == 0 || value.contains("?")) {
+                    approximation = this.existsQuery(context);
+                } else {
+                    return existsQuery(context);
+                }
+            } else {
+                approximation = matchAllTermsQuery(name(), requiredNGrams);
+            }
+            return new WildcardMatchingQuery(name(), approximation, matchPredicate, value, context, this);
         }
 
         // Package-private for testing
@@ -540,10 +547,23 @@ public Query regexpQuery(
             Automaton automaton = regExp.toAutomaton(maxDeterminizedStates);
             CompiledAutomaton compiledAutomaton = new CompiledAutomaton(automaton);
 
-            return new WildcardMatchingQuery(name(), regexpToQuery(name(), regExp), s -> {
-                BytesRef valueBytes = BytesRefs.toBytesRef(s);
-                return compiledAutomaton.runAutomaton.run(valueBytes.bytes, valueBytes.offset, valueBytes.length);
-            }, "/" + value + "/", context, this);
+            Predicate<String> regexpPredicate;
+            if (compiledAutomaton.type == CompiledAutomaton.AUTOMATON_TYPE.ALL) {
+                return existsQuery(context);
+            } else if (compiledAutomaton.type == CompiledAutomaton.AUTOMATON_TYPE.NONE) {
+                return new MatchNoDocsQuery("Regular expression matches nothing");
+            } else {
+                regexpPredicate = s -> {
+                    BytesRef valueBytes = BytesRefs.toBytesRef(s);
+                    return compiledAutomaton.runAutomaton.run(valueBytes.bytes, valueBytes.offset, valueBytes.length);
+                };
+            }
+
+            Query approximation = regexpToQuery(name(), regExp);
+            if (approximation instanceof MatchAllDocsQuery) {
+                approximation = existsQuery(context);
+            }
+            return new WildcardMatchingQuery(name(), approximation, regexpPredicate, "/" + value + "/", context, this);
         }
 
         /**
@@ -602,6 +622,8 @@ private static Query regexpToQuery(String fieldName, RegExp regExp) {
                 }
             if (query.clauses().size() == 1) {
                 return query.iterator().next().getQuery();
+            } else if (query.clauses().size() == 0) {
+                return new MatchAllDocsQuery();
             }
             return query;
         }
@@ -704,7 +726,7 @@ private WildcardMatchingQuery(
 
         @Override
         public String toString(String s) {
-            return "WildcardMatchingQuery(" + fieldName + "\"" + patternString + "\")";
+            return "WildcardMatchingQuery(" + fieldName + ":\"" + patternString + "\")";
         }
 
         @Override

diff --git a/server/src/test/java/org/opensearch/index/mapper/WildcardFieldTypeTests.java b/server/src/test/java/org/opensearch/index/mapper/WildcardFieldTypeTests.java
@@ -149,4 +149,28 @@ public void testRegexpQuery() {
         assertTrue(actualMatchingQuery.getSecondPhaseMatcher().test("abcdjk"));
         assertTrue(actualMatchingQuery.getSecondPhaseMatcher().test("abefqwertyhi"));
     }
+
+    public void testWildcardMatchAll() {
+        String pattern = "???";
+        MappedFieldType ft = new WildcardFieldMapper.WildcardFieldType("field");
+        Query actual = ft.wildcardQuery(pattern, null, null);
+        assertEquals(new WildcardFieldMapper.WildcardMatchingQuery("field", ft.existsQuery(null), "???"), actual);
+
+        pattern = "*";
+        actual = ft.wildcardQuery(pattern, null, null);
+        assertEquals(ft.existsQuery(null), actual);
+    }
+
+    public void testRegexpMatchAll() {
+        // The following matches any string of length exactly 3. We do need to evaluate the predicate.
+        String pattern = "...";
+        MappedFieldType ft = new WildcardFieldMapper.WildcardFieldType("field");
+        Query actual = ft.regexpQuery(pattern, 0, 0, 1000, null, null);
+        assertEquals(new WildcardFieldMapper.WildcardMatchingQuery("field", ft.existsQuery(null), "/.../"), actual);
+
+        // The following pattern has a predicate that matches everything. We can just return the field exists query.
+        pattern = ".*";
+        actual = ft.regexpQuery(pattern, 0, 0, 1000, null, null);
+        assertEquals(ft.existsQuery(null), actual);
+    }
 }