Wildcard field use only 3-gram to index (opensearch-project#17349)

* support 3gram wildcard Signed-off-by: gesong.samuel <[email protected]> * add changelog-3 Signed-off-by: gesong.samuel <[email protected]> * add rolling upgrade test for wildcard field Signed-off-by: gesong.samuel <[email protected]> * remove test case added in opensearch-project#16827 Signed-off-by: gesong.samuel <[email protected]> --------- Signed-off-by: gesong.samuel <[email protected]> Co-authored-by: gesong.samuel <[email protected]>
jainankitk · Feb 18, 2025 · e62bf1a · e62bf1a
1 parent 91a93da
commit e62bf1a
Show file tree

Hide file tree

Showing 7 changed files with 715 additions and 130 deletions.
diff --git a/CHANGELOG-3.0.md b/CHANGELOG-3.0.md
@@ -37,6 +37,7 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 - Stop minimizing automata used for case-insensitive matches ([#17268](https://github.com/opensearch-project/OpenSearch/pull/17268))
 - Refactor the `:server` module `org.opensearch.client` to `org.opensearch.transport.client` to eliminate top level split packages for JPMS support ([#17272](https://github.com/opensearch-project/OpenSearch/pull/17272))
 - Use Lucene `BM25Similarity` as default since the `LegacyBM25Similarity` is marked as deprecated ([#17306](https://github.com/opensearch-project/OpenSearch/pull/17306))
+- Wildcard field index only 3gram of the input data [#17349](https://github.com/opensearch-project/OpenSearch/pull/17349)
 
 ### Deprecated
 

diff --git a/qa/rolling-upgrade/src/test/resources/rest-api-spec/test/mixed_cluster/40_wildcard.yml b/qa/rolling-upgrade/src/test/resources/rest-api-spec/test/mixed_cluster/40_wildcard.yml
@@ -0,0 +1,200 @@
+# refactored from rest-api-spec/src/main/resources/rest-api-spec/test/search/270_wildcard_fieldtype_queries.yml
+---
+"search on mixed state":
+  # "term query matches exact value"
+  - do:
+      search:
+        index: test
+        body:
+          query:
+            term:
+              my_field: "AbCd"
+  - match: { hits.total.value: 1 }
+  - match: { hits.hits.0._id: "5" }
+
+  - do:
+      search:
+        index: test
+        body:
+          query:
+            term:
+              my_field.doc_values: "AbCd"
+  - match: { hits.total.value: 1 }
+  - match: { hits.hits.0._id: "5" }
+
+  # term query matches lowercase-normalized value
+  - do:
+      search:
+        index: test
+        body:
+          query:
+            term:
+              my_field.lower: "abcd"
+  - match: { hits.total.value: 2 }
+  - match: { hits.hits.0._id: "5" }
+  - match: { hits.hits.1._id: "7" }
+
+  - do:
+      search:
+        index: test
+        body:
+          query:
+            term:
+              my_field.lower: "ABCD"
+  - match: { hits.total.value: 2 }
+  - match: { hits.hits.0._id: "5" }
+  - match: { hits.hits.1._id: "7" }
+
+  - do:
+      search:
+        index: test
+        body:
+          query:
+            term:
+              my_field: "abcd"
+  - match: { hits.total.value: 0 }
+
+  # wildcard query matches
+  - do:
+      search:
+        index: test
+        body:
+          query:
+            wildcard:
+              my_field:
+                value: "*Node*Exception*"
+  - match: { hits.total.value: 1 }
+  - match: { hits.hits.0._id: "1" }
+
+  # wildcard query matches lowercase-normalized field
+  - do:
+      search:
+        index: test
+        body:
+          query:
+            wildcard:
+              my_field.lower:
+                value: "*node*exception*"
+  - match: { hits.total.value: 1 }
+  - match: { hits.hits.0._id: "1" }
+
+  - do:
+      search:
+        index: test
+        body:
+          query:
+            wildcard:
+              my_field.lower:
+                value: "*NODE*EXCEPTION*"
+  - match: { hits.total.value: 1 }
+  - match: { hits.hits.0._id: "1" }
+
+  - do:
+      search:
+        index: test
+        body:
+          query:
+            wildcard:
+              my_field:
+                value: "*node*exception*"
+  - match: { hits.total.value: 0 }
+
+  # prefix query matches
+  - do:
+      search:
+        index: test
+        body:
+          query:
+            prefix:
+              my_field:
+                value: "[2024-06-08T"
+  - match: { hits.total.value: 3 }
+
+  # regexp query matches
+  - do:
+      search:
+        index: test
+        body:
+          query:
+            regexp:
+              my_field:
+                value: ".*06-08.*cluster-manager node.*"
+  - match: { hits.total.value: 2 }
+
+  # regexp query matches lowercase-normalized field
+  - do:
+      search:
+        index: test
+        body:
+          query:
+            regexp:
+              my_field.lower:
+                value: ".*06-08.*Cluster-Manager Node.*"
+  - match: { hits.total.value: 2 }
+
+  - do:
+      search:
+        index: test
+        body:
+          query:
+            regexp:
+              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: 6 }
+
+  # regexp match-all works
+  - do:
+      search:
+        index: test
+        body:
+          query:
+            regexp:
+              my_field:
+                value: ".*"
+  - match: { hits.total.value: 6 }
+
+  # terms query on wildcard field matches
+  - do:
+      search:
+        index: test
+        body:
+          query:
+            terms: { my_field: [ "AbCd" ] }
+  - match: { hits.total.value: 1 }
+  - match: { hits.hits.0._id: "5" }
+
+  # case insensitive query on wildcard field
+  - do:
+      search:
+        index: test
+        body:
+          query:
+            wildcard:
+              my_field:
+                value: "AbCd"
+  - match: { hits.total.value: 1 }
+  - match: { hits.hits.0._id: "5" }
+
+  - do:
+      search:
+        index: test
+        body:
+          query:
+            wildcard:
+              my_field:
+                value: "AbCd"
+                case_insensitive: true
+  - match: { hits.total.value: 2 }
+  - match: { hits.hits.0._id: "5" }
+  - match: { hits.hits.1._id: "7" }