From ab9bec0a6dc1d50dab7989372c2aa31ab284920c Mon Sep 17 00:00:00 2001
From: Spade  A <71589810+SpadeA-Tang@users.noreply.github.com>
Date: Fri, 19 Dec 2025 16:13:19 +0800
Subject: [PATCH] fix: some fixes for ngram index (#46405)

issue: https://github.com/milvus-io/milvus/issues/42053

The splitted literals in `match` execution should be handled in `and`
manner rather than `or`.

Signed-off-by: SpadeA <tangchenjie1210@gmail.com>
---
 internal/core/src/index/NgramInvertedIndex.cpp | 6 ++++--
 internal/core/src/index/NgramInvertedIndex.h   | 5 +++++
 internal/core/src/index/ScalarIndex.h          | 3 +++
 3 files changed, 12 insertions(+), 2 deletions(-)

diff --git a/internal/core/src/index/NgramInvertedIndex.cpp b/internal/core/src/index/NgramInvertedIndex.cpp
index 5bfc8007c0..6395a28c54 100644
--- a/internal/core/src/index/NgramInvertedIndex.cpp
+++ b/internal/core/src/index/NgramInvertedIndex.cpp
@@ -395,13 +395,15 @@ NgramInvertedIndex::MatchQuery(const std::string& literal,
         root_span->SetAttribute("match_query_min_gram", min_gram_);
         root_span->SetAttribute("match_query_max_gram", max_gram_);
     }
-    TargetBitmap bitset{static_cast<size_t>(Count())};
+    TargetBitmap bitset(static_cast<size_t>(Count()), true);
     auto literals = split_by_wildcard(literal);
     for (const auto& l : literals) {
         if (l.length() < min_gram_) {
             return std::nullopt;
         }
-        wrapper_->ngram_match_query(l, min_gram_, max_gram_, &bitset);
+        TargetBitmap tmp_bitset(static_cast<size_t>(Count()), false);
+        wrapper_->ngram_match_query(l, min_gram_, max_gram_, &tmp_bitset);
+        bitset &= tmp_bitset;
     }
 
     TargetBitmapView res(bitset);
diff --git a/internal/core/src/index/NgramInvertedIndex.h b/internal/core/src/index/NgramInvertedIndex.h
index 43bc828b0d..65b68e545c 100644
--- a/internal/core/src/index/NgramInvertedIndex.h
+++ b/internal/core/src/index/NgramInvertedIndex.h
@@ -49,6 +49,11 @@ class NgramInvertedIndex : public InvertedIndexTantivy<std::string> {
                  proto::plan::OpType op_type,
                  exec::SegmentExpr* segment);
 
+    ScalarIndexType
+    GetIndexType() const override {
+        return ScalarIndexType::NGRAM;
+    }
+
     void
     finish() {
         this->wrapper_->finish();
diff --git a/internal/core/src/index/ScalarIndex.h b/internal/core/src/index/ScalarIndex.h
index 404880fadf..d56bddbf29 100644
--- a/internal/core/src/index/ScalarIndex.h
+++ b/internal/core/src/index/ScalarIndex.h
@@ -38,6 +38,7 @@ enum class ScalarIndexType {
     HYBRID,
     JSONSTATS,
     RTREE,
+    NGRAM,
 };
 
 inline std::string
@@ -57,6 +58,8 @@ ToString(ScalarIndexType type) {
             return "HYBRID";
         case ScalarIndexType::RTREE:
             return "RTREE";
+        case ScalarIndexType::NGRAM:
+            return "NGRAM";
         default:
             return "UNKNOWN";
     }