From d796391464cc6e044c91cff1f9209220364370f6 Mon Sep 17 00:00:00 2001
From: Harpreet Sangar <happy_san@protonmail.com>
Date: Tue, 11 Apr 2023 17:44:32 +0530
Subject: [PATCH] Add `approx_filter_ids_length` field.

---
 include/filter_result_iterator.h | 10 ++++++++--
 src/filter_result_iterator.cpp   |  6 ++++--
 src/index.cpp                    |  3 ++-
 src/or_iterator.cpp              |  2 +-
 4 files changed, 15 insertions(+), 6 deletions(-)

diff --git a/include/filter_result_iterator.h b/include/filter_result_iterator.h
index b67d67ca..f5ae22a6 100644
--- a/include/filter_result_iterator.h
+++ b/include/filter_result_iterator.h
@@ -125,12 +125,18 @@ private:
 
 public:
     uint32_t seq_id = 0;
-    // Collection name -> references
+    /// Collection name -> references
     std::map<std::string, reference_filter_result_t> reference;
     Option<bool> status = Option(true);
 
+    /// Holds the upper-bound of the number of seq ids this iterator would match.
+    /// Useful in a scenario where we need to differentiate between filter iterator not matching any document v/s filter
+    /// iterator reaching it's end. (is_valid would be false in both these cases)
+    uint32_t approx_filter_ids_length;
+
     explicit filter_result_iterator_t(const std::string collection_name,
-                                      Index const* const index, filter_node_t const* const filter_node);
+                                      Index const* const index, filter_node_t const* const filter_node,
+                                      uint32_t approx_filter_ids_length = UINT32_MAX);
 
     ~filter_result_iterator_t();
 
diff --git a/src/filter_result_iterator.cpp b/src/filter_result_iterator.cpp
index 21a4128f..f5bd8b5f 100644
--- a/src/filter_result_iterator.cpp
+++ b/src/filter_result_iterator.cpp
@@ -835,10 +835,12 @@ uint32_t filter_result_iterator_t::and_scalar(const uint32_t* A, const uint32_t&
 }
 
 filter_result_iterator_t::filter_result_iterator_t(const std::string collection_name, const Index *const index,
-                                                   const filter_node_t *const filter_node)  :
+                                                   const filter_node_t *const filter_node,
+                                                   uint32_t approx_filter_ids_length)  :
         collection_name(collection_name),
         index(index),
-        filter_node(filter_node) {
+        filter_node(filter_node),
+        approx_filter_ids_length(approx_filter_ids_length) {
     if (filter_node == nullptr) {
         is_valid = false;
         return;
diff --git a/src/index.cpp b/src/index.cpp
index 29938911..ab8f8576 100644
--- a/src/index.cpp
+++ b/src/index.cpp
@@ -2713,7 +2713,8 @@ Option<bool> Index::search(std::vector<query_tokens_t>& field_query_tokens, cons
         return rearrange_op;
     }
 
-    auto filter_result_iterator = filter_result_iterator_t(collection_name, this, filter_tree_root);
+    auto filter_result_iterator = filter_result_iterator_t(collection_name, this, filter_tree_root,
+                                                           approx_filter_ids_length);
     auto filter_init_op = filter_result_iterator.init_status();
     if (!filter_init_op.ok()) {
         return filter_init_op;
diff --git a/src/or_iterator.cpp b/src/or_iterator.cpp
index df4404cb..8dd9d487 100644
--- a/src/or_iterator.cpp
+++ b/src/or_iterator.cpp
@@ -208,7 +208,7 @@ bool or_iterator_t::take_id(result_iter_state_t& istate, uint32_t id, bool& is_e
         return false;
     }
 
-    if (istate.fit != nullptr) {
+    if (istate.fit != nullptr && istate.fit->approx_filter_ids_length > 0) {
         return (istate.fit->valid(id) == 1);
     }