From 90d17f1bfdb11a85fcdd8f6eb7a15da42669ad9f Mon Sep 17 00:00:00 2001
From: Josip Mrden <josip.mrden@memgraph.io>
Date: Fri, 5 May 2023 15:34:25 +0200
Subject: [PATCH] Merge from match + load csv invalid behaviour

---
 src/query/plan/operator.cpp             | 22 ++++++++++-----------
 src/query/plan/pretty_print.cpp         | 26 ++++++++++++++++++++-----
 src/query/plan/rewrite/index_lookup.hpp | 10 ++++++++++
 src/utils/csv_parsing.cpp               |  2 +-
 src/utils/csv_parsing.hpp               |  2 +-
 tests/unit/utils_csv_parsing.cpp        |  2 +-
 6 files changed, 45 insertions(+), 19 deletions(-)

diff --git a/src/query/plan/operator.cpp b/src/query/plan/operator.cpp
index fb729e392..dbe3fe0f0 100644
--- a/src/query/plan/operator.cpp
+++ b/src/query/plan/operator.cpp
@@ -4598,7 +4598,7 @@ LoadCsv::LoadCsv(std::shared_ptr<LogicalOperator> input, Expression *file, bool
   MG_ASSERT(file_, "Something went wrong - '{}' member file_ shouldn't be a nullptr", __func__);
 }
 
-bool LoadCsv::Accept(HierarchicalLogicalOperatorVisitor &visitor) { return false; };
+ACCEPT_WITH_INPUT(LoadCsv)
 
 class LoadCsvCursor;
 
@@ -4649,14 +4649,12 @@ TypedValue CsvRowToTypedMap(csv::Reader::Row &row, csv::Reader::Header header) {
 class LoadCsvCursor : public Cursor {
   const LoadCsv *self_;
   const UniqueCursorPtr input_cursor_;
-  bool input_is_once_;
+  bool did_pull_;
   std::optional<csv::Reader> reader_{};
 
  public:
   LoadCsvCursor(const LoadCsv *self, utils::MemoryResource *mem)
-      : self_(self), input_cursor_(self_->input_->MakeCursor(mem)) {
-    input_is_once_ = dynamic_cast<Once *>(self_->input_.get());
-  }
+      : self_(self), input_cursor_(self_->input_->MakeCursor(mem)), did_pull_{false} {}
 
   bool Pull(Frame &frame, ExecutionContext &context) override {
     SCOPED_PROFILE_OP("LoadCsv");
@@ -4674,12 +4672,14 @@ class LoadCsvCursor : public Cursor {
 
     bool input_pulled = input_cursor_->Pull(frame, context);
 
-    // If the input is Once, we have to keep going until we read all the rows,
-    // regardless of whether the pull on Once returned false.
-    // If we have e.g. MATCH(n) LOAD CSV ... AS x SET n.name = x.name, then we
-    // have to read at most cardinality(n) rows (but we can read less and stop
-    // pulling MATCH).
-    if (!input_is_once_ && !input_pulled) return false;
+    if (input_pulled) {
+      if (did_pull_) {
+        throw QueryRuntimeException(
+            "LOAD CSV can be executed only once, please check if the cardinality of the operator before LOAD CSV is 1");
+      }
+      did_pull_ = true;
+    }
+
     auto row = reader_->GetNextRow(context.evaluation_context.memory);
     if (!row) {
       return false;
diff --git a/src/query/plan/pretty_print.cpp b/src/query/plan/pretty_print.cpp
index 1d0512d85..3b5c2303b 100644
--- a/src/query/plan/pretty_print.cpp
+++ b/src/query/plan/pretty_print.cpp
@@ -874,11 +874,27 @@ bool PlanToJsonVisitor::PreVisit(query::plan::CallProcedure &op) {
 bool PlanToJsonVisitor::PreVisit(query::plan::LoadCsv &op) {
   json self;
   self["name"] = "LoadCsv";
-  self["file"] = ToJson(op.file_);
-  self["with_header"] = op.with_header_;
-  self["ignore_bad"] = op.ignore_bad_;
-  self["delimiter"] = ToJson(op.delimiter_);
-  self["quote"] = ToJson(op.quote_);
+
+  if (op.file_) {
+    self["file"] = ToJson(op.file_);
+  }
+
+  if (op.with_header_) {
+    self["with_header"] = op.with_header_;
+  }
+
+  if (op.ignore_bad_) {
+    self["ignore_bad"] = op.ignore_bad_;
+  }
+
+  if (op.delimiter_) {
+    self["delimiter"] = ToJson(op.delimiter_);
+  }
+
+  if (op.quote_) {
+    self["quote"] = ToJson(op.quote_);
+  }
+
   self["row_variable"] = ToJson(op.row_var_);
 
   op.input_->Accept(*this);
diff --git a/src/query/plan/rewrite/index_lookup.hpp b/src/query/plan/rewrite/index_lookup.hpp
index 1bcf2cb09..feac431fe 100644
--- a/src/query/plan/rewrite/index_lookup.hpp
+++ b/src/query/plan/rewrite/index_lookup.hpp
@@ -477,6 +477,16 @@ class IndexLookupRewriter final : public HierarchicalLogicalOperatorVisitor {
     return true;
   }
 
+  bool PreVisit(LoadCsv &op) override {
+    prev_ops_.push_back(&op);
+    return true;
+  }
+
+  bool PostVisit(LoadCsv & /*op*/) override {
+    prev_ops_.pop_back();
+    return true;
+  }
+
   std::shared_ptr<LogicalOperator> new_root_;
 
  private:
diff --git a/src/utils/csv_parsing.cpp b/src/utils/csv_parsing.cpp
index 49d8a0949..4744f2100 100644
--- a/src/utils/csv_parsing.cpp
+++ b/src/utils/csv_parsing.cpp
@@ -1,4 +1,4 @@
-// Copyright 2022 Memgraph Ltd.
+// Copyright 2023 Memgraph Ltd.
 //
 // Use of this software is governed by the Business Source License
 // included in the file licenses/BSL.txt; by using this file, you agree to be bound by the terms of the Business Source
diff --git a/src/utils/csv_parsing.hpp b/src/utils/csv_parsing.hpp
index 37d438b41..928654ca8 100644
--- a/src/utils/csv_parsing.hpp
+++ b/src/utils/csv_parsing.hpp
@@ -1,4 +1,4 @@
-// Copyright 2022 Memgraph Ltd.
+// Copyright 2023 Memgraph Ltd.
 //
 // Use of this software is governed by the Business Source License
 // included in the file licenses/BSL.txt; by using this file, you agree to be bound by the terms of the Business Source
diff --git a/tests/unit/utils_csv_parsing.cpp b/tests/unit/utils_csv_parsing.cpp
index 3c852b171..9fef48af1 100644
--- a/tests/unit/utils_csv_parsing.cpp
+++ b/tests/unit/utils_csv_parsing.cpp
@@ -1,4 +1,4 @@
-// Copyright 2022 Memgraph Ltd.
+// Copyright 2023 Memgraph Ltd.
 //
 // Use of this software is governed by the Business Source License
 // included in the file licenses/BSL.txt; by using this file, you agree to be bound by the terms of the Business Source