From 48862c678442c38afcc961631834b0d7ad8fd8b9 Mon Sep 17 00:00:00 2001
From: PHILO-HE <feilong.he@intel.com>
Date: Fri, 17 Dec 2021 10:52:37 +0800
Subject: [PATCH] Add check for input date string (#59)

---
 cpp/src/gandiva/function_registry_datetime.cc |  4 ++++
 cpp/src/gandiva/precompiled/time.cc           | 17 +++++++++++++++++
 cpp/src/gandiva/precompiled/types.h           |  3 +++
 3 files changed, 24 insertions(+)

diff --git a/cpp/src/gandiva/function_registry_datetime.cc b/cpp/src/gandiva/function_registry_datetime.cc
index 1e003f818db2a..b222967e53d78 100644
--- a/cpp/src/gandiva/function_registry_datetime.cc
+++ b/cpp/src/gandiva/function_registry_datetime.cc
@@ -69,6 +69,10 @@ std::vector<NativeFunction> GetDateTimeFunctionRegistry() {
                      kResultNullIfNull, "castTIMESTAMP_utf8",
                      NativeFunction::kNeedsContext | NativeFunction::kCanReturnErrors),
 
+      NativeFunction("castTIMESTAMP_with_validation_check", {}, DataTypeVector{utf8()}, timestamp(),
+                     kResultNullInternal, "castTIMESTAMP_with_validation_check_utf8",
+                     NativeFunction::kNeedsContext | NativeFunction::kCanReturnErrors),
+
       NativeFunction("castTIMESTAMP_withCarrying", {}, DataTypeVector{utf8()}, timestamp(),
                      kResultNullInternal, "castTIMESTAMP_withCarrying_utf8",
                      NativeFunction::kNeedsContext | NativeFunction::kCanReturnErrors),
diff --git a/cpp/src/gandiva/precompiled/time.cc b/cpp/src/gandiva/precompiled/time.cc
index 48601048ad82b..525b60178428c 100644
--- a/cpp/src/gandiva/precompiled/time.cc
+++ b/cpp/src/gandiva/precompiled/time.cc
@@ -871,6 +871,23 @@ gdv_timestamp castTIMESTAMP_utf8(int64_t context, const char* input, gdv_int32 l
   return std::chrono::time_point_cast<milliseconds>(date_time).time_since_epoch().count();
 }
 
+gdv_timestamp castTIMESTAMP_with_validation_check_utf8(int64_t context, const char* input,
+                                                       gdv_int32 length, bool in_valid,
+                                                       bool* out_valid) {
+  if (!in_valid) {
+    *out_valid = false;
+    return 0;
+  }
+  // Suppose the input is yyyy-MM-dd.
+  // TODO: check validaition during/after parsing.
+  if (length > 10) {
+    *out_valid = false;
+    return 0;
+  }
+  return castTIMESTAMP_withCarrying_utf8(context, input, length, in_valid, out_valid);
+}
+
+
 /*
  * Input consists of mandatory and optional fields.
  * Mandatory fields are year, month and day.
diff --git a/cpp/src/gandiva/precompiled/types.h b/cpp/src/gandiva/precompiled/types.h
index 2e360d3841642..208f71476196a 100644
--- a/cpp/src/gandiva/precompiled/types.h
+++ b/cpp/src/gandiva/precompiled/types.h
@@ -425,6 +425,9 @@ gdv_timestamp castTIMESTAMP_utf8(int64_t execution_context, const char* input,
 gdv_timestamp castTIMESTAMP_withCarrying_utf8(int64_t context, const char* input,
                                               gdv_int32 length, bool in_valid,
                                               bool* out_valid);
+gdv_timestamp castTIMESTAMP_with_validation_check_utf8(int64_t context, const char* input,
+                                              gdv_int32 length, bool in_valid,
+                                              bool* out_valid);
 gdv_timestamp castTIMESTAMP_date64(gdv_date64);
 gdv_timestamp castTIMESTAMP_int64(gdv_int64);
 gdv_date64 castDATE_timestamp(gdv_timestamp);