From a9e6e71267d0569563d4d3acb8429a13082f9610 Mon Sep 17 00:00:00 2001
From: mwish <maplewish117@gmail.com>
Date: Tue, 7 May 2024 00:15:31 +0800
Subject: [PATCH] Add test inside ArrowBinaryDirectPut

---
 cpp/src/parquet/encoding_test.cc | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/cpp/src/parquet/encoding_test.cc b/cpp/src/parquet/encoding_test.cc
index b91fcb0839cba..cc76689abb53c 100644
--- a/cpp/src/parquet/encoding_test.cc
+++ b/cpp/src/parquet/encoding_test.cc
@@ -577,6 +577,11 @@ TEST(PlainEncodingAdHoc, ArrowBinaryDirectPut) {
     auto decoder = MakeTypedDecoder<ByteArrayType>(Encoding::PLAIN);
 
     ASSERT_NO_THROW(encoder->Put(*values));
+    // For Plain encoding, the estimated size should be at least the total byte size
+    auto& string_array = dynamic_cast<const ::arrow::StringArray&>(*values);
+    EXPECT_GE(encoder->EstimatedDataEncodedSize(), string_array.total_values_length())
+        << "Estimated size should be at least the total byte size";
+
     auto buf = encoder->FlushValues();
 
     int num_values = static_cast<int>(values->length() - values->null_count());
@@ -2160,6 +2165,10 @@ TEST(DeltaLengthByteArrayEncodingAdHoc, ArrowBinaryDirectPut) {
 
   auto CheckSeed = [&](std::shared_ptr<::arrow::Array> values) {
     ASSERT_NO_THROW(encoder->Put(*values));
+    auto& string_array = checked_cast<const ::arrow::StringArray&>(*values);
+    // For DeltaLength encoding, the estimated size should be at least the total byte size
+    EXPECT_GE(encoder->EstimatedDataEncodedSize(), string_array.total_values_length())
+        << "Estimated size should be at least the total byte size";
     auto buf = encoder->FlushValues();
 
     int num_values = static_cast<int>(values->length() - values->null_count());