From a03caef6e30376243638f680a3103a1c2f0c5c73 Mon Sep 17 00:00:00 2001
From: Gang Wu <ustcwg@gmail.com>
Date: Mon, 24 Jun 2024 23:34:17 +0800
Subject: [PATCH] fix tests

---
 cpp/src/parquet/page_index.cc           | 4 ++--
 cpp/src/parquet/size_statistics.cc      | 2 +-
 cpp/src/parquet/size_statistics_test.cc | 9 +++++----
 3 files changed, 8 insertions(+), 7 deletions(-)

diff --git a/cpp/src/parquet/page_index.cc b/cpp/src/parquet/page_index.cc
index b9959f8b4aaf2..3175bfe8f6f35 100644
--- a/cpp/src/parquet/page_index.cc
+++ b/cpp/src/parquet/page_index.cc
@@ -519,7 +519,7 @@ class ColumnIndexBuilderImpl final : public ColumnIndexBuilder {
       column_index_.null_counts.clear();
     }
 
-    if (size_stats) {
+    if (size_stats != nullptr) {
       const auto& page_ref_level_hist = size_stats->repetition_level_histogram();
       const auto& page_def_level_hist = size_stats->definition_level_histogram();
       column_index_.repetition_level_histograms.insert(
@@ -696,7 +696,7 @@ class OffsetIndexBuilderImpl final : public OffsetIndexBuilder {
         if (offset_index_.page_locations.size() ==
             offset_index_.unencoded_byte_array_data_bytes.size()) {
           offset_index_.__isset.unencoded_byte_array_data_bytes = true;
-        } else {
+        } else if (!offset_index_.unencoded_byte_array_data_bytes.empty()) {
           std::stringstream ss;
           ss << "Invalid count of unencoded BYTE_ARRAY data bytes: "
              << offset_index_.unencoded_byte_array_data_bytes.size()
diff --git a/cpp/src/parquet/size_statistics.cc b/cpp/src/parquet/size_statistics.cc
index 819b45a57616f..45273d123fdff 100644
--- a/cpp/src/parquet/size_statistics.cc
+++ b/cpp/src/parquet/size_statistics.cc
@@ -113,7 +113,7 @@ std::unique_ptr<SizeStatistics> SizeStatistics::Make(const void* size_statistics
 
 class SizeStatisticsBuilder::SizeStatisticsBuilderImpl {
  public:
-  SizeStatisticsBuilderImpl(const ColumnDescriptor* descr)
+  explicit SizeStatisticsBuilderImpl(const ColumnDescriptor* descr)
       : rep_level_histogram_(descr->max_repetition_level() + 1, 0),
         def_level_histogram_(descr->max_definition_level() + 1, 0) {
     if (descr->physical_type() == Type::BYTE_ARRAY) {
diff --git a/cpp/src/parquet/size_statistics_test.cc b/cpp/src/parquet/size_statistics_test.cc
index 3184a92ecd482..b8073b65cba24 100644
--- a/cpp/src/parquet/size_statistics_test.cc
+++ b/cpp/src/parquet/size_statistics_test.cc
@@ -23,6 +23,7 @@
 #include <gtest/gtest.h>
 
 #include "arrow/buffer.h"
+#include "arrow/testing/builder.h"
 #include "arrow/testing/gtest_util.h"
 #include "arrow/util/bit_util.h"
 #include "parquet/schema.h"
@@ -32,14 +33,13 @@
 
 namespace parquet {
 
-using namespace parquet::schema;
-
 TEST(SizeStatistics, WriteBatchLevels) {
   std::vector<int64_t> expected_def_level_histogram = {256, 128, 64, 32, 16, 8, 4, 2, 2};
   std::vector<int64_t> expected_rep_level_histogram = {256, 128, 64, 32, 32};
   constexpr int16_t kMaxDefLevel = 8;
   constexpr int16_t kMaxRefLevel = 4;
-  auto descr = std::make_unique<ColumnDescriptor>(Int32("a"), kMaxDefLevel, kMaxRefLevel);
+  auto descr =
+      std::make_unique<ColumnDescriptor>(schema::Int32("a"), kMaxDefLevel, kMaxRefLevel);
   auto builder = SizeStatisticsBuilder::Make(descr.get());
 
   auto write_batch_levels =
@@ -73,7 +73,8 @@ TEST(SizeStatistics, WriteBatchLevels) {
 TEST(SizeStatistics, WriteRepeatedLevels) {
   constexpr int16_t kMaxDefLevel = 2;
   constexpr int16_t kMaxRepLevel = 3;
-  auto descr = std::make_unique<ColumnDescriptor>(Int32("a"), kMaxDefLevel, kMaxRepLevel);
+  auto descr =
+      std::make_unique<ColumnDescriptor>(schema::Int32("a"), kMaxDefLevel, kMaxRepLevel);
   auto builder = SizeStatisticsBuilder::Make(descr.get());
 
   constexpr int64_t kNumRounds = 10;