GH-34375: [C++][Parquet] Ignore page header stats when page index enabled

Author: wgtmac

cpp/src/parquet/arrow/arrow_reader_writer_test.cc

@@ -5252,17 +5252,34 @@ class ParquetPageIndexRoundTripTest : public ::testing::Test {
       auto row_group_index_reader = page_index_reader->RowGroup(rg);
       ASSERT_NE(row_group_index_reader, nullptr);
 
+      auto row_group_reader = reader->RowGroup(rg);
+      ASSERT_NE(row_group_reader, nullptr);
+
       for (int col = 0; col < metadata->num_columns(); ++col) {
         auto column_index = row_group_index_reader->GetColumnIndex(col);
         column_indexes_.emplace_back(column_index.get());
 
+        bool expect_no_page_index =
+            expect_columns_without_index.find(col) != expect_columns_without_index.cend();
+
         auto offset_index = row_group_index_reader->GetOffsetIndex(col);
-        if (expect_columns_without_index.find(col) !=
-            expect_columns_without_index.cend()) {
+        if (expect_no_page_index) {
           ASSERT_EQ(offset_index, nullptr);
         } else {
           CheckOffsetIndex(offset_index.get(), expect_num_pages, &offset_lower_bound);
         }
+
+        // Verify page stats are not written to page header if page index is enabled.
+        auto page_reader = row_group_reader->GetColumnPageReader(col);
+        ASSERT_NE(page_reader, nullptr);
+        std::shared_ptr<Page> page = nullptr;
+        while ((page = page_reader->NextPage()) != nullptr) {
+          if (page->type() == PageType::DATA_PAGE ||
+              page->type() == PageType::DATA_PAGE_V2) {
+            ASSERT_EQ(std::static_pointer_cast<DataPage>(page)->statistics().is_set(),
+                      expect_no_page_index);
+          }
+        }
       }
     }
   }

cpp/src/parquet/column_writer.cc

@@ -460,7 +460,11 @@ class SerializedPageWriter : public PageWriter {
         ToThrift(page.definition_level_encoding()));
     data_page_header.__set_repetition_level_encoding(
         ToThrift(page.repetition_level_encoding()));
-    data_page_header.__set_statistics(ToThrift(page.statistics()));
+
+    // Write page statistics only when page index is not enabled.
+    if (column_index_builder_ == nullptr) {
+      data_page_header.__set_statistics(ToThrift(page.statistics()));
+    }
 
     page_header.__set_type(format::PageType::DATA_PAGE);
     page_header.__set_data_page_header(data_page_header);
@@ -479,7 +483,11 @@ class SerializedPageWriter : public PageWriter {
         page.repetition_levels_byte_length());
 
     data_page_header.__set_is_compressed(page.is_compressed());
-    data_page_header.__set_statistics(ToThrift(page.statistics()));
+
+    // Write page statistics only when page index is not enabled.
+    if (column_index_builder_ == nullptr) {
+      data_page_header.__set_statistics(ToThrift(page.statistics()));
+    }
 
     page_header.__set_type(format::PageType::DATA_PAGE_V2);
     page_header.__set_data_page_header_v2(data_page_header);