target multiple fragments per page to attempt to stay under page size…

… limit

cpp/src/io/parquet/writer_impl.cu

@@ -1463,10 +1463,18 @@ void writer::impl::write(table_view const& table, std::vector<partition_info> co
   auto iter = thrust::make_counting_iterator<size_type>(0);
   if (table.num_rows() > 0 && max_page_fragment_size_ == cudf::io::default_max_page_fragment_size) {
     std::for_each(iter, iter + single_streams_table.num_columns(), [&](size_type index) {
-      auto const avg_len = util::div_rounding_up_safe(
-        column_size(single_streams_table.column(index), stream), table.num_rows());
+      // dividing page size by average row length will tend to overshoot the desired page
+      // size when there's high variability in the row lengths. instead, shoot for multiple
+      // fragments per page to smooth things out. using 2 was too unbalanced in final page
+      // sizes, so using 4 which seems to be a good compromise at smoothing things out without
+      // getting fragment sizes too small.
+      constexpr int target_frags_per_page = 4;
+      auto const avg_len =
+        target_frags_per_page *
+        util::div_rounding_up_safe(column_size(single_streams_table.column(index), stream),
+                                   table.num_rows());
       if (avg_len > 0) {
-        auto const frag_size    = util::div_rounding_up_safe<size_type>(max_page_size_bytes, avg_len);
+        auto const frag_size = util::div_rounding_up_safe<size_type>(max_page_size_bytes, avg_len);
         column_frag_size[index] = std::min<size_type>(max_page_fragment_size_, frag_size);
       }
     });