review changes

cpp/src/io/orc/reader_impl.cu

@@ -426,12 +426,16 @@ table_with_metadata reader::impl::read(size_type skip_rows,
   std::vector<data_type> column_types;
   for (const auto &col : _selected_columns) {
     auto col_type = to_type_id(_metadata->ff.types[col], _use_np_dtypes, _timestamp_type.id());
-    auto scale    = (col_type == type_id::DECIMAL64) ? _metadata->ff.types[col].scale : 0;
     CUDF_EXPECTS(col_type != type_id::EMPTY, "Unknown type");
     // Remove this once we support Decimal128 data type
     CUDF_EXPECTS((col_type != type_id::DECIMAL64) or (_metadata->ff.types[col].precision <= 18),
                  "Decimal data has precision > 18, Decimal64 data type doesn't support it.");
-    column_types.emplace_back(col_type, -1 * static_cast<int32_t>(scale));
+    // sign of the scale is changed since cuDF follows c++ libraries like CNL
+    // which uses negative scaling, but liborc and other libraries
+    // follow positive scaling.
+    auto scale =
+      (col_type == type_id::DECIMAL64) ? -static_cast<int32_t>(_metadata->ff.types[col].scale) : 0;
+    column_types.emplace_back(col_type, scale);
 
     // Map each ORC column to its column
     orc_col_map[col] = column_types.size() - 1;

cpp/src/io/orc/stripe_data.cu

@@ -991,6 +991,8 @@ static const __device__ __constant__ int64_t kPow5i[28] = {1,
  *
  * @param[in] bs Input byte stream
  * @param[in,out] vals on input: scale from secondary stream, on output: value
+ * @param[in] val_scale Scale of each value
+ * @param[in] col_scale Scale from schema to which value will be adjusted
  * @param[in] numvals Number of values to decode
  * @param[in] t thread id
  *
@@ -1027,7 +1029,10 @@ static __device__ int Decode_Decimals(orc_bytestream_s *bs,
     if (t >= num_vals_read and t < num_vals_to_read) {
       auto const pos = static_cast<int>(vals.i64[t]);
       int128_s v     = decode_varint128(bs, pos);
-      int32_t scale  = (t < numvals) ? col_scale - val_scale : 0;
+      // Since cuDF column stores just one scale, value needs to
+      // be adjusted to col_scale from val_scale. So the difference
+      // of them will be used to add 0s or remove digits.
+      int32_t scale = (t < numvals) ? col_scale - val_scale : 0;
       if (scale >= 0) {
         scale       = min(scale, 27);
         vals.i64[t] = ((int64_t)v.lo * kPow5i[scale]) << scale;