pitrou commented on a change in pull request #8136: URL: https://github.com/apache/arrow/pull/8136#discussion_r487921544
########## File path: cpp/src/parquet/arrow/schema.cc ########## @@ -682,48 +689,78 @@ Status GetOriginSchema(const std::shared_ptr<const KeyValueMetadata>& metadata, // Restore original Arrow field information that was serialized as Parquet metadata // but that is not necessarily present in the field reconstitued from Parquet data // (for example, Parquet timestamp types doesn't carry timezone information). -Status ApplyOriginalMetadata(std::shared_ptr<Field> field, const Field& origin_field, - std::shared_ptr<Field>* out) { + +Status ApplyOriginalStorageMetadata(const Field& origin_field, SchemaField* inferred) { auto origin_type = origin_field.type(); - if (field->type()->id() == ::arrow::Type::TIMESTAMP) { + auto inferred_type = inferred->field->type(); + + if (inferred_type->id() == ::arrow::Type::TIMESTAMP) { // Restore time zone, if any - const auto& ts_type = static_cast<const ::arrow::TimestampType&>(*field->type()); + const auto& ts_type = static_cast<const ::arrow::TimestampType&>(*inferred_type); const auto& ts_origin_type = static_cast<const ::arrow::TimestampType&>(*origin_type); // If the unit is the same and the data is tz-aware, then set the original // time zone, since Parquet has no native storage for timezones if (ts_type.unit() == ts_origin_type.unit() && ts_type.timezone() == "UTC" && ts_origin_type.timezone() != "") { - field = field->WithType(origin_type); + inferred->field = inferred->field->WithType(origin_type); } } + if (origin_type->id() == ::arrow::Type::DICTIONARY && - field->type()->id() != ::arrow::Type::DICTIONARY && - IsDictionaryReadSupported(*field->type())) { + inferred_type->id() != ::arrow::Type::DICTIONARY && + IsDictionaryReadSupported(*inferred_type)) { const auto& dict_origin_type = static_cast<const ::arrow::DictionaryType&>(*origin_type); - field = field->WithType( - ::arrow::dictionary(::arrow::int32(), field->type(), dict_origin_type.ordered())); + inferred->field = inferred->field->WithType( + ::arrow::dictionary(::arrow::int32(), inferred_type, dict_origin_type.ordered())); + } + + // Restore field metadata + std::shared_ptr<const KeyValueMetadata> field_metadata = origin_field.metadata(); + if (field_metadata != nullptr) { + if (inferred->field->metadata()) { + // Prefer the metadata keys (like field_id) from the current metadata + field_metadata = field_metadata->Merge(*inferred->field->metadata()); + } + inferred->field = inferred->field->WithMetadata(field_metadata); } if (origin_type->id() == ::arrow::Type::EXTENSION) { // Restore extension type, if the storage type is as read from Parquet const auto& ex_type = checked_cast<const ::arrow::ExtensionType&>(*origin_type); - if (ex_type.storage_type()->Equals(*field->type())) { - field = field->WithType(origin_type); + if (ex_type.storage_type()->Equals(*inferred_type)) { + inferred->field = inferred->field->WithType(origin_type); } } - // Restore field metadata - std::shared_ptr<const KeyValueMetadata> field_metadata = origin_field.metadata(); - if (field_metadata != nullptr) { - if (field->metadata()) { - // Prefer the metadata keys (like field_id) from the current metadata - field_metadata = field_metadata->Merge(*field->metadata()); + // TODO Should apply metadata recursively, but for that we need to move metadata + // application inside NodeToSchemaField (ARROW-9943) + + return Status::OK(); +} + +Status ApplyOriginalMetadata(const Field& origin_field, SchemaField* inferred) { + auto origin_type = origin_field.type(); + auto inferred_type = inferred->field->type(); + + if (origin_type->id() == ::arrow::Type::EXTENSION) { + const auto& ex_type = checked_cast<const ::arrow::ExtensionType&>(*origin_type); + auto origin_storage_field = origin_field.WithType(ex_type.storage_type()); + + // Apply metadata recursively to storage type Review comment: I should make the TODO more precise then, it is about applying to children types. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org