supporting more physical types of parquet time columns with time zone…

… info
Tmonster · Dec 7, 2023 · c6bf4c6 · c6bf4c6
1 parent baf670f
commit c6bf4c6
Show file tree

Hide file tree

Showing 3 changed files with 26 additions and 9 deletions.
diff --git a/extension/parquet/column_reader.cpp b/extension/parquet/column_reader.cpp
@@ -1469,11 +1469,18 @@ unique_ptr<ColumnReader> ColumnReader::CreateReader(ParquetReader &reader, const
 				break;
 			}
 		}
+		throw NotImplementedException("Unsupported time encoding in Parquet file");
 	case LogicalTypeId::TIME_TZ:
 		if (schema_p.__isset.logicalType && schema_p.logicalType.__isset.TIME) {
-			if (schema_p.logicalType.TIME.unit.__isset.MICROS) {
+			if (schema_p.logicalType.TIME.unit.__isset.MILLIS) {
+				return make_uniq<CallbackColumnReader<int32_t, dtime_tz_t, ParquetIntToTimeMsTZ>>(
+				    reader, type_p, schema_p, file_idx_p, max_define, max_repeat);
+			} else if (schema_p.logicalType.TIME.unit.__isset.MICROS) {
 				return make_uniq<CallbackColumnReader<int64_t, dtime_tz_t, ParquetIntToTimeTZ>>(
 				    reader, type_p, schema_p, file_idx_p, max_define, max_repeat);
+			} else if (schema_p.logicalType.TIME.unit.__isset.NANOS) {
+				return make_uniq<CallbackColumnReader<int64_t, dtime_tz_t, ParquetIntToTimeNsTZ>>(
+				    reader, type_p, schema_p, file_idx_p, max_define, max_repeat);
 			}
 		} else if (schema_p.__isset.converted_type) {
 			switch (schema_p.converted_type) {
@@ -1484,6 +1491,7 @@ unique_ptr<ColumnReader> ColumnReader::CreateReader(ParquetReader &reader, const
 				break;
 			}
 		}
+		throw NotImplementedException("Unsupported time encoding in Parquet file");
 	case LogicalTypeId::BLOB:
 	case LogicalTypeId::VARCHAR:
 		return make_uniq<StringColumnReader>(reader, type_p, schema_p, file_idx_p, max_define, max_repeat);

diff --git a/extension/parquet/include/parquet_timestamp.hpp b/extension/parquet/include/parquet_timestamp.hpp
@@ -25,6 +25,8 @@ date_t ParquetIntToDate(const int32_t &raw_date);
 dtime_t ParquetIntToTimeMs(const int32_t &raw_time);
 dtime_t ParquetIntToTime(const int64_t &raw_time);
 dtime_t ParquetIntToTimeNs(const int64_t &raw_time);
+dtime_tz_t ParquetIntToTimeMsTZ(const int32_t &raw_time);
 dtime_tz_t ParquetIntToTimeTZ(const int64_t &raw_time);
+dtime_tz_t ParquetIntToTimeNsTZ(const int64_t &raw_time);
 
 } // namespace duckdb
diff --git a/extension/parquet/parquet_timestamp.cpp b/extension/parquet/parquet_timestamp.cpp
@@ -54,21 +54,28 @@ date_t ParquetIntToDate(const int32_t &raw_date) {
 	return date_t(raw_date);
 }
 
-dtime_t ParquetIntToTimeMs(const int32_t &raw_time) {
-	return Time::FromTimeMs(raw_time);
+dtime_t ParquetIntToTimeMs(const int32_t &raw_millis) {
+	return Time::FromTimeMs(raw_millis);
 }
 
-dtime_t ParquetIntToTime(const int64_t &raw_time) {
-	return dtime_t(raw_time);
+dtime_t ParquetIntToTime(const int64_t &raw_micros) {
+	return dtime_t(raw_micros);
 }
 
-dtime_t ParquetIntToTimeNs(const int64_t &raw_time) {
-	return Time::FromTimeNs(raw_time);
+dtime_t ParquetIntToTimeNs(const int64_t &raw_nanos) {
+	return Time::FromTimeNs(raw_nanos);
+}
+
+dtime_tz_t ParquetIntToTimeMsTZ(const int32_t &raw_millis) {
+	return dtime_tz_t(Time::FromTimeMs(raw_millis), 0);
 }
 
 dtime_tz_t ParquetIntToTimeTZ(const int64_t &raw_micros) {
-	dtime_t t(raw_micros);
-	return dtime_tz_t(t, 0);
+	return dtime_tz_t(dtime_t(raw_micros), 0);
+}
+
+dtime_tz_t ParquetIntToTimeNsTZ(const int64_t &raw_nanos) {
+	return dtime_tz_t(Time::FromTimeNs(raw_nanos), 0);
 }
 
 } // namespace duckdb