balaji-varadarajan-ai commented on code in PR #18328: URL: https://github.com/apache/hudi/pull/18328#discussion_r2957337090
########## hudi-client/hudi-spark-client/src/main/java/org/apache/hudi/io/storage/VectorConversionUtils.java: ########## @@ -0,0 +1,205 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one + * or more contributor license agreements. See the NOTICE file + * distributed with this work for additional information + * regarding copyright ownership. The ASF licenses this file + * to you under the Apache License, Version 2.0 (the + * "License"); you may not use this file except in compliance + * with the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package org.apache.hudi.io.storage; + +import org.apache.hudi.common.schema.HoodieSchema; +import org.apache.hudi.common.schema.HoodieSchemaField; +import org.apache.hudi.common.schema.HoodieSchemaType; + +import org.apache.spark.sql.catalyst.util.GenericArrayData; +import org.apache.spark.sql.types.BinaryType$; +import org.apache.spark.sql.types.Metadata; +import org.apache.spark.sql.types.StructField; +import org.apache.spark.sql.types.StructType; + +import org.apache.spark.sql.catalyst.InternalRow; +import org.apache.spark.sql.catalyst.expressions.GenericInternalRow; + +import java.nio.ByteBuffer; +import java.util.HashMap; +import java.util.List; +import java.util.Map; + +import static org.apache.hudi.common.util.ValidationUtils.checkArgument; + +/** + * Shared utility methods for vector column handling during Parquet read/write. + * + * Vectors are stored as Parquet FIXED_LEN_BYTE_ARRAY columns. On read, Spark maps these + * to BinaryType. This class provides the canonical conversion between the binary + * representation and Spark's typed ArrayData (float[], double[], byte[]). + * + * All byte buffers use little-endian order ({@link HoodieSchema.VectorLogicalType#VECTOR_BYTE_ORDER}) + * for compatibility with common vector search libraries (FAISS, ScaNN, etc.) and to match + * native x86/ARM byte order for zero-copy reads. + */ +public final class VectorConversionUtils { + + private VectorConversionUtils() { + } + + /** + * Detects VECTOR columns in a HoodieSchema record and returns a map of field ordinal + * to the corresponding {@link HoodieSchema.Vector} schema. + * + * @param schema a HoodieSchema of type RECORD (or null) + * @return map from field index to Vector schema; empty map if schema is null or has no vectors + */ + public static Map<Integer, HoodieSchema.Vector> detectVectorColumns(HoodieSchema schema) { Review Comment: Just checking, As we are using integer ordinal position in the schema, can you check if things end to end with projections and schema evolution? ########## hudi-client/hudi-spark-client/src/main/java/org/apache/hudi/io/storage/HoodieSparkParquetReader.java: ########## @@ -142,10 +144,20 @@ public ClosableIterator<UnsafeRow> getUnsafeRowIterator(HoodieSchema requestedSc public ClosableIterator<UnsafeRow> getUnsafeRowIterator(HoodieSchema requestedSchema, List<Filter> readFilters) throws IOException { HoodieSchema nonNullSchema = requestedSchema.getNonNullType(); StructType structSchema = HoodieInternalRowUtils.getCachedSchema(nonNullSchema); + + // Detect vector columns: ordinal → Vector schema + Map<Integer, HoodieSchema.Vector> vectorColumnInfo = VectorConversionUtils.detectVectorColumns(nonNullSchema); Review Comment: seeing the pattern: 1. Detecting vector columns. 2. Replacing Schema 3. Post-process rows in HoodieSparkParquetReader, SparkFileFormatInternalRowReaderContext and HoodieFileGroupReaderBasedFileFormat. Wondering if you can bring them under one common method with specific callback. ########## hudi-client/hudi-spark-client/src/main/java/org/apache/hudi/io/storage/row/HoodieRowParquetWriteSupport.java: ########## @@ -518,6 +553,14 @@ private Type convertField(HoodieSchema fieldSchema, StructField structField, Typ .as(LogicalTypeAnnotation.decimalType(scale, precision)) .length(Decimal.minBytesForPrecision()[precision]) .named(structField.name()); + } else if (dataType instanceof ArrayType + && resolvedSchema != null + && resolvedSchema.getType() == HoodieSchemaType.VECTOR) { + HoodieSchema.Vector vectorSchema = (HoodieSchema.Vector) resolvedSchema; + int fixedSize = vectorSchema.getDimension() + * vectorSchema.getVectorElementType().getElementSize(); + return Types.primitive(FIXED_LEN_BYTE_ARRAY, repetition) Review Comment: The vectors are stored as bare FIXED_LEN_BYTE_ARRAY in Parquet with no logical type annotation or key-value metadata on the Parquet column. I think it would be useful to track this. Any thoughts? -- 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. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
