arrow-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Chaokun Yang (JIRA)" <j...@apache.org>
Subject [jira] [Created] (ARROW-5231) [Java] Arrow Java can't read union vector from ArrowStreamReader written by its own bugs
Date Mon, 29 Apr 2019 06:15:00 GMT
Chaokun Yang created ARROW-5231:
-----------------------------------

             Summary: [Java] Arrow Java can't  read union vector from ArrowStreamReader written
by its own bugs
                 Key: ARROW-5231
                 URL: https://issues.apache.org/jira/browse/ARROW-5231
             Project: Apache Arrow
          Issue Type: Bug
          Components: Java
         Environment: Mac OS 10.13.6, Arrow 0.13.0, JDK8
            Reporter: Chaokun Yang


Similar  to https://issues.apache.org/jira/browse/ARROW-5230, when I write union data using 
ArrowStreamWriter in java, I can't read it back using ArrowStreamReader in java. The exception
is:
{quote}Exception in thread "main" java.lang.IllegalArgumentException: not all nodes and buffers
were consumed. nodes: [ArrowFieldNode [length=100, nullCount=0]] buffers: [ArrowBuf[14], udle:
[7 104..117], ArrowBuf[15], udle: [7 120..520]]
 at org.apache.arrow.vector.VectorLoader.load(VectorLoader.java:64)
 at org.apache.arrow.vector.ipc.ArrowReader.loadRecordBatch(ArrowReader.java:219)
 at org.apache.arrow.vector.ipc.ArrowStreamReader.loadNextBatch(ArrowStreamReader.java:121)
{quote}
The code to reproduce this exception is:

 
{code:java}
import org.apache.arrow.memory.RootAllocator;
import org.apache.arrow.vector.FieldVector;
import org.apache.arrow.vector.VectorSchemaRoot;
import org.apache.arrow.vector.complex.UnionVector;
import org.apache.arrow.vector.dictionary.DictionaryProvider;
import org.apache.arrow.vector.holders.NullableIntHolder;
import org.apache.arrow.vector.ipc.ArrowStreamReader;
import org.apache.arrow.vector.ipc.ArrowStreamWriter;
import org.apache.arrow.vector.types.UnionMode;
import org.apache.arrow.vector.types.pojo.ArrowType;
import org.apache.arrow.vector.types.pojo.Field;
import org.apache.arrow.vector.types.pojo.FieldType;
import org.apache.arrow.vector.types.pojo.Schema;

import java.io.ByteArrayInputStream;
import java.io.FileOutputStream;
import java.io.IOException;
import java.io.OutputStream;
import java.nio.file.Files;
import java.nio.file.Paths;
import java.util.Collections;
import java.util.List;

public class UnionTest {

    public static void writeUnionBatch(OutputStream os) throws IOException {
        int[] typeIds = new int[]{
                ArrowType.ArrowTypeID.Int.ordinal()};
        ArrowType.Union union = new ArrowType.Union(UnionMode.Sparse, typeIds);
        Field field = new Field("f1", FieldType.nullable(union), null);
        List<Field> fields = Collections.singletonList(field);
        Schema schema = new Schema(fields);
        VectorSchemaRoot root = VectorSchemaRoot.create(schema, new RootAllocator(Integer.MAX_VALUE));
        DictionaryProvider.MapDictionaryProvider provider = new DictionaryProvider.MapDictionaryProvider();
        ArrowStreamWriter writer = new ArrowStreamWriter(root, provider, os);
        writer.start();
        for (int i = 0; i < 2; i++) {
            root.setRowCount(100);
            List<FieldVector> vectors = root.getFieldVectors();
            UnionVector vector = (UnionVector) vectors.get(0);
            fillVector(vector, 100);
            for (int j = 0; j < 100; j++) {
                if (!vector.isNull(j)) {
                    System.out.println(vector.getObject(j));
                }
            }
            writer.writeBatch();
        }
        writer.end();
        writer.close();
    }

    private static void fillVector(UnionVector vector, int batchSize) {
        vector.setInitialCapacity(batchSize);
        vector.allocateNew();
        for (int i = 0; i < batchSize; i++) {
            NullableIntHolder intHolder = new NullableIntHolder();
            intHolder.isSet = 1;
            intHolder.value = i;
            vector.setSafe(i, intHolder);
        }
        vector.setValueCount(batchSize);
    }


    public static void main(String[] args) throws IOException {
        try(FileOutputStream fos = new FileOutputStream("result/union.arrow")) {
            writeUnionBatch(fos);
            System.out.println("write succeed");
            fos.flush();
        }

        RootAllocator allocator = new RootAllocator(1000000000);
        ByteArrayInputStream in = new ByteArrayInputStream(Files.readAllBytes(Paths.get("result/union.arrow")));
        ArrowStreamReader reader = new ArrowStreamReader(in, allocator);
        reader.loadNextBatch();
    }
}

{code}
And it can't read union data generated by python, as is reported in https://issues.apache.org/jira/browse/ARROW-1692. 

It seems strange arrow java can't read union data generated by its own. Is there any format
gap between arrow java UnionVector write and read?

 



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Mime
View raw message