use of org.apache.hadoop.hive.serde2.io.ByteWritable in project hive by apache.
the class TypedBytesRecordReader method write.
private void write(int pos, Writable inpw) throws IOException {
String typ = columnTypes.get(pos);
Writable w = (Writable) converters.get(pos).convert(inpw);
if (typ.equalsIgnoreCase(serdeConstants.BOOLEAN_TYPE_NAME)) {
tbOut.writeBoolean((BooleanWritable) w);
} else if (typ.equalsIgnoreCase(serdeConstants.TINYINT_TYPE_NAME)) {
tbOut.writeByte((ByteWritable) w);
} else if (typ.equalsIgnoreCase(serdeConstants.SMALLINT_TYPE_NAME)) {
tbOut.writeShort((ShortWritable) w);
} else if (typ.equalsIgnoreCase(serdeConstants.INT_TYPE_NAME)) {
tbOut.writeInt((IntWritable) w);
} else if (typ.equalsIgnoreCase(serdeConstants.BIGINT_TYPE_NAME)) {
tbOut.writeLong((LongWritable) w);
} else if (typ.equalsIgnoreCase(serdeConstants.FLOAT_TYPE_NAME)) {
tbOut.writeFloat((FloatWritable) w);
} else if (typ.equalsIgnoreCase(serdeConstants.DOUBLE_TYPE_NAME)) {
tbOut.writeDouble((DoubleWritable) w);
} else if (typ.equalsIgnoreCase(serdeConstants.STRING_TYPE_NAME)) {
tbOut.writeText((Text) w);
} else {
assert false;
}
}
use of org.apache.hadoop.hive.serde2.io.ByteWritable in project hive by apache.
the class VectorColumnAssignFactory method buildObjectAssign.
public static VectorColumnAssign buildObjectAssign(VectorizedRowBatch outputBatch, int outColIndex, PrimitiveCategory category) throws HiveException {
VectorColumnAssign outVCA = null;
ColumnVector destCol = outputBatch.cols[outColIndex];
if (destCol == null) {
switch(category) {
case VOID:
outVCA = new VectorLongColumnAssign() {
// This is a dummy assigner
@Override
public void assignObjectValue(Object val, int destIndex) throws HiveException {
// This is no-op, there is no column to assign to and val is expected to be null
assert (val == null);
}
};
break;
default:
throw new HiveException("Incompatible (null) vector column and primitive category " + category);
}
} else if (destCol instanceof LongColumnVector) {
switch(category) {
case BOOLEAN:
outVCA = new VectorLongColumnAssign() {
@Override
public void assignObjectValue(Object val, int destIndex) throws HiveException {
if (val == null) {
assignNull(destIndex);
} else {
BooleanWritable bw = (BooleanWritable) val;
assignLong(bw.get() ? 1 : 0, destIndex);
}
}
}.init(outputBatch, (LongColumnVector) destCol);
break;
case BYTE:
outVCA = new VectorLongColumnAssign() {
@Override
public void assignObjectValue(Object val, int destIndex) throws HiveException {
if (val == null) {
assignNull(destIndex);
} else {
ByteWritable bw = (ByteWritable) val;
assignLong(bw.get(), destIndex);
}
}
}.init(outputBatch, (LongColumnVector) destCol);
break;
case SHORT:
outVCA = new VectorLongColumnAssign() {
@Override
public void assignObjectValue(Object val, int destIndex) throws HiveException {
if (val == null) {
assignNull(destIndex);
} else {
ShortWritable bw = (ShortWritable) val;
assignLong(bw.get(), destIndex);
}
}
}.init(outputBatch, (LongColumnVector) destCol);
break;
case INT:
outVCA = new VectorLongColumnAssign() {
@Override
public void assignObjectValue(Object val, int destIndex) throws HiveException {
if (val == null) {
assignNull(destIndex);
} else {
IntWritable bw = (IntWritable) val;
assignLong(bw.get(), destIndex);
}
}
}.init(outputBatch, (LongColumnVector) destCol);
break;
case LONG:
outVCA = new VectorLongColumnAssign() {
@Override
public void assignObjectValue(Object val, int destIndex) throws HiveException {
if (val == null) {
assignNull(destIndex);
} else {
LongWritable bw = (LongWritable) val;
assignLong(bw.get(), destIndex);
}
}
}.init(outputBatch, (LongColumnVector) destCol);
break;
case TIMESTAMP:
outVCA = new VectorTimestampColumnAssign() {
@Override
public void assignObjectValue(Object val, int destIndex) throws HiveException {
if (val == null) {
assignNull(destIndex);
} else {
assignTimestamp((TimestampWritable) val, destIndex);
}
}
}.init(outputBatch, (TimestampColumnVector) destCol);
break;
case DATE:
outVCA = new VectorLongColumnAssign() {
@Override
public void assignObjectValue(Object val, int destIndex) throws HiveException {
if (val == null) {
assignNull(destIndex);
} else {
DateWritable bw = (DateWritable) val;
assignLong(bw.getDays(), destIndex);
}
}
}.init(outputBatch, (LongColumnVector) destCol);
break;
case INTERVAL_YEAR_MONTH:
outVCA = new VectorLongColumnAssign() {
@Override
public void assignObjectValue(Object val, int destIndex) throws HiveException {
if (val == null) {
assignNull(destIndex);
} else {
HiveIntervalYearMonthWritable bw = (HiveIntervalYearMonthWritable) val;
assignLong(bw.getHiveIntervalYearMonth().getTotalMonths(), destIndex);
}
}
}.init(outputBatch, (LongColumnVector) destCol);
break;
case INTERVAL_DAY_TIME:
outVCA = new VectorIntervalDayTimeColumnAssign() {
@Override
public void assignObjectValue(Object val, int destIndex) throws HiveException {
if (val == null) {
assignNull(destIndex);
} else {
HiveIntervalDayTimeWritable bw = (HiveIntervalDayTimeWritable) val;
assignIntervalDayTime(bw.getHiveIntervalDayTime(), destIndex);
}
}
}.init(outputBatch, (IntervalDayTimeColumnVector) destCol);
break;
default:
throw new HiveException("Incompatible Long vector column and primitive category " + category);
}
} else if (destCol instanceof DoubleColumnVector) {
switch(category) {
case DOUBLE:
outVCA = new VectorDoubleColumnAssign() {
@Override
public void assignObjectValue(Object val, int destIndex) throws HiveException {
if (val == null) {
assignNull(destIndex);
} else {
DoubleWritable bw = (DoubleWritable) val;
assignDouble(bw.get(), destIndex);
}
}
}.init(outputBatch, (DoubleColumnVector) destCol);
break;
case FLOAT:
outVCA = new VectorDoubleColumnAssign() {
@Override
public void assignObjectValue(Object val, int destIndex) throws HiveException {
if (val == null) {
assignNull(destIndex);
} else {
FloatWritable bw = (FloatWritable) val;
assignDouble(bw.get(), destIndex);
}
}
}.init(outputBatch, (DoubleColumnVector) destCol);
break;
default:
throw new HiveException("Incompatible Double vector column and primitive category " + category);
}
} else if (destCol instanceof BytesColumnVector) {
switch(category) {
case BINARY:
outVCA = new VectorBytesColumnAssign() {
@Override
public void assignObjectValue(Object val, int destIndex) throws HiveException {
if (val == null) {
assignNull(destIndex);
} else {
BytesWritable bw = (BytesWritable) val;
byte[] bytes = bw.getBytes();
assignBytes(bytes, 0, bw.getLength(), destIndex);
}
}
}.init(outputBatch, (BytesColumnVector) destCol);
break;
case STRING:
outVCA = new VectorBytesColumnAssign() {
@Override
public void assignObjectValue(Object val, int destIndex) throws HiveException {
if (val == null) {
assignNull(destIndex);
} else {
Text bw = (Text) val;
byte[] bytes = bw.getBytes();
assignBytes(bytes, 0, bw.getLength(), destIndex);
}
}
}.init(outputBatch, (BytesColumnVector) destCol);
break;
case VARCHAR:
outVCA = new VectorBytesColumnAssign() {
@Override
public void assignObjectValue(Object val, int destIndex) throws HiveException {
if (val == null) {
assignNull(destIndex);
} else {
// We store VARCHAR type stripped of pads.
HiveVarchar hiveVarchar;
if (val instanceof HiveVarchar) {
hiveVarchar = (HiveVarchar) val;
} else {
hiveVarchar = ((HiveVarcharWritable) val).getHiveVarchar();
}
byte[] bytes = hiveVarchar.getValue().getBytes();
assignBytes(bytes, 0, bytes.length, destIndex);
}
}
}.init(outputBatch, (BytesColumnVector) destCol);
break;
case CHAR:
outVCA = new VectorBytesColumnAssign() {
@Override
public void assignObjectValue(Object val, int destIndex) throws HiveException {
if (val == null) {
assignNull(destIndex);
} else {
// We store CHAR type stripped of pads.
HiveChar hiveChar;
if (val instanceof HiveChar) {
hiveChar = (HiveChar) val;
} else {
hiveChar = ((HiveCharWritable) val).getHiveChar();
}
byte[] bytes = hiveChar.getStrippedValue().getBytes();
assignBytes(bytes, 0, bytes.length, destIndex);
}
}
}.init(outputBatch, (BytesColumnVector) destCol);
break;
default:
throw new HiveException("Incompatible Bytes vector column and primitive category " + category);
}
} else if (destCol instanceof DecimalColumnVector) {
switch(category) {
case DECIMAL:
outVCA = new VectorDecimalColumnAssign() {
@Override
public void assignObjectValue(Object val, int destIndex) throws HiveException {
if (val == null) {
assignNull(destIndex);
} else {
if (val instanceof HiveDecimal) {
assignDecimal((HiveDecimal) val, destIndex);
} else {
assignDecimal((HiveDecimalWritable) val, destIndex);
}
}
}
}.init(outputBatch, (DecimalColumnVector) destCol);
break;
default:
throw new HiveException("Incompatible Decimal vector column and primitive category " + category);
}
} else {
throw new HiveException("Unknown vector column type " + destCol.getClass().getName());
}
return outVCA;
}
use of org.apache.hadoop.hive.serde2.io.ByteWritable in project hive by apache.
the class VectorDeserializeRow method convertRowColumn.
/**
* Convert one row column value that is the current value in deserializeRead.
*
* We deserialize into a writable and then pass that writable to an instance of VectorAssignRow
* to convert the writable to the target data type and assign it into the VectorizedRowBatch.
*
* @param batch
* @param batchIndex
* @param logicalColumnIndex
* @throws IOException
*/
private void convertRowColumn(VectorizedRowBatch batch, int batchIndex, int logicalColumnIndex) throws IOException {
final int projectionColumnNum = projectionColumnNums[logicalColumnIndex];
Writable convertSourceWritable = convertSourceWritables[logicalColumnIndex];
switch(sourceCategories[logicalColumnIndex]) {
case PRIMITIVE:
{
switch(sourcePrimitiveCategories[logicalColumnIndex]) {
case VOID:
convertSourceWritable = null;
break;
case BOOLEAN:
((BooleanWritable) convertSourceWritable).set(deserializeRead.currentBoolean);
break;
case BYTE:
((ByteWritable) convertSourceWritable).set(deserializeRead.currentByte);
break;
case SHORT:
((ShortWritable) convertSourceWritable).set(deserializeRead.currentShort);
break;
case INT:
((IntWritable) convertSourceWritable).set(deserializeRead.currentInt);
break;
case LONG:
((LongWritable) convertSourceWritable).set(deserializeRead.currentLong);
break;
case TIMESTAMP:
((TimestampWritable) convertSourceWritable).set(deserializeRead.currentTimestampWritable);
break;
case DATE:
((DateWritable) convertSourceWritable).set(deserializeRead.currentDateWritable);
break;
case FLOAT:
((FloatWritable) convertSourceWritable).set(deserializeRead.currentFloat);
break;
case DOUBLE:
((DoubleWritable) convertSourceWritable).set(deserializeRead.currentDouble);
break;
case BINARY:
if (deserializeRead.currentBytes == null) {
LOG.info("null binary entry: batchIndex " + batchIndex + " projection column num " + projectionColumnNum);
}
((BytesWritable) convertSourceWritable).set(deserializeRead.currentBytes, deserializeRead.currentBytesStart, deserializeRead.currentBytesLength);
break;
case STRING:
if (deserializeRead.currentBytes == null) {
throw new RuntimeException("null string entry: batchIndex " + batchIndex + " projection column num " + projectionColumnNum);
}
// Use org.apache.hadoop.io.Text as our helper to go from byte[] to String.
((Text) convertSourceWritable).set(deserializeRead.currentBytes, deserializeRead.currentBytesStart, deserializeRead.currentBytesLength);
break;
case VARCHAR:
{
// that does not use Java String objects.
if (deserializeRead.currentBytes == null) {
throw new RuntimeException("null varchar entry: batchIndex " + batchIndex + " projection column num " + projectionColumnNum);
}
int adjustedLength = StringExpr.truncate(deserializeRead.currentBytes, deserializeRead.currentBytesStart, deserializeRead.currentBytesLength, maxLengths[logicalColumnIndex]);
((HiveVarcharWritable) convertSourceWritable).set(new String(deserializeRead.currentBytes, deserializeRead.currentBytesStart, adjustedLength, Charsets.UTF_8), -1);
}
break;
case CHAR:
{
// that does not use Java String objects.
if (deserializeRead.currentBytes == null) {
throw new RuntimeException("null char entry: batchIndex " + batchIndex + " projection column num " + projectionColumnNum);
}
int adjustedLength = StringExpr.rightTrimAndTruncate(deserializeRead.currentBytes, deserializeRead.currentBytesStart, deserializeRead.currentBytesLength, maxLengths[logicalColumnIndex]);
((HiveCharWritable) convertSourceWritable).set(new String(deserializeRead.currentBytes, deserializeRead.currentBytesStart, adjustedLength, Charsets.UTF_8), -1);
}
break;
case DECIMAL:
((HiveDecimalWritable) convertSourceWritable).set(deserializeRead.currentHiveDecimalWritable);
break;
case INTERVAL_YEAR_MONTH:
((HiveIntervalYearMonthWritable) convertSourceWritable).set(deserializeRead.currentHiveIntervalYearMonthWritable);
break;
case INTERVAL_DAY_TIME:
((HiveIntervalDayTimeWritable) convertSourceWritable).set(deserializeRead.currentHiveIntervalDayTimeWritable);
break;
default:
throw new RuntimeException("Primitive category " + sourcePrimitiveCategories[logicalColumnIndex] + " not supported");
}
}
break;
default:
throw new RuntimeException("Category " + sourceCategories[logicalColumnIndex] + " not supported");
}
/*
* Convert our source object we just read into the target object and store that in the
* VectorizedRowBatch.
*/
convertVectorAssignRow.assignConvertRowColumn(batch, batchIndex, logicalColumnIndex, convertSourceWritable);
}
use of org.apache.hadoop.hive.serde2.io.ByteWritable in project hive by apache.
the class VectorExtractRow method extractRowColumn.
/**
* Extract a row's column object from the ColumnVector at batchIndex in the VectorizedRowBatch.
*
* @param batch
* @param batchIndex
* @param logicalColumnIndex
* @return
*/
public Object extractRowColumn(VectorizedRowBatch batch, int batchIndex, int logicalColumnIndex) {
final int projectionColumnNum = projectionColumnNums[logicalColumnIndex];
ColumnVector colVector = batch.cols[projectionColumnNum];
if (colVector == null) {
// may ask for them..
return null;
}
int adjustedIndex = (colVector.isRepeating ? 0 : batchIndex);
if (!colVector.noNulls && colVector.isNull[adjustedIndex]) {
return null;
}
Category category = categories[logicalColumnIndex];
switch(category) {
case PRIMITIVE:
{
Writable primitiveWritable = primitiveWritables[logicalColumnIndex];
PrimitiveCategory primitiveCategory = primitiveCategories[logicalColumnIndex];
switch(primitiveCategory) {
case VOID:
return null;
case BOOLEAN:
((BooleanWritable) primitiveWritable).set(((LongColumnVector) batch.cols[projectionColumnNum]).vector[adjustedIndex] == 0 ? false : true);
return primitiveWritable;
case BYTE:
((ByteWritable) primitiveWritable).set((byte) ((LongColumnVector) batch.cols[projectionColumnNum]).vector[adjustedIndex]);
return primitiveWritable;
case SHORT:
((ShortWritable) primitiveWritable).set((short) ((LongColumnVector) batch.cols[projectionColumnNum]).vector[adjustedIndex]);
return primitiveWritable;
case INT:
((IntWritable) primitiveWritable).set((int) ((LongColumnVector) batch.cols[projectionColumnNum]).vector[adjustedIndex]);
return primitiveWritable;
case LONG:
((LongWritable) primitiveWritable).set(((LongColumnVector) batch.cols[projectionColumnNum]).vector[adjustedIndex]);
return primitiveWritable;
case TIMESTAMP:
((TimestampWritable) primitiveWritable).set(((TimestampColumnVector) batch.cols[projectionColumnNum]).asScratchTimestamp(adjustedIndex));
return primitiveWritable;
case DATE:
((DateWritable) primitiveWritable).set((int) ((LongColumnVector) batch.cols[projectionColumnNum]).vector[adjustedIndex]);
return primitiveWritable;
case FLOAT:
((FloatWritable) primitiveWritable).set((float) ((DoubleColumnVector) batch.cols[projectionColumnNum]).vector[adjustedIndex]);
return primitiveWritable;
case DOUBLE:
((DoubleWritable) primitiveWritable).set(((DoubleColumnVector) batch.cols[projectionColumnNum]).vector[adjustedIndex]);
return primitiveWritable;
case BINARY:
{
BytesColumnVector bytesColVector = ((BytesColumnVector) batch.cols[projectionColumnNum]);
byte[] bytes = bytesColVector.vector[adjustedIndex];
int start = bytesColVector.start[adjustedIndex];
int length = bytesColVector.length[adjustedIndex];
if (bytes == null) {
LOG.info("null binary entry: batchIndex " + batchIndex + " projection column num " + projectionColumnNum);
}
BytesWritable bytesWritable = (BytesWritable) primitiveWritable;
bytesWritable.set(bytes, start, length);
return primitiveWritable;
}
case STRING:
{
BytesColumnVector bytesColVector = ((BytesColumnVector) batch.cols[projectionColumnNum]);
byte[] bytes = bytesColVector.vector[adjustedIndex];
int start = bytesColVector.start[adjustedIndex];
int length = bytesColVector.length[adjustedIndex];
if (bytes == null) {
nullBytesReadError(primitiveCategory, batchIndex, projectionColumnNum);
}
// Use org.apache.hadoop.io.Text as our helper to go from byte[] to String.
((Text) primitiveWritable).set(bytes, start, length);
return primitiveWritable;
}
case VARCHAR:
{
BytesColumnVector bytesColVector = ((BytesColumnVector) batch.cols[projectionColumnNum]);
byte[] bytes = bytesColVector.vector[adjustedIndex];
int start = bytesColVector.start[adjustedIndex];
int length = bytesColVector.length[adjustedIndex];
if (bytes == null) {
nullBytesReadError(primitiveCategory, batchIndex, projectionColumnNum);
}
int adjustedLength = StringExpr.truncate(bytes, start, length, maxLengths[logicalColumnIndex]);
HiveVarcharWritable hiveVarcharWritable = (HiveVarcharWritable) primitiveWritable;
hiveVarcharWritable.set(new String(bytes, start, adjustedLength, Charsets.UTF_8), -1);
return primitiveWritable;
}
case CHAR:
{
BytesColumnVector bytesColVector = ((BytesColumnVector) batch.cols[projectionColumnNum]);
byte[] bytes = bytesColVector.vector[adjustedIndex];
int start = bytesColVector.start[adjustedIndex];
int length = bytesColVector.length[adjustedIndex];
if (bytes == null) {
nullBytesReadError(primitiveCategory, batchIndex, projectionColumnNum);
}
int adjustedLength = StringExpr.rightTrimAndTruncate(bytes, start, length, maxLengths[logicalColumnIndex]);
HiveCharWritable hiveCharWritable = (HiveCharWritable) primitiveWritable;
hiveCharWritable.set(new String(bytes, start, adjustedLength, Charsets.UTF_8), maxLengths[logicalColumnIndex]);
return primitiveWritable;
}
case DECIMAL:
// The HiveDecimalWritable set method will quickly copy the deserialized decimal writable fields.
((HiveDecimalWritable) primitiveWritable).set(((DecimalColumnVector) batch.cols[projectionColumnNum]).vector[adjustedIndex]);
return primitiveWritable;
case INTERVAL_YEAR_MONTH:
((HiveIntervalYearMonthWritable) primitiveWritable).set((int) ((LongColumnVector) batch.cols[projectionColumnNum]).vector[adjustedIndex]);
return primitiveWritable;
case INTERVAL_DAY_TIME:
((HiveIntervalDayTimeWritable) primitiveWritable).set(((IntervalDayTimeColumnVector) batch.cols[projectionColumnNum]).asScratchIntervalDayTime(adjustedIndex));
return primitiveWritable;
default:
throw new RuntimeException("Primitive category " + primitiveCategory.name() + " not supported");
}
}
default:
throw new RuntimeException("Category " + category.name() + " not supported");
}
}
use of org.apache.hadoop.hive.serde2.io.ByteWritable in project hive by apache.
the class TestOrcFile method testSeek.
@Test
public void testSeek() throws Exception {
ObjectInspector inspector;
synchronized (TestOrcFile.class) {
inspector = ObjectInspectorFactory.getReflectionObjectInspector(BigRow.class, ObjectInspectorFactory.ObjectInspectorOptions.JAVA);
}
Writer writer = OrcFile.createWriter(testFilePath, OrcFile.writerOptions(conf).inspector(inspector).stripeSize(200000).bufferSize(65536).rowIndexStride(1000));
Random rand = new Random(42);
final int COUNT = 32768;
long[] intValues = new long[COUNT];
double[] doubleValues = new double[COUNT];
String[] stringValues = new String[COUNT];
BytesWritable[] byteValues = new BytesWritable[COUNT];
String[] words = new String[128];
for (int i = 0; i < words.length; ++i) {
words[i] = Integer.toHexString(rand.nextInt());
}
for (int i = 0; i < COUNT / 2; ++i) {
intValues[2 * i] = rand.nextLong();
intValues[2 * i + 1] = intValues[2 * i];
stringValues[2 * i] = words[rand.nextInt(words.length)];
stringValues[2 * i + 1] = stringValues[2 * i];
}
for (int i = 0; i < COUNT; ++i) {
doubleValues[i] = rand.nextDouble();
byte[] buf = new byte[20];
rand.nextBytes(buf);
byteValues[i] = new BytesWritable(buf);
}
for (int i = 0; i < COUNT; ++i) {
writer.addRow(createRandomRow(intValues, doubleValues, stringValues, byteValues, words, i));
}
writer.close();
Reader reader = OrcFile.createReader(testFilePath, OrcFile.readerOptions(conf).filesystem(fs));
assertEquals(COUNT, reader.getNumberOfRows());
RecordReader rows = reader.rows();
OrcStruct row = null;
for (int i = COUNT - 1; i >= 0; --i) {
// we load the previous buffer of rows
if (i % COUNT == COUNT - 1) {
rows.seekToRow(i - (COUNT - 1));
}
rows.seekToRow(i);
row = (OrcStruct) rows.next(row);
BigRow expected = createRandomRow(intValues, doubleValues, stringValues, byteValues, words, i);
assertEquals(expected.boolean1.booleanValue(), ((BooleanWritable) row.getFieldValue(0)).get());
assertEquals(expected.byte1.byteValue(), ((ByteWritable) row.getFieldValue(1)).get());
assertEquals(expected.short1.shortValue(), ((ShortWritable) row.getFieldValue(2)).get());
assertEquals(expected.int1.intValue(), ((IntWritable) row.getFieldValue(3)).get());
assertEquals(expected.long1.longValue(), ((LongWritable) row.getFieldValue(4)).get());
assertEquals(expected.float1, ((FloatWritable) row.getFieldValue(5)).get(), 0.0001);
assertEquals(expected.double1, ((DoubleWritable) row.getFieldValue(6)).get(), 0.0001);
assertEquals(expected.bytes1, row.getFieldValue(7));
assertEquals(expected.string1, row.getFieldValue(8));
List<InnerStruct> expectedList = expected.middle.list;
List<OrcStruct> actualList = (List<OrcStruct>) ((OrcStruct) row.getFieldValue(9)).getFieldValue(0);
compareList(expectedList, actualList);
compareList(expected.list, (List<OrcStruct>) row.getFieldValue(10));
}
rows.close();
Iterator<StripeInformation> stripeIterator = reader.getStripes().iterator();
long offsetOfStripe2 = 0;
long offsetOfStripe4 = 0;
long lastRowOfStripe2 = 0;
for (int i = 0; i < 5; ++i) {
StripeInformation stripe = stripeIterator.next();
if (i < 2) {
lastRowOfStripe2 += stripe.getNumberOfRows();
} else if (i == 2) {
offsetOfStripe2 = stripe.getOffset();
lastRowOfStripe2 += stripe.getNumberOfRows() - 1;
} else if (i == 4) {
offsetOfStripe4 = stripe.getOffset();
}
}
boolean[] columns = new boolean[reader.getStatistics().length];
// long colulmn
columns[5] = true;
// text column
columns[9] = true;
rows = reader.rowsOptions(new Reader.Options().range(offsetOfStripe2, offsetOfStripe4 - offsetOfStripe2).include(columns));
rows.seekToRow(lastRowOfStripe2);
for (int i = 0; i < 2; ++i) {
row = (OrcStruct) rows.next(row);
BigRow expected = createRandomRow(intValues, doubleValues, stringValues, byteValues, words, (int) (lastRowOfStripe2 + i));
assertEquals(expected.long1.longValue(), ((LongWritable) row.getFieldValue(4)).get());
assertEquals(expected.string1, row.getFieldValue(8));
}
rows.close();
}
Aggregations