use of org.apache.arrow.vector.types.pojo.ArrowType in project textdb by TextDB.
the class NltkSentimentOperator method convertToTexeraSchema.
private Schema convertToTexeraSchema(org.apache.arrow.vector.types.pojo.Schema arrowSchema) {
List<Attribute> texeraAttributes = new ArrayList<>();
for (Field f : arrowSchema.getFields()) {
String attributeName = f.getName();
AttributeType attributeType;
ArrowType arrowType = f.getFieldType().getType();
switch(arrowType.getTypeID()) {
case Int:
attributeType = INTEGER;
break;
case FloatingPoint:
attributeType = DOUBLE;
break;
case Bool:
attributeType = BOOLEAN;
break;
case Utf8:
case Null:
attributeType = TEXT;
break;
case Date:
attributeType = DATE;
break;
case Struct:
// For now only Struct of DateTime
attributeType = DATETIME;
break;
case List:
attributeType = LIST;
break;
default:
throw (new DataflowException("Unsupported data type " + arrowType.getTypeID() + " when converting back to Texera table."));
}
texeraAttributes.add(new Attribute(attributeName, attributeType));
}
return new Schema(texeraAttributes);
}
use of org.apache.arrow.vector.types.pojo.ArrowType in project flink by apache.
the class ArrowUtilsTest method init.
@BeforeClass
public static void init() {
testFields = new ArrayList<>();
testFields.add(Tuple5.of("f1", new TinyIntType(), new ArrowType.Int(8, true), TinyIntWriter.TinyIntWriterForRow.class, ArrowTinyIntColumnVector.class));
testFields.add(Tuple5.of("f2", new SmallIntType(), new ArrowType.Int(8 * 2, true), SmallIntWriter.SmallIntWriterForRow.class, ArrowSmallIntColumnVector.class));
testFields.add(Tuple5.of("f3", new IntType(), new ArrowType.Int(8 * 4, true), IntWriter.IntWriterForRow.class, ArrowIntColumnVector.class));
testFields.add(Tuple5.of("f4", new BigIntType(), new ArrowType.Int(8 * 8, true), BigIntWriter.BigIntWriterForRow.class, ArrowBigIntColumnVector.class));
testFields.add(Tuple5.of("f5", new BooleanType(), new ArrowType.Bool(), BooleanWriter.BooleanWriterForRow.class, ArrowBooleanColumnVector.class));
testFields.add(Tuple5.of("f6", new FloatType(), new ArrowType.FloatingPoint(FloatingPointPrecision.SINGLE), FloatWriter.FloatWriterForRow.class, ArrowFloatColumnVector.class));
testFields.add(Tuple5.of("f7", new DoubleType(), new ArrowType.FloatingPoint(FloatingPointPrecision.DOUBLE), DoubleWriter.DoubleWriterForRow.class, ArrowDoubleColumnVector.class));
testFields.add(Tuple5.of("f8", new VarCharType(), ArrowType.Utf8.INSTANCE, VarCharWriter.VarCharWriterForRow.class, ArrowVarCharColumnVector.class));
testFields.add(Tuple5.of("f9", new VarBinaryType(), ArrowType.Binary.INSTANCE, VarBinaryWriter.VarBinaryWriterForRow.class, ArrowVarBinaryColumnVector.class));
testFields.add(Tuple5.of("f10", new DecimalType(10, 3), new ArrowType.Decimal(10, 3), DecimalWriter.DecimalWriterForRow.class, ArrowDecimalColumnVector.class));
testFields.add(Tuple5.of("f11", new DateType(), new ArrowType.Date(DateUnit.DAY), DateWriter.DateWriterForRow.class, ArrowDateColumnVector.class));
testFields.add(Tuple5.of("f13", new TimeType(0), new ArrowType.Time(TimeUnit.SECOND, 32), TimeWriter.TimeWriterForRow.class, ArrowTimeColumnVector.class));
testFields.add(Tuple5.of("f14", new TimeType(2), new ArrowType.Time(TimeUnit.MILLISECOND, 32), TimeWriter.TimeWriterForRow.class, ArrowTimeColumnVector.class));
testFields.add(Tuple5.of("f15", new TimeType(4), new ArrowType.Time(TimeUnit.MICROSECOND, 64), TimeWriter.TimeWriterForRow.class, ArrowTimeColumnVector.class));
testFields.add(Tuple5.of("f16", new TimeType(8), new ArrowType.Time(TimeUnit.NANOSECOND, 64), TimeWriter.TimeWriterForRow.class, ArrowTimeColumnVector.class));
testFields.add(Tuple5.of("f17", new LocalZonedTimestampType(0), new ArrowType.Timestamp(TimeUnit.SECOND, null), TimestampWriter.TimestampWriterForRow.class, ArrowTimestampColumnVector.class));
testFields.add(Tuple5.of("f18", new LocalZonedTimestampType(2), new ArrowType.Timestamp(TimeUnit.MILLISECOND, null), TimestampWriter.TimestampWriterForRow.class, ArrowTimestampColumnVector.class));
testFields.add(Tuple5.of("f19", new LocalZonedTimestampType(4), new ArrowType.Timestamp(TimeUnit.MICROSECOND, null), TimestampWriter.TimestampWriterForRow.class, ArrowTimestampColumnVector.class));
testFields.add(Tuple5.of("f20", new LocalZonedTimestampType(8), new ArrowType.Timestamp(TimeUnit.NANOSECOND, null), TimestampWriter.TimestampWriterForRow.class, ArrowTimestampColumnVector.class));
testFields.add(Tuple5.of("f21", new TimestampType(0), new ArrowType.Timestamp(TimeUnit.SECOND, null), TimestampWriter.TimestampWriterForRow.class, ArrowTimestampColumnVector.class));
testFields.add(Tuple5.of("f22", new TimestampType(2), new ArrowType.Timestamp(TimeUnit.MILLISECOND, null), TimestampWriter.TimestampWriterForRow.class, ArrowTimestampColumnVector.class));
testFields.add(Tuple5.of("f23", new TimestampType(4), new ArrowType.Timestamp(TimeUnit.MICROSECOND, null), TimestampWriter.TimestampWriterForRow.class, ArrowTimestampColumnVector.class));
testFields.add(Tuple5.of("f24", new TimestampType(8), new ArrowType.Timestamp(TimeUnit.NANOSECOND, null), TimestampWriter.TimestampWriterForRow.class, ArrowTimestampColumnVector.class));
testFields.add(Tuple5.of("f25", new ArrayType(new VarCharType()), ArrowType.List.INSTANCE, ArrayWriter.ArrayWriterForRow.class, ArrowArrayColumnVector.class));
RowType rowFieldType = new RowType(Arrays.asList(new RowType.RowField("a", new IntType()), new RowType.RowField("b", new VarCharType()), new RowType.RowField("c", new ArrayType(new VarCharType())), new RowType.RowField("d", new TimestampType(2)), new RowType.RowField("e", new RowType((Arrays.asList(new RowType.RowField("e1", new IntType()), new RowType.RowField("e2", new VarCharType())))))));
testFields.add(Tuple5.of("f26", rowFieldType, ArrowType.Struct.INSTANCE, RowWriter.RowWriterForRow.class, ArrowRowColumnVector.class));
List<RowType.RowField> rowFields = new ArrayList<>();
for (Tuple5<String, LogicalType, ArrowType, Class<?>, Class<?>> field : testFields) {
rowFields.add(new RowType.RowField(field.f0, field.f1));
}
rowType = new RowType(rowFields);
allocator = ArrowUtils.getRootAllocator().newChildAllocator("stdout", 0, Long.MAX_VALUE);
}
Aggregations