use of org.apache.drill.exec.vector.accessor.writer.UnionWriterImpl.VariantObjectWriter in project drill by apache.
the class BaseWriterBuilder method buildUnion.
private AbstractObjectWriter buildUnion(UnionVector vector, VectorDescrip descrip) {
if (vector == null) {
throw new UnsupportedOperationException("Dummy variant writer not yet supported");
}
final AbstractObjectWriter[] variants = new AbstractObjectWriter[MinorType.values().length];
final MetadataProvider mdProvider = descrip.childProvider();
int i = 0;
for (final MinorType type : vector.getField().getType().getSubTypeList()) {
// This call will create the vector if it does not yet exist.
// Will throw an exception for unsupported types.
// so call this only if the MajorType reports that the type
// already exists.
final ValueVector memberVector = vector.getMember(type);
final VectorDescrip memberDescrip = new VectorDescrip(mdProvider, i++, memberVector.getField());
variants[type.ordinal()] = buildVectorWriter(memberVector, memberDescrip);
}
return new VariantObjectWriter(new UnionWriterImpl(descrip.metadata, vector, variants));
}
use of org.apache.drill.exec.vector.accessor.writer.UnionWriterImpl.VariantObjectWriter in project drill by apache.
the class ColumnBuilder method buildUnionList.
/**
* Create a list based on a (possible) union. The list starts empty here. The client
* can then add types as they are discovered. The list itself will transition from a
* list of nulls (no child type), to a list of a single type, to a list of unions.
* The writer interface will consistently present the list as a list of unions, even
* when the list itself has no subtype or a single subtype.
* <p>
* List vectors (lists of unions) are not supported in
* Drill. The code here works up through the scan operator. But, other operators do
* not support the {@code ListVector} type.
*
* @param parent the parent (tuple, union or list) that holds this list
* @param columnSchema description of the list (must be empty of
* subtypes)
* @return the column state for the list
*/
private ColumnState buildUnionList(ContainerState parent, ColumnMetadata columnSchema) {
// The variant must start out empty.
assert columnSchema.variantSchema().size() == 0;
// Create the union writer, bound to an empty list shim.
final UnionWriterImpl unionWriter = new UnionWriterImpl(columnSchema);
unionWriter.bindShim(new EmptyListShim());
final VariantObjectWriter unionObjWriter = new VariantObjectWriter(unionWriter);
// Create the list vector. Starts with the default (dummy) data
// vector which corresponds to the empty union shim above.
// Don't get the list vector from the vector cache. List vectors may
// have content that varies from batch to batch. Only the leaf
// vectors can be cached.
final ListVector listVector = new ListVector(columnSchema.schema(), parent.loader().allocator(), null);
// Create the list vector state that tracks the list vector lifecycle.
final ListVectorState vectorState = new ListVectorState(unionWriter, listVector);
// Create the list writer: an array of unions.
final AbstractObjectWriter listWriter = new ArrayObjectWriter(new ListWriterImpl(columnSchema, listVector, unionObjWriter));
// Create the manager for the columns within the list (which may or
// may not be grouped into a union.)
final ListState listState = new ListState(parent.loader(), parent.vectorCache().childCache(columnSchema.name()));
// Bind the union state to the union writer to handle column additions.
unionWriter.bindListener(listState);
// Assemble it all into a union column state.
return new UnionColumnState(parent.loader(), listWriter, vectorState, listState);
}
use of org.apache.drill.exec.vector.accessor.writer.UnionWriterImpl.VariantObjectWriter in project drill by apache.
the class ColumnBuilder method buildUnion.
/**
* Builds a union column.
* <p>
* The union vector type is not well supported in Drill. The idea is that
* arbitrary operators can absorb schema changes by converting vectors to
* unions so that an operator can handle, say, a nullable int and a varchar.
* In practice, most operators don't support this feature. (Sort does -- but
* does not manage memory for the union case.) In principal, union can't solve
* the problem because ODBC and JDBC don't support unions, and it is easy
* to envision changes that unions won't solve (int and varchar types combining
* in a join column, say.) Still, Drill supports unions, so the code here
* does so. Unions are fully tested in the row set writer mechanism.
*
* @param parent container of vectors
* @param columnSchema implied projection type for the column
* @return column
*/
private ColumnState buildUnion(ContainerState parent, ColumnMetadata columnSchema) {
assert columnSchema.isVariant() && !columnSchema.isArray();
// vectors can be cached.
assert columnSchema.variantSchema().size() == 0;
final UnionVector vector = new UnionVector(columnSchema.schema(), parent.loader().allocator(), null);
// Then the union writer.
final UnionWriterImpl unionWriter = new UnionWriterImpl(columnSchema, vector, null);
final VariantObjectWriter writer = new VariantObjectWriter(unionWriter);
// The union vector state which manages the types vector.
final UnionVectorState vectorState = new UnionVectorState(vector, unionWriter);
// Create the manager for the columns within the union.
final UnionState unionState = new UnionState(parent.loader(), parent.vectorCache().childCache(columnSchema.name()));
// Bind the union state to the union writer to handle column additions.
unionWriter.bindListener(unionState);
// Assemble it all into a union column state.
return new UnionColumnState(parent.loader(), writer, vectorState, unionState);
}
Aggregations