Impala
Impalaistheopensource,nativeanalyticdatabaseforApacheHadoop.
|
Public Attributes | |
ScopedAvroSchemaT | schema |
The record field schema from the file. More... | |
std::vector< SchemaElement > | children |
Complex types, e.g. records, may have nested child types. More... | |
int | null_union_position |
const SlotDescriptor * | slot_desc |
Internal representation of a field/column schema. The schema for a data file is a record with a field for each column. For a given file, we create a SchemaElement for each field of the record.
Definition at line 126 of file hdfs-avro-scanner.h.
std::vector<SchemaElement> impala::HdfsAvroScanner::SchemaElement::children |
Complex types, e.g. records, may have nested child types.
Definition at line 131 of file hdfs-avro-scanner.h.
int impala::HdfsAvroScanner::SchemaElement::null_union_position |
Avro supports nullable types via unions of the form [<type>, "null"]. We special case nullable primitives by storing which position "null" occupies in the union and setting 'schema' to the non-null type's schema, rather than the union schema. null_union_position is set to 0 or 1 accordingly if this type is a union between a primitive type and "null", and -1 otherwise.
Definition at line 138 of file hdfs-avro-scanner.h.
Referenced by impala::HdfsAvroScanner::ConvertSchema(), and impala::HdfsAvroScanner::MaterializeTuple().
ScopedAvroSchemaT impala::HdfsAvroScanner::SchemaElement::schema |
The record field schema from the file.
Definition at line 128 of file hdfs-avro-scanner.h.
Referenced by impala::HdfsAvroScanner::ConvertSchema(), impala::HdfsAvroScanner::MaterializeTuple(), and impala::HdfsAvroScanner::ResolveSchemas().
const SlotDescriptor* impala::HdfsAvroScanner::SchemaElement::slot_desc |
The slot descriptor corresponding to this element. NULL if this element does not correspond to a materialized column.
Definition at line 142 of file hdfs-avro-scanner.h.
Referenced by impala::HdfsAvroScanner::MaterializeTuple(), and impala::HdfsAvroScanner::ResolveSchemas().