object HadoopRDDWriter
Linear Supertypes
Ordering
- Alphabetic
- By Inheritance
Inherited
- HadoopRDDWriter
- AnyRef
- Any
- Hide All
- Show All
Visibility
- Public
- All
Type Members
-
class
MultiMapWriter extends AnyRef
When record being written would exceed the block size of the current MapFile opens a new file to continue writing.
When record being written would exceed the block size of the current MapFile opens a new file to continue writing. This allows to split partition into block-sized chunks without foreknowledge of how big it is.
Value Members
-
final
val
DefaultIndexInterval: Int(4)
Index innterval at which map files should store an offset into sequence file.
Index innterval at which map files should store an offset into sequence file. This value is picked as a compromize between in-memory footprint and IO cost of retreiving a single record.
- def write[K, V](rdd: RDD[(K, V)], path: Path, keyIndex: KeyIndex[K], indexInterval: Int = DefaultIndexInterval, existenceCheck: Boolean = true)(implicit arg0: AvroRecordCodec[K], arg1: ClassTag[K], arg2: AvroRecordCodec[V], arg3: ClassTag[V]): Unit