co.cask.cdap.data.file

Class ReadFilter

java.lang.Object

co.cask.cdap.data.file.ReadFilter

Direct Known Subclasses:

AndReadFilter, TimeRangeReadFilter, TTLReadFilter

public abstract class ReadFilter
extends Object

Filter for reading from FileReader. This class is experimental is still expanding.

See StreamDataFileWriter for the file format.

The methods on this class will be called in the following sequence:

  reset() - Called before filtering a new data block
  acceptTimestamp() - Called when the timestamp of a data block is read
  getNextTimestampHint() - Called if acceptTimestamp() return false
  acceptOffset() - Called at each stream event begin offset
 

Field Summary

Fields

Modifier and Type	Field and Description
static ReadFilter	ALWAYS_ACCEPT Always accept what it sees.
static ReadFilter	ALWAYS_REJECT_OFFSET Always reject offset.

Constructor Summary

Constructors

Constructor and Description
ReadFilter()

Method Summary

Methods

Modifier and Type	Method and Description
boolean	acceptOffset(long offset) Accept or reject based on file offset.
boolean	acceptTimestamp(long timestamp) Accept or reject based on event timestamp.
long	getNextTimestampHint() If the acceptTimestamp(long) returns false, the filter can provide a hint for the next timestamp the reader should start reading from.
void	reset() Invoked before filtering a new data block.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

ALWAYS_ACCEPT

public static final ReadFilter ALWAYS_ACCEPT

Always accept what it sees.

ALWAYS_REJECT_OFFSET

public static final ReadFilter ALWAYS_REJECT_OFFSET

Always reject offset.

Constructor Detail

ReadFilter

public ReadFilter()

Method Detail

reset

public void reset()

Invoked before filtering a new data block.

acceptOffset

public boolean acceptOffset(long offset)

Accept or reject based on file offset.

Parameters:

offset - The file offset.

Returns:

true to accept, false to reject.

acceptTimestamp

public boolean acceptTimestamp(long timestamp)

Accept or reject based on event timestamp.

Parameters:

timestamp - The timestamp of the event.

Returns:

true to accept, false to reject.

getNextTimestampHint

public long getNextTimestampHint()

If the acceptTimestamp(long) returns false, the filter can provide a hint for the next timestamp the reader should start reading from.

Returns:

the timestamp for the next event should be or a negative value if no hint is provided.