com.pervasive.datarush.operators.io

Interface SplitParsingContext

public interface SplitParsingContext

An object representing the context of a data split parsing operation. Parsers use the context to:

• Get the split to be parsed
• Publish parsed records
• Handle parsing errors which may arise

Method Summary

Methods

Modifier and Type	Method and Description
void	bulkPublish(RecordTokenSequence records) Provides a set of records to publish.
void	discardRecord() Signals that the current record should be ignored.
DataSplit	getSplit() Gets the split being parsed.
void	handleExtraField(String message) Reports an extra field being found in a record.
void	handleFieldError(String message) Reports a field parsing error.
void	handleMissingFields(String message) Reports a record being found to having missing fields.
void	handleParseException(long offsetInSplit, Exception e) Reports an exception occurring during parsing of the split.
void	publishRecord() Signals that the current record is ready to be published.
void	startRecord(long offsetInSplit) Establishes the context for the current record.

Method Detail

getSplit

DataSplit getSplit()

Gets the split being parsed.

Returns:

the currently parsed split

startRecord

void startRecord(long offsetInSplit)

Establishes the context for the current record. Errors messages and published records will be associated with this information.

Parameters:

offsetInSplit - the offset within the split at which the record begins. This offset should be either in bytes or characters as appropriate for the format.

publishRecord

void publishRecord()

Signals that the current record is ready to be published. Field values are set in the buffers provided to DataFormat.DataParser.bindOutput(RecordSettable).

discardRecord

void discardRecord()

Signals that the current record should be ignored. Field values in the buffers provided to DataFormat.DataParser.bindOutput(RecordSettable) should be discarded.

bulkPublish

void bulkPublish(RecordTokenSequence records)

Provides a set of records to publish.

This method is intended only for column oriented block formats which assemble multiple records at once. For row oriented formats, publishRecord() should be used instead.

Parameters:

records -

handleFieldError

void handleFieldError(String message)

Reports a field parsing error.

Parameters:

message - additional information about the error. The message is interpreted within the context of the current split, so this data need not be included.

handleMissingFields

void handleMissingFields(String message)

Reports a record being found to having missing fields.

Parameters:

message - additional information about the error. The message is interpreted within the context of the current split, so this data need not be included.

handleExtraField

void handleExtraField(String message)

Reports an extra field being found in a record.

Parameters:

message - additional information about the error. The message is interpreted within the context of the current split, so this data need not be included.

handleParseException

void handleParseException(long offsetInSplit,
                        Exception e)

Reports an exception occurring during parsing of the split. A parser should invoke this when an exception occurs within DataFormat.DataParser.parseSplit(SplitParsingContext).

Parameters:

offsetInSplit - the current offset, in bytes or characters, within the split when the error occurred. The appropriate units for the format should be used.

e - the exception that occurred