Defines an abstract class for accepting data and outputting optionally transformed or filtered data.
More...
#include <AbstractDataProcessor.qc.dox.h>
|
|
*hash< auto > | thread_local_data |
| | Thread-local data to set before running the processor.
|
| |
|
*list< string > | thread_local_data_keys |
| | Thread-local data keys from thread_local_data.
|
| |
Defines an abstract class for accepting data and outputting optionally transformed or filtered data.
◆ flush()
| DataProvider::AbstractDataProcessor::flush |
( |
code |
enqueue | ) |
|
This method is called for each pipeline element to permit flushing of any pending data at the end of processing.
- Parameters
-
| enqueue | a closure taking a single argument that enqueues the processed data for the next step in the pipeline; if no data should be processed onwards, do not call enqueue; if only one record should be processed onwards, then enqueue should be called only once; if multiple records are generated, then call it once for each generated record; prototype: code enqueue = sub (auto qdata) {}
|
- Note
- Calls flushImpl() on the data to do the actual processing
- Accept and return type information is not enforced in this method; it must be enforced in flushImpl()
◆ flushImpl()
| DataProvider::AbstractDataProcessor::flushImpl |
( |
code |
enqueue | ) |
|
|
private |
This method is called for each pipeline element to permit flushing of any pending data at the end of processing.
- Parameters
-
| enqueue | a closure taking a single argument that enqueues the processed data for the next step in the pipeline; if no data should be processed onwards, do not call enqueue; if only one record should be processed onwards, then enqueue should be called only once; if multiple records are generated, then call it once for each generated record; prototype: code enqueue = sub (auto qdata) {}
|
- Note
- The default implementation in the base class does nothing; to submit pending data for processing by subsequent elements in the pipeline when pipeline processing is finished, call enqueue with the pending data
- No data should be cached or pending after this call; if the object requires closing or purging of internal state to make it reusable, it should be performed in this method
◆ getExpectedType()
Returns the expected type of data to be submitted, if available.
- Returns
- the expected type of data to be submitted, if available
- Note
- Calls getExpectedTypeImpl() to provide the return value
◆ getExpectedTypeImpl()
Returns the expected type of data to be submitted, if available.
This base class method returns NOTHING; reimplement in subclasses to provide a type
◆ getReturnType()
Returns the type of data that will be returned, if available.
- Returns
- the type of data that will be returned, if available
- Note
- Calls getReturnTypeImpl() to provide the return value
◆ getReturnTypeImpl()
Returns the type of data that will be returned, if available.
This base class method returns NOTHING; reimplement in subclasses to provide a type
◆ submit()
| DataProvider::AbstractDataProcessor::submit |
( |
code |
enqueue, |
|
|
auto |
_data |
|
) |
| |
Submits the data for processing.
- Parameters
-
| enqueue | a closure taking a single argument that enqueues the processed data for the next step in the pipeline; if no data should be processed onwards, do not call enqueue; if only one record should be processed onwards, then enqueue should be called only once; if multiple records are generated from the input data, then call it once for each generated record; prototype: code enqueue = sub (auto qdata) {}
|
| _data | the data to process |
Sets thread-local data before running and clears on exit if any is set
- Note
- Calls submitImpl() on the data to do the actual processing
- Accept and return type information is not enforced in this method; it must be enforced in submitImpl()
◆ submitImpl()
| abstract DataProvider::AbstractDataProcessor::submitImpl |
( |
code |
enqueue, |
|
|
auto |
_data |
|
) |
| |
|
privatepure virtual |
Submits the data for processing.
- Parameters
-
| enqueue | a closure taking a single argument that enqueues the processed data for the next step in the pipeline; if no data should be processed onwards, do not call enqueue; if only one record should be processed onwards, then enqueue should be called only once; if multiple records are generated from the input data, then call it once for each generated record; prototype: code enqueue = sub (auto qdata) {}
|
| _data | the data to process |
◆ supportsBulkApi()
| bool DataProvider::AbstractDataProcessor::supportsBulkApi |
( |
| ) |
|
Returns True if the data processor supports bulk operation.
- Returns
- True if the data processor supports bulk operation
- Note
- Calls supportsBulkApiImpl() to return the answer
◆ supportsBulkApiImpl()
| abstract bool DataProvider::AbstractDataProcessor::supportsBulkApiImpl |
( |
| ) |
|
|
privatepure virtual |
