chaoqin-li1123 commented on code in PR #45023: URL: https://github.com/apache/spark/pull/45023#discussion_r1491463836
########## python/pyspark/sql/datasource.py: ########## @@ -298,6 +320,117 @@ def read(self, partition: InputPartition) -> Iterator[Union[Tuple, Row]]: ... +class DataSourceStreamReader(ABC): + """ + A base class for streaming data source readers. Data source stream readers are responsible + for outputting data from a streaming data source. + + .. versionadded: 4.0.0 + """ + + def initialOffset(self) -> dict: + """ + Return the initial offset of the streaming data source. + A new streaming query starts reading data from the initial offset. + If Spark is restarting an existing query, it will restart from the check-pointed offset + rather than the initial one. + + Returns + ------- + dict + A dict whose key and values are str type. + """ + ... + raise PySparkNotImplementedError( + error_class="NOT_IMPLEMENTED", + message_parameters={"feature": "initialOffset"}, + ) + + def latestOffset(self) -> dict: + """ + Returns the most recent offset available. + + Returns + ------- + dict + A dict whose key and values are str type. + """ + ... + raise PySparkNotImplementedError( + error_class="NOT_IMPLEMENTED", + message_parameters={"feature": "latestOffset"}, + ) + + def partitions(self, start: dict, end: dict) -> Sequence[InputPartition]: + """ + Returns a list of InputPartition given the start and end offsets. Each InputPartition + represents a data split that can be processed by one Spark task. + + Parameters + ---------- + start : dict + The start offset of the microbatch to plan partitioning. + end : dict + The end offset of the microbatch to plan partitioning. + + Returns + ------- + Sequence[InputPartition] + A sequence of partitions for this data source. Each partition value + must be an instance of `InputPartition` or a subclass of it. + """ + ... + raise PySparkNotImplementedError( + error_class="NOT_IMPLEMENTED", + message_parameters={"feature": "partitions"}, + ) + + @abstractmethod + def read(self, partition) -> Iterator[Union[Tuple, Row]]: + """ + Generates data for a given partition and returns an iterator of tuples or rows. + + This method is invoked once per partition to read the data. Implementing + this method is required for stream reader. You can initialize any + non-serializable resources required for reading data from the data source + within this method. + This method is static and stateless. You shouldn't access mutable class member + or keep in memory state between different invocations of read(). + + Parameters + ---------- + partition : object + The partition to read. It must be one of the partition values returned by + ``partitions()``. + + Returns + ------- + Iterator[Tuple] or Iterator[Row] + An iterator of tuples or rows. Each tuple or row will be converted to a row + in the final DataFrame. + """ + ... + + def commit(self, end: dict): Review Comment: This function is not supposed to return anything, how do we express that in python? -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org