aokolnychyi commented on a change in pull request #315: [WIP] Incremental
processing prototype
URL: https://github.com/apache/incubator-iceberg/pull/315#discussion_r347104255
##########
File path: api/src/main/java/org/apache/iceberg/Table.java
##########
@@ -44,6 +44,15 @@
*/
TableScan newScan();
+
+ /**
+ * @param fromSnapshotId - the last snapshot id read by the user, exclusive
+ * @param toSnapshotId - read incremental data upto this snapshot id
+ * @return a table scan which can read incremental data from {@param
fromSnapshotId}
+ * exclusive and up to {@toSnapshotId} inclusive
+ */
+ TableScan newIncrementalScan(long fromSnapshotId, long toSnapshotId);
Review comment:
Splitting a snapshot into multiple batches will require some thinking.
Maybe, this can be done by streaming sources. For example, our implementation
of `Offset` in Spark can store additional information. Storing consumed files
in offset JSON files isn't an option but maybe we can store a pointer to a list
of consumed manifests or something (which can be an Avro or a Parquet file, for
example).
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
