Multi-streams analytics
View Patent ↗Multi-stream analytics is disclosed. An example method of multi-stream analytics with a query engine includes punctuating unbounded streaming data into data chunks, each of the data chunks representing a bounded data set in the unbounded streaming data. The method also includes processing one of the data chunks. The method also includes rewinding a query instance for processing another of the data chunks.
1. A method of multi-stream analytics with a query engine, comprising:
punctuating unbounded streaming data into data chunks, each of the data chunks representing a bounded data set in the unbounded streaming data;
processing one of the data chunks,
cutting query execution based on cycle specification; and
rewinding a state of a query instance without shutting down the query instance between execution cycles for processing another of the data chunks in a next execution cycle.
2. The method of claim 1 , further comprising executing a sequence of query cycles on a cycle-by-cycle basis on a sequence of the data chunks in a chunk-by-chunk manner.
3. The method of claim 1 , further comprising joining multiple streams and self-joining a single stream using a data chunk-based window or sliding window, wherein self-joining, joins a most recent chunk and a predetermined past chunk.
4. The method of claim 1 , further comprising joining multiple input streams by buffering consecutive data chunks in a sliding window.
5. The method of claim 1 , further comprising joining a stream window and a static table.
6. The method of claim 1 , further comprising window-joining multiple input streams.
7. The method of claim 1 , further comprising joining streams in sliding windows.
8. The method of claim 1 , further comprising self-joining a stream in a sliding window.
9. The method of claim 1 , wherein the query uses structured query language (SQL).
10. The method of claim 1 , further comprising when an end-of-cycle event is signaled from a stream capture function (SCF), the query engine completes a current query execution cycle and then rewinds the query instance for the next execution cycle.
11. The method of claim 10 , wherein the end-of-cycle event is determined when a first stream element belonging to the next execution cycle is received.
12. The method of claim 11 , wherein the query instance remains active even after the end-of-cycle event.
13. The method of claim 12 , further comprising discarding intermediate results of SQL operators associated with a current chunk of data, and maintaining application context at the end-of-cycle event.
14. The method of claim 1 , further comprising maintaining continuity of the query instance for retaining the buffered data for history-sensitive applications.
15. A system haying a query engine executing program code for multi-stream analytics, the program code stored on a non-transient computer readable medium and executed to:
punctuate unbounded streaming data into data chunks representing bounded data sets;
cutting query execution based on a cut point specified in cycle specification; and
execute a query instance on one of the data chunks, and then rewind a state of the query instance without shutting down the query instance between execution cycles for processing another of the data chunks in a next cycle; and
execute a sequence of query cycles on a sequence of the data chunks.
16. The system of claim 15 , wherein the program code is further executed to join multiple streams and self-join a single stream using a data chunk-based window or sliding window.
17. The system of claim 15 , wherein the program code is further executed to join multiple input streams.
18. The system of claim 15 , wherein the program code is further executed to join a stream window and a static table.
19. The system of claim 15 , wherein the program code is further executed to join streams in sliding windows.
20. The system of claim 15 , wherein the program code is further executed to self-join a stream in a sliding window.