Sorting algorithms are one of the most studied in computer science.
It is extremely hard to invent something new and yet every year there are new papers published about the subject. There are several well-known sorting algorithms: Insertion, Selection, Bubble, Shell, Merge, Heap and Quicksort. More information about sorting algorithms:
http://www.sorting-algorithms.com/
At though it is all nice and clear there is just one problem: the compiler does not know anything about sorting.
So all the code generated must be carefully reviewed and optimized.
Our developers have just completed data sorting optimisation and the result is quite positive:
Version 5.8.0.18 is 2 times faster than version 5.8.0.12 and 12 times than version 5.8.0.6
On top of that, it sorts data faster than Pentaho Kettle and it should be possible to make it even faster
Testing results:
Version | Records per second | Time Taken |
---|---|---|
5.8.0.18 | 75,222 | 39 sec |
5.8.0.12 | 32,832 | 1 min 31 sec |
5.8.0.6 | 5,451 | 7 min 45 sec |
5.8.0.18 |
5.8.0.125.8.0.6 |
Sorting data using Pentaho Kettle |
Notes:
- The performance also depends on the hardware configuration
- A bigger sorting buffer improves performance
- Please use our support forum to provide us with feedback