The new Adaptive ELT feature in Ajilius 3 uses intelligent algorithm selection at runtime to optimise ELT performance.
Combined with the new SQL+Scripts architecture for code generation and execution, we’re seeing orders of magnitude improvement in some use cases.
Here’s a sample of batch output from the standard Ajilius demonstration:
The first two jobs – stage_location and dim_location – used to take around 0.06 and 0.08 seconds to run. Now they’re talking less than .001 seconds.
They’re trivially small tables, but we’re using them as an example because you can replicate the experience in an Ajilius demonstration.
Not every job will benefit to this extent, it depends on factors like the algorithm, and the number of rows in the source and target tables, but every job that processes different data on each run will be faster.