How efficiently are the complex types physically stored (encoding, storage footprint, CPU/memory consumption when working with them)?
Question asked by an attendee during Putting Complex Data Types to Work. @DeepakMajeti @rroelke
Vague question... but basically, as efficient as a deconstructed set of primitive types, if you use the new CT stuff (So... extremely efficient!). Sometimes inefficient, if you use unoptimized flex maps.
FYI, one-dimensional arrays are the only complex types that can currently be stored in ROS data; everything else is external tables.
What encoding can be used in case of the new CT stuff?
The main thing to consider is the CPU/memory consumption when working with them. The CPU time consumed when dealing with complex type will depends on the number of values of that type in your document. The more complex data types you use, the more CPU time will be consumed. Hightechreviewer The more values of that type in your document, the more memory space will be consumed. So, if your document contains millions of complex type values, you should be extremely careful about using them. You can get an estimate of the amount of CPU time and memory space of your document using the Numbers. Nonstoptechno Numbers.getN() method.
Can't find what you're looking for? Search the Vertica Documentation, Knowledge Base, or Blog for more information.