On 04/12/2017 10:48 AM, Yair Altman wrote:
...
> Categorical data and tables in general use more memory and are less
> performant than the corresponding implementation using simple arrays. So
> the actual tradeoff here is not memory vs. performance, but rather
> maintainability and development time/cost vs. performance.
> Yair Altman http://UndocumentedMatlab.com Author: "Accelerating MATLAB
> Performance" (CRC Press, 2014)
Yabbut... :)
If performance is _too_ laggardly, the net result may be that one has to
end up throwing out the feature and reverting to the lower-level
implementation. It depends on what's the more important in any given
situation; RAD and exploration vis a vis a resulting more
production-like use of an algorithm/analysis or the need to handle
datasets of a size that the feature while making writing the analysis
much easier just turns out to not be practical for end use.
--
...
> Categorical data and tables in general use more memory and are less
> performant than the corresponding implementation using simple arrays. So
> the actual tradeoff here is not memory vs. performance, but rather
> maintainability and development time/cost vs. performance.
> Yair Altman http://UndocumentedMatlab.com Author: "Accelerating MATLAB
> Performance" (CRC Press, 2014)
Yabbut... :)
If performance is _too_ laggardly, the net result may be that one has to
end up throwing out the feature and reverting to the lower-level
implementation. It depends on what's the more important in any given
situation; RAD and exploration vis a vis a resulting more
production-like use of an algorithm/analysis or the need to handle
datasets of a size that the feature while making writing the analysis
much easier just turns out to not be practical for end use.
--
