Oh yeah, I can see how the different filtering spots would make a difference. Was just a bit surprised at the overhead of adding a command to the pipeline. A s

Oh yeah, I can see how the different filtering spots would make a difference. Was just a bit surprised at the overhead of adding a command to the pipeline. A similar setup on linux would be like the following I guess (on a folder with 4600 files, 170 matching files). 20ms time ls *.pdf 35ms time find -maxdepth 1 -iname '*.pdf' 60ms time ls | egrep '.pdf$' I was more wondering if adding each additional command added so much overhead. Your numbers looks much more reasonable. Maybe the article I read had a big step due to filesystem / caching or something.