If I calculate the average file size of the list, as it is, the average may be too high or too low for the target average I'm seeking. What I am trying to do is generate a new list that will match (or be within ~500 bytes) of a target average file size.size filename ----- -------- 4329 file1 12311 file2 ... 657 file100000
One idea I had was to generate buckets of files at a resolution of 1024 bytes and then create the new list by plucking files from there, and chucking some from the new list to meet the efficiency goal (see below). I would also like to maintain some randomness to the file sizes. Don't want to have clumping near the target average or something like that. Fast would be good, but efficient is more important. i.e. efficient = maintain as many of the elements of the original as possible../prog 25000 file.lst > new.25000.lst
In reply to Creating a sub-list constrained by an average while maintaining maximum number of original elements by Jack B. Nymbol
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |