It doesn't "decorate" the original records (thus saving memory) and what would be the "undecorate" step is invariant and so doesn't have to be coded each time. It also allows one to sort indices which can be quite useful, such as when one has multiple lists to be sorted in sync.
I see now that one could view it as applying GRT to the list of indices. Though, using pack on the indices adds some minor advantages (especially if you use an old Perl where sort isn't stable).
I don't really care to split hairs about whether any particular person considers my technique to be something that they'd call "the GRT", but I have never seen anybody else suggest my technique under any name.
I can see my technique as combining three techniques, one of which is GRT (though that wasn't how I ever thought about it before now). But the selection of the three techniques and the details of precisely how to combine them and the fact that that combination has several important advantages (over any expression of GRT that I've seen) makes me consider this as something worth categorizing separately. Certainly, searching for "GRT" isn't enough to ensure somebody discovers this [combination of] technique[s] and the resulting advantages, IME.
- tye
In reply to Re^3: perl ST sort performance issue for large file? (>GRT)
by tye
in thread perl ST sort performance issue for large file?
by rkshyam
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |