it’s so confusing that the order changes when adding IDENTICAL strings to BOTH filenames. Is this really how it’s supposed to be?
Yes. The periods are just part of the name like any other letter, so 5 is compared to m, and numbers sort before letters. You can add something like ‘.0’ to make it sort more naturally. Look up an ASCII table to get a feeling for how strings are sorted.
What you expecting called natural sorting. Mac employed natural sorting back in 90s. What you get is legitimate Alphabetical sorting which used by Linux and Windows. Natural sorting parses tokens in the string and compares them. Alphabetical sorting compares two strings by comparing individual characters at same index (position). Alphabetical sorting is quite common as it simpler to implement (or rather harder to screw up) and yields predictable results
One of many libraries for Python which implements natural sorting https://github.com/SethMMorton/natsort
Humans order by strings, computers order by characters.
I believe it’s correct. If you sort say “A”, “AA”, “AAA” then you get
- A
- AA
- AAA
Because the first character is compared, which are all the same, then the second. The first one has no second character, so it comes first. The second has no third character, so it comes before the third item.
In your scenario, you have:
- 5
- 5.5
The first characters are the same, so it looks at the second character. Item 1 has no second character so it comes first.
Scenario 2:
- 5.5 A
- 5 A
The first character is the same, so it looks at the second character. The second characters are “.” and " ". The “.” comes first in the character ranking so is shown first.
yes yes I get what you’re saying but it’s still odd. Didn’t humans do this differently in the old analog days? I’m sure any human when working with a real paper archive in front of him, order 5 A before 5.5 A. Perhaps it has something to do with viewing 5 as 5.0 and 5.00, since they are mathematically equivalent, and come before 5.5. Although humans would also be inconsistent because they would order 5.9 before 5.11 if the context were to be chapters going from 5.9 -> 5.10 -> 5.11. But if these papers were to represent values, humans would order 5.9 AFTER 5.11. And computers obviously don’t make exceptions based on context like humans do.
edit: if I understand correctly, I’d be cleaner if spaces come first in character ranking of ANY character. Perhaps that’d make it more human readable.
Humans aren’t sorting this though. A computer is.
How should 5.2 and 5.12 be sorted? Numerically 5.12 is less than 5.2. But if it’s a version then it’s “five dot twelve” and thus 5.12 is greater.
These are contextual things that are very difficult for a computer to know. And trying to guess often just makes things weirder. So they often sort in a way that is at least consistent.
If a person was ordering them, they would do it in numerical order. Despite these being numbers, the computer is still ordering in alphabetical order.
Doing it the way a person would requires the file manager to understand context, which requires a lot more logic for arguably little benefit.
I note that your season and episode start with 0 as well (S01E05), in order to ensure the alphabetical ordering works. Perhaps you should use 5.0 to solve this in the same way.
Doing it the way a person would requires the file manager to understand context, which requires a lot more logic for arguably little benefit.
I’m so glad KDE Dolphin has a “natutal sorting” option. Not sure about this specific case, but I have never been surprised by the order with that setting.
Would be interesting to check the code behind it.
It’s an API call which emails a guy who just does it real fast by hand
I encounter this mostly with manga. (I’ll not rehash what others have said).
FWIW, and in that use case, I deal with it by renaming x5 to x5.0 so it will sort before x5.5. And then usually put both into an x5 directory and then zip that into a CBZ.
that all makes sense to me. how would you want it to work?
i’m not quite sure whether i understood your question but this seems to be right. the
5fromS01E05.*5*.mkvis higher in the alphabet thenmfromS01E05.*m*kvso it belongs above that entry.The correct solution is to make sure all files to be sorted have equivalent numerical structure, like 5.0 and 5.5
Same with eg. 05 and 10
Always been weird, this is why for instance my ll alias is:
alias ll='LC_COLLATE=C ls -alFh' alias ls='ls --color=auto'
