Sort order of alphanumeric file names in Finder

Question

Level 1

0 points

Sort order of alphanumeric file names in Finder

File Name Sorting
I often using dates for file names - for letters and for photos in particular, with a format something like yymmdd, e.g., today's date, December 27, would be 071227. Occasionally I may use 20071227. If I have the year and month but do not have a specific date, the file name might look like 0712xx. The sort sequence of files in the various Finder views does not seem to be what I would expect, based on standard computer alphanumeric sorting rules. Here is what I am finding - three files beginning with file names 0708xx, 070212 and 070521 will appear in the Finder list as follows:

0708xx
070212
070521

whereas I would expect the proper sort order to be:

070212
070521
0708xx

If I change "xx" to "00" then the sort order is strictly numerical, i.e.,

070212
070521
070800

No one can explain why the files sort per the first list above, if I use the "xx" convention for the date. Is there a Leopard bug, or is the sort order based on an algorithm I am unfamiliar with.

iMac, Mac OS X (10.5.1)

Posted on Dec 31, 2007 9:18 AM

Reply

Answer 1

Best reply

Camelot

Level 9

54,333 points

Dec 31, 2007 9:36 AM in response to new-ken

The Finder does have its own rules on sorting, but I can't reproduce what you describe, so I'm not sure if that's the issue or not.

The Finder implements a numeric sort for file names that begin with digits. This is counter to a purely alphabetic sort that you might expect?

What's the difference? Create a series of files named from 1 through 100.
In an alphabetic sort you'll see:

1
10
11
12
13
14
15
16
17
18
19
2
20
21
etc.

In other words, the file '2' appears after all the files that begin with '1'.

In comparison, the Finder will list this as:

1
2
3
4
5
6
7
8
9
10
11
12
etc.

In other words, the Finder considers numerical value over alphabetic characters.

Now, that said, your '0708' should sort after '0702' whether you're sorting numerically or alphabetically. As I said, I can't reproduce what you're seeing, so there must be some other force at work.

Reply

Answer 2

biovizier

Level 5

7,925 points

Dec 31, 2007 9:54 AM in response to Camelot

..." As I said, I can't reproduce what you're seeing, so there must be some other force at work."...

Could it be that there is in fact something atypical about your system? I had a thread in the unix section regarding folder permissions where you were unable to reproduce a behaviour that I was seeing on many machines, and which another user was subsequently able to confirm.

Here, my system behaves as the OPs - the number 708 is smaller than 70212 is smaller than 70521 so 0708xx sorts before 070212, which sorts before 070521. The numerical part of 0708xx is interpreted separately.

I consider it counter-intuitive and makes keyboard based file selection cumbersome compared to straight alphanumeric sorting, but this has been how OS X "Finder" has sorted, probably since 10.0.

Reply

Answer 3

Tom Gewecke

Level 10

115,910 points

Dec 31, 2007 10:20 AM in response to new-ken

No one can explain why the files sort per the first list above, if I use the "xx" convention for the date. Is there a Leopard bug, or is the sort order based on an algorithm I am unfamiliar with.

This is standard OS X file sorting behavior. Numerical strings are sorted by their arithmetic value. See this note:

http://discussions.apple.com/thread.jspa?threadID=121936&tstart=105

Reply

Answer 4

R C-R

Level 6

17,782 points

Dec 31, 2007 12:05 PM in response to Tom Gewecke

Tom's comprehensive article provides the answer but is perhaps a bit much for those looking for a simple, easy to remember rule. The brief one at File System Overview: Sorting Rules may be useful in this respect:

"Digit sub-strings are sorted by numeric value rather than as characters."

IOW, much like a human would naturally do, the Finder considers numeric expressions as potentially multi-character entities.

Except ... just like humans, it seems not to know what to do with leading zeros. Try this:

Create a new folder & add to it 4 sub-folders named as follows (all "0" characters are zeros):

2
2aa
02
02aa

Note that in list view, Leopard sorts them by name as follows:

02
2
02aa
2aa

But, try selecting them in order by hitting the tab key repeatedly, which should move you down the list one item after the other ... but doesn't.

Reply