awk: find the oldest and newest files by date

Question

I'm using an awk script to print the oldest and newest files created based on the date (the 6th 7th and 8th field).
my sample list, date in the format (YYYY/MM/DD):

file was created 2020/10/10 20:18:42 its name is output1
file was created on 2020/09/10 12:13:22 its name is foobar.awk
file was created on 2020/10/10 20:12:43 its name is output2
file was created on 2020/12/10 18:11:38 its name is foobar.bash
file was created on 2020/12/10 22:32:13 its name is output.txt

what it should look like:

Oldest file date :
2020/09/10 12:13:22
file name: foobar.awk

Newest file date :
2020/12/10 22:32:13
file name: output,txt

Note: I can only use awk I only want the oldest and newest files to be printed.
Also: I am NOT sorting my files, just simply printing the oldest and newest files.

This will never work well. You can't depend on the format of `ls` and it is generally a bad idea to parse the output of `ls`. See https://mywiki.wooledge.org/ParsingLs and [Why \*not\* parse \`ls\` (and what to do instead)?](https://unix.stackexchange.com/q/128985). Would you be open to other solutions, using something like `stat` instead of `ls`? — terdon, Oct 11 '20 at 21:33
I've edited to the format to be used with the modified `ls -la --time-style` — bull, Oct 11 '20 at 22:02
This really won't work. Just make a file with a space in its name and try running your script. You really want to look at `stat` instead of `ls` for this. — terdon, Oct 11 '20 at 22:09
Likewise, since you "know about" `ls -t`, is there any reason you couldn't just do something like `ls -tr | awk 'NR==1{print} END{print}'`? — steeldriver, Oct 11 '20 at 22:37
dont worry about piping in `ls` anymore, I made changes to the file format so that it can act like a regular text file. — bull, Oct 11 '20 at 23:08

score 0 · Answer 1 · answered Oct 12 '20 at 00:33

Assuming GNU awk:

gawk '
    {created[$NF] = $5 " " $6}
    END {
        PROCINFO["sorted_in"] = "@val_str_asc"
        for (f in created) {print "oldest:", f, created[f]; break}
        PROCINFO["sorted_in"] = "@val_str_desc"
        for (f in created) {print "newest:", f, created[f]; break}
    }
' file

Ref: Controlling Array Traversal, and Using Predefined Array Scanning Orders with gawk

score 0 · Answer 2 · answered Oct 12 '20 at 04:06

Using stat to get the files and their modification times, awk to get min/max without sorting and format the output.

stat -c "%Y %n" * | awk '
    min == "" || min > $1 {min = $1; fmin = substr($0, index($0, FS)+1)}
    max == "" || max < $1 {max = $1; fmax = substr($0, index($0, FS)+1)}
    END {
        form = "%Y/%m/%d %H:%M:%S"
        printf "Oldest\nDate: %s\nFile: %s\n", strftime(form, min), fmin
        printf "Newest\nDate: %s\nFile: %s\n", strftime(form, max), fmax
    }'

Sample output:

Oldest
Date: 2020/08/29 01:52:23
File: test.java
Newest
Date: 2020/10/12 06:52:21
File: test.sh

Notes:

The above handles whitespaces into files, not newlines.
I think awk strftime is GNU only

awk: find the oldest and newest files by date

2 Answers2