How do I get the name of the newest file via the T

Question

I'm trying to create a macro for Keyboard Maestro for OS X doing the following:

1. Get name of newest file in a directory on my disk based on date created;
2. Paste the text "newest file: " plus the name of the newest file.

One of its options is to "Execute a shell script", so I thought that would do it for 1. After Googling around a bit I came up with this:

cd /path/to/directory/
ls -t | head -n1

This sorts it right, and returns the first filename. However, it also seems to includes a line break, which I do not want. As for 2: I can output the text "newest file: " with a different action in the app, and paste the filename behind that. But I'm wondering if you can't return "random text" + the outcome of the ls command.

So my question is: can I do this only using the ls command? And how do I get just the name of the latest file without any linebreaks or returns?

Answer 1

cd /path/to/directory/
echo -n "random text goes here" $(ls -t | head -n1)

If you want, you can add more text on the end in the same way.

Answer 2

Since you're already using pipes, just throw another one in there:

ls -t | head -n1 |awk '{printf("newest file: %s",$0)}'

(Note that the "printf" does not include a '\n' at the end; that gets rid of the linebreak)

Edit:

With Arkku's suggestion to exit awk after the first line, it looks like:

ls -t | awk '{printf("newest file: %s",$0);exit}'

Answer 3

Since this comes up first when I search for "find newest file on mac shell" then I thought I would post for others to help...

If you are trying to find the file in just that directory ls is good, but I found that find works quite well, and you might learn something new in the process. I however had to get the GNU find on my mac in order to use the line of code that diimdeep posted above.

If you run Homebrew, which you should if you spend any time in terminal, or maybe one of the alternatives. You can run "brew install findutils" Once this is finished the code from above will work like this...

sudo /usr/local/opt/findutils/bin/gfind -type f -printf '%T+ %p\n' | sort -r | head -n 1

Just change the last bit to a 10 if you want more. gfine will take a path if you need to specify that as well. I usually cd into the folder that I want to start in though. Find is automatically recursive.

Answer 4

Using output of ls is bad practice.

find -type f -printf '%T+ %p\n' | sort -r | head -n 1

Answer 5

You can't do it with only ls. However, as echo is generally built into the shell, it doesn't really add any overhead into the script. To get just the name of the file, I'd suggest:

echo -n "newest: $(ls -t1 | head -n1)"

If, for some reason, you really want to eliminate the head, then I suppose you could go for something like:

ls -t1 | ( read n; echo -n "newest: $n")

(read is built into the shell, head isn't.)

Note that these solutions do not recurse into subdirectories, since that was not specified in the question. In fact, subdirectories may be printed as the newest "file".

Answer 6

You can do that in bash in a single statement like so:

echo -n "newest file: $(ls -t |head -n1)"

You can also remove that newline without echo:

ls -t |head -n1 |tr -d '\n'

Make sure ls doesn't output colors to non-tty streams (i.e. specify color by ls --color=never or ls --color=auto or not at all).

The ls solution will output files of any kind sorted by modification time. If you want only regular files or if you don't want directories then you can use find and xargs:

echo -n "newest file: $(find . -maxdepth 1 -type f -print0 |xargs -0 ls -t)"