How to find the file with the less rows in directory

NewGuy95 · 10-17-2016, 06:43 AM

Code:

#!/bin/bash

rows=""
name=""
for f in $1
do
	[ -f "$f" ] || continue
	cnt='wc -l < "$f"'
	if [ -z "$rows" ] || [ $cnt -lt $rows ]
		then
		rows=$cnt
		name=$f
	fi
done
echo "$file has $rows

Turbocapitalist · 10-17-2016, 06:55 AM

Thanks. There are three hard to see typos. The first is that single quotes ( ' ) were used instead of backticks ( ` ) That is one reason why it is a good idea to do command substitution with parenthesis instead. The second is the wrong variable name in the final line and missing a closing double quote.

Code:

#!/bin/bash

rows=""
name=""
for f in $1
do
	[ -f "$f" ] || continue
	cnt=$(wc -l < "$f")
	if [ -z "$rows" ] || [ $cnt -lt $rows ]
		then
		rows=$cnt
		name=$f
	fi
done
echo "$name has $rows"

Note that the script will try to look at all files in the given directory, even if they are subdirectories or binaries.

allend · 10-17-2016, 07:29 AM

Quote:

Name of directory will be passed as argument, also when there will be no arguments than script should return error (like echo "You forgot arument").

and

Quote:

I don't know where to put directory name?

For this you want to pass a positional parameter.
If you call your script and supply the directory name e.g. scriptname.sh </your/directory/name>
then I suggest a construction like:

Code:

if [ -d "$1"] ; then
  cd "$1"
else
  echo "Invalid directory name - exiting. Usage: ${0##*/} <directory>"
  exit(1)
fi

The '${0##*/}' is a bash parameter expansion that finesses the $0 positional parameter to leave just the script name.

NewGuy95 · 10-17-2016, 08:44 AM

Quote:

Originally Posted by allend

and

For this you want to pass a positional parameter.
If you call your script and supply the directory name e.g. scriptname.sh </your/directory/name>
then I suggest a construction like:

Code:

if [ -d "$1"] ; then
  cd "$1"
else
  echo "Invalid directory name - exiting. Usage: ${0##*/} <directory>"
  exit(1)
fi

The '${0##*/}' is a bash parameter expansion that finesses the $0 positional parameter to leave just the script name.

Can you be a little bit more specific where to put that code?
After

Code:

 for f in $1
do ....

or where?

Thank you all for help!

Turbocapitalist · 10-17-2016, 08:59 AM

The square brace is another way of writing "test", which is built-in with "bash"

It checks the existence of the directory you are going to be looking for. The manual page for "bash" is daunting. Really. But over time you'll learn to navigate it. See if you can hunt for "test" and its option -d in there:

Code:

man bash

It's under "conditional expressions"
So you'll need that test in your script before you try to use the directory.

Try it in a separate script:

Code:

#!/bin/bash

if [ -d "$1"] ; then
  cd "$1"
else
  echo "Invalid directory name - exiting. Usage: ${0##*/} directory"
  exit(1)
fi

exit(0)

Edit: fixed shell

NewGuy95 · 10-17-2016, 09:14 AM

I get errors at

Code:

 exit(1)

line

But if I try to paste those if statements inside my code it is working with directory, but without directory it displays me error like 20 times in a row.

Code:

#!/bin/bash

rows=""
name=""
for f in $1/*
do
        if [ -d "$1" ] ; then
	[ -f "$f" ] || continue
	cnt=$(wc -l < "$f")
	if [ -z "$rows" ] || [ $cnt -lt $rows ]
		then
		rows=$cnt
		name=$f
	fi
        else
           echo "Invalid directory name - exiting. Usage: ${0##*/} <directory>"
        fi
done
echo "$name has $rows"

Turbocapitalist · 10-17-2016, 09:22 AM

The for and done mark a loop. Stuff inside that will be repeated for each file found. Try moving the test before it, since you will need to test the directory.

PS. I fixed the shell in the script above, since we're sticking with "bash" for this task.

NewGuy95 · 10-17-2016, 09:30 AM

Code:

#!/bin/bash

rows=""
name=""
if [ -d "$1" ] ; then
   cd "$1"
else
   echo "Invalid directory name - exiting. Usage: ${0##*/} <directory>"
fi
for f in $1/*
do
       
	[ -f "$f" ] || continue
	cnt=$(wc -l < "$f")
	if [ -z "$rows" ] || [ $cnt -lt $rows ]
		then
		rows=$cnt
		name=$f
	fi
done
echo "$name has $rows"

If I try this way, than I again got

Code:

     has    rows

in other way, when I should get error I got

Code:

Row 13: /vmlinuz: Permission denied
Row 14: [: -lt: Expecting simple operator (Idk translation again)
/vmlinuz.old has 24188 rows.

Turbocapitalist · 10-17-2016, 09:36 AM

Where did the exit 1 go and

why is it needed?

Also, a test is needed to make sure that $1 is not empty.

NewGuy95 · 10-17-2016, 09:42 AM

Like I told you before I got error if I use exit(1) in row 9 (where exit is) and it wont let me any further.
Exit (1) (usually) indicates unsucessful termination if I got it right

grail · 10-17-2016, 09:44 AM

I'm kind of curious at what point the OP is going to do any work? So far he/she has been spoon fed every line of code and as far as I can see done zero to work anything out except ask another question
on how to fix code that is clearly not understood.

Maybe the OP would be better served by searching for the solutions instead of asking for a hand out. I would add that running a script on your system when you have no idea what it is doing is fraught with
danger.

Turbocapitalist has provided the 'wooledge' website which if processed would cover off many of the questions being asked. You have also been told repeatedly that operating on non-text files (see your
latest error messages) will cause issues.

NewGuy95 · 10-17-2016, 09:48 AM

Dear grail, thanks for your response.
As an non-english speaker I find it more simple if someone can explain me my problem in a way I will understand what is going on. As a totaly newbie (like few days) I don't understand all terms in literature that has been given to me.

Thats why I am asking experts like you guys are for some help. I you find it too hard to help, than please avoid this topic

grail · 10-17-2016, 10:19 AM

I have no issue with helping and am happy to see you take your first steps in this arena, but currently I have not seen any feedback from you that was not written by someone else and you will
not learn by regurgitating someone else's code / issues.

The comments you have made show that you are not completely understanding the first 'if' clause and why it requires the exit command. You should get to grips with this first and understand
what entries make you hit either part of the 'if' clause and what you need to enter at the command line to progress correctly past this structure.

Once you have understood the above it may then be prudent to walk your self the same way through each part of the code inside the 'for loop' as again, the questions you are coming up against are from a lack of understanding which others will fix but you will become no better at coding in bash without understanding it.

I would reiterate my earlier information too in that you really need to click on every link and work your way through the entire 'wooledge' site as the information is invaluable. I know this will be daunting as the sire is only in English, but the concepts are sound and will help you should you learn what is there

Good luck ... and I hope you keep at it

szboardstretcher · 10-17-2016, 10:33 AM

Does it have to be a script? This is really a one-liner kind of thing:

Code:

find /root/ -type f -exec wc -l {} \; | sort -g > sorted_list; clear; tail -n 1 sorted_list

If you want to break that out into a whole script, with parameters:

Code:

#!/bin/bash
if [ -d "$1" ] ; then
find $1 -type f -exec wc -l {} \; | sort -g > /tmp/sorted_list; clear; tail -n 1 /tmp/sorted_list   
else
echo "USAGE: ./sortbylines.sh [path to directory]"
echo ""
fi

Sefyir · 10-17-2016, 02:37 PM

Quote:

What if there are other text files like .odt ?
...
OpenDocument Format files have text in them but it is buried in XML which is in turn archived in a ZIP archive, so it's not a straight up task to read the number of lines in them.

Once you learn a lot of the terminal commands to manage plain text, a frequent task is figuring out to convert those "complex" files like xml into a more readable plain text.
You will lose a lot / all formatting data but this is typically not needed (finding number of lines, grepping for strings, etc)

In this case, odt2txt can be used to convert odt files to text files
https://stackoverflow.com/questions/...s-to-txt-files