hi fellas!!!!

i am trying to write a script in which the output should be named after its parent directory.

Ex.
i have 3 directory named as 2018-10-01, 2018-11-11, 2018-06-05 . In each directory having n number of files.

I run the command as below :
for D in */; do wc -l > response.txt; done

I got output in each folder but with same name ie response.txt

THE THING I WANT HERE response.txt SHOULD BE NAMED AFTER IT'S PARENT DIRECTORY'S NAME.

This looks like homework.
The current directory path is held in the PWD environment variable. The subdirectory can be obtained with the parameter expansion ${PWD##*/}.

Don't yell; all caps and bolding isn't pleasant. And you've been asking about scripting for two years now...your previous posting history indicates that. After two years, have you not yet learned how to define a variable? There are LOTS of easily-found bash scripting tutorials, and like allend said...this looks like homework.

Hmm...you say you are getting some output with that command??
Actually, what you should get is just a new empty response.txt file (only one and not in the different folders themselves) because the redirection is processed. But the command itself should return nothing as is...

1 members found this post helpful.

The loop variable D cycles thru the directories. Use $D inside the loop!
In your sample the wc -l seems to miss a filename.
The following uses $D to construct filenames for wc -l and an output file

hi used the command ==>> for D in /; do wc -l ${D}/ > ${D%/}Request;done which is working fine . But in each directory there are two file named as .CUD and .CUR , I want to read both files separately . Please suggest

How is it related to the original question? How do you want to read those files at all?

this can be tricky. D is just a var, it could just as well be G to get the same info.

depending on where your dir are in the tree, you have to adjust your code to get the dir you want then it is just as simple as creating a file with its same name then putting that file where ever you want it.

that is a 3 step process.
1. first develop your code until you get it returning the exact dir(s) you want to use.
2. then using that information, further develop your code to create a file by the same name.
3. then further develop your code to create the files somewhere else, ie where you tell them to be created.


this is doing exactly what you are telling it to do, it does not show you using that guild line for development of code.
put the output of wc -l into a file called 'response.txt' and to over write it every time that is ran.

to create a file try using 'touch'

1. they cannot be in the same dir, having the same name therefor what do you think needs to be done to find out where they are located?

2. you cannot by any means read two files at the same time. one then the other.

3. I do not even see how you are getting files in that code.
what is this suppose to do?

Why did you replace for D in */ with for D in /? That command was working properly... Do you want to list contents of root folder?
What are you trying to achieve with your wc -l? You know you cannot apply that command to a folder, right?

OP wants to trim the shortest substring ending with a "/" in $D (read from right to left)
It's OK for me as long as $D is a folder and so contains an ending "/", hence */ in the for loop...

I fingured that might be it, but I wanted to be sure, as you forgot the star *
going from right to left shortest distance to first /
chops off everything to the right of the / including the /

Yes, BW-userx, you're right
I should have mentioned that ${D%/} can only trim the ending one-character "/".
If OP wants to trim the shortest substring ending with a "/" in $D (read from right to left), (s)he must rather write as you did: ${D%/*}
But I'm not sure OP needs that

And just to clarify for OP:
has nothing to do with the root folder. It's just a one-run loop with one-element, namely "/"
If OP wants to iterate through root folder, (s)he needs:

As was mentioned, the trailing / should be chopped
The /* does not include /.[!.]* filenames.
But instead of
it is perhaps better to use find