Is it possible make the shell read functions 1 by 1 and calling an other function?

Shell Programming and Scripting
1

Greetings,

I m wondering if it's possible do do the following :

I have a simple function called "FindMoveDelete" which does the following :

FindMoveDelete()
{
find . -iname "$FILENAME*.ext" -exec mv {} "$PATH/$VAR" \; &&
find . -maxdepth 1 -type d -iname "$FILENAME*" -exec rm -rf {} \;
}

So basicly it moves the variable "$FILENAME*.ext" to "$PATH/$VAR" and if the move succeeds, deletes the previous of said file directory. (which is basicly the same name)

Now, that's working but what I want to do next is load a configuration file containing a long list of functions. Each functions declaring new variables for $FILENAME $PATH and $VAR as follow :

Function1()
{
FILENAME=function1
PATH="/volume1/my/path/"
VAR=myvar
}

That way when I want to modify or add new functions I simply have to edit my configuration file with all the functions...

Only problem is : How do I make the shell read function 1 then apply the MoveAndDelete then read function 2, apply MoveAndDelete.

I guess i could call the MoveAndDelete function within each function like so :

Function1()
{
FILENAME=function1
PATH="/volume1/my/path/"
VAR=myvar
MoveAndDelete
}

But it looks ugly :smiley:

Thanks for taking the time to read

2

Instead having 9,999 functions which do the same thing with slightly different values, just have your function take parameters.

Never use PATH as a variable. That's a special variable, changing its value will mess up tons of things.

That code couldn't possibly work. '$FILENAME' in single quotes won't expand. 'FILENAME*' won't expand properly either.

Loading a file is easy:

. configfile

Why do you have -type d in one and not in the other?

function FindMoveDelete
{
        find . -iname "${1}*.ext' -exec mv {} "${2}/${3}" \; &&
        find . -type d -iname 'FILENAME*' -exec rm -rf {} \;
}

FindMoveDelete FILENAME PATH DEST
3

"Instead having 9,999 functions which do the same thing with slightly different values, just have your function take parameters."

  • Well it's a good point and i would do that If I only wanted to do this action on only 1 particular kind of file but having "9.999 functions" allows me to run them all at once. If i don't want to use one I simply comment it. If I have to modify the subdirectory ($VAR) I simply edit the function.

Good point for the $PATH tho hum.. *hides :slight_smile: switched it to $DIR.

"That code couldn't possibly work. '$FILENAME' in single quotes won't expand. 'FILENAME*' won't expand properly either."

  • Indeed it's under double quotes in my script, wrote it back while writing this post, editing previous post..

"Why do you have -type d in one and not in the other?"

  • To find my "$FILENAME*" directory and delete it. The name of the directory is basicly the same as the filename. It's set in each function.
4

You already found and moved your FILENAME directory. The only time you'd ever delete it is when mv didn't work :wall:

If you'd tell me exactly what that function takes as inputs and exactly what it's supposed to do, I can probably find a better way.

How about this -- make a big list of things you want moved/whatever, then have your program handle this list.

while IFS="," read FILENAME DIR VAR
do
        # Skip comments
        [ "${FILENAME:0:1}" = "#" ] && continue
        # skip blank lines
        [ -z "$FILENAME" ] && continue

        function "$FILENAME" "$DIR" "$VAR"
done < ~/.thingstomove

# ~/.thingstomove
# one thing per line
# FILENAME,DIR,VAR
a,b,c
d,e,f
g,h,i
j,k,l
# skip this one
# m,n,o
p,q,r
1 Like
5

Well as mentionned above it's like that :

You misread.

At first I only move file with a specific extension (.ext) and if it succeeds, deletes the directory it was contained it.

For now the code is like this

 
#!/bin/sh

FindMoveDelete()
{
find . -iname "$FILENAME*.ext" -exec mv {} "$DIR/$SUBDIR" \; &&
find . -maxdepth 1 -type d -iname "FILENAME*" -exec rm -rf {} \;
}

Function1()
{
FILENAME=name.of.file.and.dir
DIR="/volume1/my/dir"
SUBDIR=Subname
FindMoveDelete
}

Function1
Function2
Function3

What it does :
Move the files "name.of.file.and.dir*.ext" to "/volume1/my/dir/Subname" and then delete the directory "name.of.file.and.dir*"

I call every functions manualy and each of them contains the function FindMoveDelete.
It kinda looks ugly but it works. If you have a better idea I m all hears, at first I only wanted to run a file containing each of those functions for a script to read and end at the end of the file (loop?)

Anyways don't go knock your head on a wall for this, it was just a simple question at first. it's working as it is :slight_smile:

Edit : Just took an other look at your last bit of code and it's interesting. Playing with IFS.. I ll test it out!
Thanks mate

6

It'll run rm no matter what, then :stuck_out_tongue:

I take it you're trying to search inside ${DIR}/${SUBDIR} ? Are there any sub-sub-dirs or is it flat inside?

7

Nop I run the script from my "in" directory only but yes for safety I ll modify the "find ." to "find /data/in" to avoid any problems.. If i happen to run it from other directories.. :slight_smile:

And no, I m not doing any "find" under $DIR/$SUBDIR I m just moving files there. Directories are already created.