scripting error

nanda22 · 09-04-2008, 05:28 AM

Hi
new bie for scripting in linux
i want to write a script which takes the string as the argument and remove the files by going into that specific folder.
the script looks like this:

set JAN FEB MAR APR
echo $1 $2 $3 $4
i=1
while [$i -le 4]
do
find /home/susee/test/$i/*/data/ -type f -name "*.247" -exec rm {} \;
i='expr $i + 1'
done

but i'm getting the following error

test.sh: line 5: [1: command not found

where am i going wrong?

colucix · 09-04-2008, 05:45 AM

You have to put a space between the square brackets and the expression inside:

Code:

while [ $i -le 4 ]

Moreover, use the code tags to embed lines of code. In this way the formatting is preserved. To put code tags either go in advanced mode, select the lines of code and press the # button, or explicitly write [CODE] and [/CODE] before and after the lines respectively.

nanda22 · 09-04-2008, 06:00 AM

Now the error is like this

--------------------------------------------------------------
find: /home/susee/test/1/*/data/: No such file or directory
test.sh: line 5: [: expr $i + 1: integer expression expected
--------------------------------------------------------------
why it is taking '1' as argument instead of 'AU'?

and i didnt get what is "integer expression expected"

colucix · 09-04-2008, 06:05 AM

What is 'AU'? Please, post the last version of your script, using code tags.

The error should be

Code:

i='expr $i + 1'

here you have to use back ticks (not single quotes) to do command substitution. Or better the $(command) syntax which has some advantages over back ticks:

Code:

i=`expr $i + 1`    # or the equivalent
i=$(expr $i + 1)

You may also consider the arithmetic expansion with the double parentheses construct, in bash.

nanda22 · 09-05-2008, 03:13 AM

Code:

IFS=:
set A B C D
echo $1 $2 $3 $4
i=1
while [ $i -le 3 ]
do
find /home/susee/test/$i/*/data/ -type f -name "*.247"  -exec rm {} \;
i=$(expr $i + 1)
done

Now the error is
-----------------------------------------------------------
find: /home/susee/test/1/*/data/: No such file or directory
find: /home/susee/test/2/*/data/: No such file or directory
find: /home/susee/test/3/*/data/: No such file or directory
------------------------------------------------------------

The A,B,C and D are folders, which contain the data.

colucix · 09-05-2008, 04:34 AM

You have to carefully check the path: are you sure you have some directories under 1, 2 and 3...? Moreover, for testing purposes better put an echo in front of the rm command, otherwise you may end up with unwanted deleting of some files:

Code:

find /home/susee/test/$i/*/data/ -type f -name "*.247"  -exec echo rm {} \;

or you can use the -ok action instead of -exec. This will ask you confirmation before actually executing the rm command:

Code:

find /home/susee/test/$i/*/data/ -type f -name "*.247" -ok rm {} \;

matthewg42 · 09-05-2008, 04:34 AM

Rather than using set to replace the positional parameters, you might consider this simpler and clearer by to iterative over the list of months:

Code:

#!/bin/bash

for month in JAN FEB MAR APR; do
    find /home/susee/test/$month/*/data/ -type f -name "*.247" -exec rm {} \;
done

Note that file names on unix-like systems (such as Linux) are case sensitive. If you are getting "file not found" messages, check the path is correct - including the case of all characters.

If there are a large number of files to delete, the method above can be somewhat inefficient, as an rm process is invoked for every file. There is a more efficient method using xargs, although it is a little harder to understand how it works. If you are interested in this other method, please ask. I don't want to post it right away to avoid overwhelming you. :-)

nanda22 · 09-05-2008, 06:51 AM

Dear matthewg42,

I would be obviously grateful to you, if you are going to teach me something more.

matthewg42 · 09-05-2008, 07:00 AM

OK, it's quite a long explanation, but it should give you some idea how it works:

http://www.linuxquestions.org/questi...1/#post2985881