find comand - executing >1 statement

[Danny Fang]

Hi,

I'm attempting to execute >1 statement in within the -exec option of the find command:

bash-2.03$ find . -depth -type f -name "*.xml" -exec grep -l "summaryid" |awk -F/ '{print("cp" $0 "/tmp/" $NF)}' {} \;
find: incomplete statement
awk: can't open {}

From the command line above, I'm attempting to list the instances of files containing the string "simmaryid" within the current directory, and copy those files to another directory if the "grep" command succeeds. However, I obtained the error message shown above.

I had to resort to writting it in a longer method, which eventually works:
#!/bin/sh

find . -type f -name ".xml" -print |while read i
do
if [! `grep -l summaryid`]; then
echo "PATTERN DNE"
else
cp $i /tmp
fi
done

I was wondering if anyone could show me the method of executing >1 statement within the -exec option of the find command? I'm trying to do it within 1 line.

Thanks in advance.
Danny

[Ermin Borovac]

How about

find . -depth -type f -name "*.xml" -exec grep -l "summaryid" {} \; |awk -F/ '{print("cp" $0 "/tmp/" $NF)}'

[Stephen Keane]

Write a shell script containing all the lines you want. Call the shell script from the exec in the find comand.

[Pete Randall]

Danny,

Try using xargs instead:

$ find . -depth -type f -name "*.xml" | xargs grep -l "summaryid" |awk -F/ '{print("cp" $0 "/tmp/" $NF)}'


Pete

[Danny Fang]

Hi everyone,

I was wondering if there's another way of doing the same thing shown below, still the one-liner of the -exec option; but without using awk ?

Greatly appreciate any ideas.

Danny

[Kent Ostby]

Well this still uses awk, but it encapsulates it into a file so its cleaner.

Create useme.awk as:

{print"cp", $0, "/tmp/" $NF}

Script file:
rm ./useit

find . -depth -type f -name "*.xml" | xargs grep -l "summaryid" |awk -f useme.awk > useit

chmod +x useit
./useit

[Ermin Borovac · Accepted solution]

You can do it with multiple exec statements.

$ find . -depth -type f -name "*.xml" -exec grep -l "summaryid" {} \; -exec cp {} /tmp \;

Second set of curly braces ({}) will only have files that passed the grep.

[Ermin Borovac]

You may want to replace 'grep -l' with 'grep -q' in my last post.

[Sandman!]

You could use the following one liner w/o awk

# find . -depth -type f -name "*.xml" -exec grep -il "summaryid" {} \; | xargs -i cp {} /tmp/{}