How to search "hello" but exclude "test" from find grep files-CodePudding

I want to search all files include text "hello" but exclude the result contains "test".

Here is the example files:

mkdir -p /tmp/test
cd /tmp/test

echo "foo hello" > foo.txt

echo "bar world" > bar.txt
echo "test hello" >> bar.txt
echo "world hello" >> bar.txt

Here is the search for "hello":

# find /tmp/test -type f -name '*' -exec grep -H -i "hello" {} \;
/tmp/test/bar.txt:test hello
/tmp/test/bar.txt:world hello
/tmp/test/foo.txt:foo hello

Now I want to exclude "test" from the above search output:

# find /tmp/test -type f -name '*' -exec grep -H -i "hello" {} \; | grep -v "test"
...Nothing here...

Try other pattern:

# find /tmp/test -type f -name '*' -exec grep -H -i "hello" -v "test" {} \;
grep: test: No such file or directory
/tmp/test/bar.txt:bar world
grep: test: No such file or directory

Here is the expected output:

# find /tmp/test -type f -name '*' -exec grep [commands argumensts here] {} \;
/tmp/test/bar.txt:world hello
/tmp/test/foo.txt:foo hello

How to do this search and exclude for find in files?

CodePudding user response：

The issue is that you have test in the file path.

You can match for things after the : but that will only work if you do not have : in the text of the file.

find /tmp/test -type f -name '*' -exec grep -H -i

Example:

/t/test  ❯❯❯ find /tmp/test -type f -name '*' -exec grep -H -i "hello" {} \; | grep -v -E ".*\:.*test"
/tmp/test/foo.txt:foo hello
/tmp/test/bar.txt:world hello

CodePudding user response：

You can use awk which can parse file contents very quickly. The following sample script

#!/bin/sh

file1="grepTest1.txt"
echo "dum dum
dum hello dum
crumb test crumb
dum dum" >${file1}

file2="grepTest2.txt"
echo "dum dum
dum hello dum
dum dum" >${file2}

file3="grepTest3.txt"
echo "dum dum
dum test dum
crumb hello crumb
dum dum" >${file3}

file4="grepTest4.txt"
echo "dum dum
crumb jello crumb
crumb bellow crumb
dum test dum
dum dum" >${file4}

for file in grepTest?.txt
do
    awk -v acc="hello" -v rej="test" 'BEGIN{
        good=0 ;
    }
    {
        if( index($0, rej) != 0 ){
            good=2 ;
            exit ;
        }else{
            if( index($0, acc) != 0 ){
                good=1 ;
                line=$0 ;
            } ;
        } ;
    }
    END{
        if( good == 1 ){
            printf("%s: %s\n", FILENAME, line ) ;
        }else{
            if( good == 2 ){
                printf("REJECT: Found %s at line %s in %s.\n", rej, NR, FILENAME ) | "cat >&2" ;
            } ;
        } ;
    }' ${file}
done

gives the output

REJECT: Found test at line 3 in grepTest1.txt.
grepTest2.txt: dum hello dum
REJECT: Found test at line 2 in grepTest3.txt.
REJECT: Found test at line 4 in grepTest4.txt.

BUT ... note that the 3 reject lines go to stderr. if you redirect stdout to a file, you will only have the list of good files in that list.