Compare files to get string before special character

Question

I am trying to evaluate two different files (.tf files) to check if the resource names are the same in a bash script. I know that in those files resources are declared as such:

resource "service_name" "resource_name" {
 #resource config
}

One approach would be something like:

while read line
do
    if word_file1 == "resource"; then
        #save string in array1 before "{" character
    fi
    while read line
    do
        if word_file2 == "resource"; then
            #save string in array2 before "{" character
            if array1 == array2; then
                break
            else 
                # write diff to another file, including all config 
                # info between {} for the missing or different resource
            fi
        fi
    done < filename2
done < filename1

From a test file (file1) an example input would be:

resource "service_name" "resource_name_AA" {
 #resource config
 # policy_config = << POLICY
 { 
     policy_definition

 } POLICY
}

From a test file (file2) an example input would be:

resource "service_name" "resource_name_AA" {
 #resource config
 # policy_config = << POLICY
 { 
     policy_definition

 } POLICY
}

resource "service_name" "resource_name_BB" {
 #resource config
 # policy_config = << POLICY
 { 
     policy_definition

 } POLICY
}

The desired output would be (diff_file):

resource "service_name" "resource_name_BB" {
 #resource config
 # policy_config = << POLICY
 { 
     policy_definition

 } POLICY
}

Answer 1

This might do the job. I guess as per the code you showed, the contents of the resource also need to to printed in cases of mismatch. If only difference needs to be pointed out, diff is better and enough. Anyways I still like writing awk(still a bit new, learning) scripts, so wrote one.

#! /bin/bash

awk '{
    if (FNR == NR)
    {
        if ($1 == "resource")
            resource_name=$3
        else
            contents[resource_name]=contents[resource_name]"\n"

        contents[resource_name]=contents[resource_name]$0
    }
    else
    {
        if (($1 == "}") && (flag == 1))
        {
            flag=0
            next
        }
        else
        if (($1 == "resource") && (contents[$3] != ""))
        {
            flag=1
            contents[$3]=""
            next
        }   
        
        if (flag == 1)
            next
        
        print
    }
} 
END {
    for (resource in contents) 
    {
        if (contents[resource] != "")
            print contents[resource]
    }
}
' file2 file1 > otherfile

---

UPDATE:

#! /bin/bash

awk '{
    if (FNR == NR)
    {
        if ($1 == "resource")
            resource_name=$3
        else
            contents[resource_name]=contents[resource_name]"\n"

        contents[resource_name]=contents[resource_name]$0
    }
    else
    {
        if (($1 == "}") && (flag == 1))
        {
            flag=0
            next
        }
        else
        if (($1 == "resource") && (contents[$3] == ""))
        {
            flag=1
            contents[$3]=""
            next
        }   

        if (flag == 1)
            next

        print
    }
}' file1 file2 > same_resources

---

UPDATE-2:

#! /bin/bash

awk '{
    if (FNR == NR)
    {
        if ($1 == "resource")
            resource_name=$3
        else
            contents[resource_name]=contents[resource_name]"\n"

        contents[resource_name]=contents[resource_name]$0
    }
    else
    {
        if ($1 == "resource")
        {
            if (flag == 1)
            {
                flag=0
            }
            if (contents[$3] != "") 
            {
                flag=1
                contents[$3]=""
            }
        }   

        if (flag == 1)
            next

        print
    }
} 
END {
    for (resource in contents) 
    {
        if (contents[resource] != "")
            print contents[resource]
    }
}' file2 file1 > someotherfile

Answer 2

I think I would try to keep it simpler than that:

 grep 'resource' file1 > file1_resources.txt 
 grep 'resource' file2 > file2_resources.txt 
 diff file{1,2}_resources.txt 

if the word "resource" shows up in different contexts, then you could use a regexp grep instead:

egrep "resource.*\{" fileX