How to search and replace long HTML in multiple files on Linux

Question

I need to recursively find all files with this HTML:

<html id="blx-5fb3c619e82a2863d6567c52-000000001" class="blx-5fb3c619e82a2863d6567c52"><head>
    <meta charset="utf-8">
    <meta name="google" value="notranslate">

And replace it with this HTML:

<html id="blx-5fb3c619e82a2863d6567c52-000000001" class="blx-5fb3c619e82a2863d6567c52"><head>
    <meta charset="utf-8">
    <meta name="google" value="notranslate">
    <meta name="format-detection" content="telephone=no">
    <meta name="format-detection" content="date=no">
    <meta name="format-detection" content="address=no">
    <meta name="format-detection" content="email=no">

This is my unsuccessful attempt of a grep command piped to a sed:

grep --include="index.html" -PRwzl -e '<html id="blx-5fb3c619e82a2863d6567c52-000000001" class="blx-5fb3c619e82a2863d6567c52"><head>\n    <meta charset="utf-8">\n    <meta name="google" value="notranslate">\n' | xargs -i@ sed -i 's/<html id="blx-5fb3c619e82a2863d6567c52-000000001" class="blx-5fb3c619e82a2863d6567c52"><head>\n    <meta charset="utf-8">\n    <meta name="google" value="notranslate">\n/<html id="blx-5fb3c619e82a2863d6567c52-000000001" class="blx-5fb3c619e82a2863d6567c52"><head>\n    <meta charset="utf-8">\n    <meta name="google" value="notranslate">\n    <meta name="google" value="notranslate">\n    <meta name="format-detection" content="telephone=no">\n    <meta name="format-detection" content="date=no">\n    <meta name="format-detection" content="address=no">\n    <meta name="format-detection" content="email=no">\n/g' @

The grep command alone works perfectly.

For clarity, here is the command split into many sections.:

grep --include="index.html" \
    -PRwzl \
    -e '<html id="blx-5fb3c619e82a2863d6567c52-000000001" class="blx-5fb3c619e82a2863d6567c52"><head>
        \n    <meta charset="utf-8">
        \n    <meta name="google" value="notranslate">
        \n' \
    | xargs -i@ sed -i 's/<html id="blx-5fb3c619e82a2863d6567c52-000000001" class="blx-5fb3c619e82a2863d6567c52"><head>
                            \n    <meta charset="utf-8">
                            \n    <meta name="google" value="notranslate">
                            \n
                        /<html id="blx-5fb3c619e82a2863d6567c52-000000001" class="blx-5fb3c619e82a2863d6567c52"><head>
                            \n    <meta charset="utf-8">
                            \n    <meta name="google" value="notranslate">
                            \n    <meta name="google" value="notranslate">
                            \n    <meta name="format-detection" content="telephone=no">
                            \n    <meta name="format-detection" content="date=no">
                            \n    <meta name="format-detection" content="address=no">
                            \n    <meta name="format-detection" content="email=no">
                            \n
                        /g' @

Your attempt is a `grep` command which you claim is unsuccessful but works perfectly? Clarify. — Nic3500, Oct 20 '21 at 02:18
@Nic3500 My attempt is a grep command piped to a sed command. The grep command alone works fine. — standac, Oct 20 '21 at 11:48

score 0 · Accepted Answer · answered Oct 22 '21 at 01:19

Your command is very complex for nothing. You can run your sed on the file, without the grep and xargs before. Typically an inline edit of a file with sed looks like:

sed -i 's/TO_FIND/REPLACE/' FILE.txt

Another comment, sed is not a great tool to edit HTML. Look at RegEx match open tags except XHTML self-contained tags.

That being said I propose this script to meet your requirement.

#!/bin/bash
#
find . -type f -name "*.html" -print0 | while IFS= read -r -d '' file
do
    if [[ $(grep -c 'id="blx-5fb3c619e82a2863d6567c52-000000001" class="blx-5fb3c619e82a2863d6567c52"' $file) -ne 0 ]]
    then
        # Add the content...
        echo "Adding in file $file"
        sed -i 's#</head>#    <meta name="format-detection" content="telephone=no">\n    <meta name="format-detection" content="date=no">\n    <meta name="format-detection" content="address=no">\n    <meta name="format-detection" content="email=no">\n    </head>#' "$file"
    else
        echo "Nothing to do on $file"
    fi
done

Using find with while and read cover cases where you have HTML files in sub-directories.
The grep has been highly simplified. If the id and class values are present, it is enough to identify valid files.
Then in the sed, you can just add the new lines. Your sed replaced lines with these same lines.
I used # as a separator in sed instead of / to avoid confusion with HTML code.
This is based on a file I created myself, since you did not provide a sample. You should provide samples in your questions.
The order of tags within the <head> section is not relevant, so adding lines just before the closing </head> works.
Obviously the else section is optional.
<opinion>I find this type of script easier to understand and debug in the future than long single liners.</opinion>.

Assuming that index.html is:

<html id="blx-5fb3c619e82a2863d6567c52-000000001" class="blx-5fb3c619e82a2863d6567c52"><head>
    <meta charset="utf-8">
    <meta name="google" value="notranslate">
    <title>TITRE</title>
</head>
    <body>
        <p>PARAGRAPH</p>
    </body>
</html>

The result is:

<html id="blx-5fb3c619e82a2863d6567c52-000000001" class="blx-5fb3c619e82a2863d6567c52"><head>
    <meta charset="utf-8">
    <meta name="google" value="notranslate">
    <title>TITRE</title>
    <meta name="format-detection" content="telephone=no">
    <meta name="format-detection" content="date=no">
    <meta name="format-detection" content="address=no">
    <meta name="format-detection" content="email=no">
    </head>
    <body>
        <p>PARAGRAPH</p>
    </body>
</html>

How to search and replace long HTML in multiple files on Linux

1 Answers1