使用find和sed将文件名插入文件

Question

I have hundreds .htm files and i need to replace <title>sometext</title> with the actual file name. For example: i have file called records-england.htm and i need to replace <title>sometext</title> in this file with file name like so <title>records-england.htm</title>

Even better if without extension - but that does not matter. I'v tried like this but it's not working:

find . -name *.htm | while read file; do sed "s|<title>sometext</title>|<title>$file</title>|g" {} \; done

Any advice?

Answer 1

You can use awk s FILENAME. Needs to be gnu awk >= 4.1 due to -i inplace

awk -i inplace '{a=FILENAME;sub(/\.htm$/,"",a);gsub(/<title>[^<]*<\/title>/,"<title>"a"</title>")}1' *.htm

This will change <title>sometext</title> (where sometext can be anything) to <title>filename(without extension)</title>

Answer 2

To fix the problem in your code, you can try this :

find -name "*.htm" | while read file; do sed "s|<title>sometext</title>|<title>${file##*/}</title>|g" -i $file; done

Sed need the -i option to specify which files you want to edit and I use a parameter expansion to only display the basename of the file.

Answer 3

这样尝试

's/\<title\>.*\<title\>/title>$file<title/g'

Answer 4

This would work for any arbitrary title text. If you don't want it to be any text but some fixed text then change <title>.*</title> to <title>sometext</title> .

 find -type f -name '*.htm' -printf '%P\0%f\0' | xargs -0 -n2 sh -c 'fil="${2/\.htm/}"; sed -i -e "s;<title>.*</title>;<title>$fil</title>;" $1' replace

It transfers the result of find to sh using xargs ; bash strips the extension off and give the arguments to sed which does the replacement.

Arguments:

To find

• -type f find files
• -name '*.htm' which has extension htm
• printf '%P\\0%f\\0' output full file path then a ASCII null, followed by just the filename (stripping the path) and again an ASCII nul

To xargs

• -0 use ASCII nul as the delimiter
• -n2 pick up at most 2 arguments from the list for each execution of the command

To sh

• -c reads commands from the first non-option argument
• a bash script which takes $1 (file path) and $2 (file name) as arguments; does regex replacement to snip the extension. Passing a dummy string replace for it to take $0

To sed

• -i in-place replace in file
• -e the script to execute

Answer 5

You could also try this:

#!/bin/bash
outFile=/tmp/myOutFile.out
touch $outFile
while read line
do
    name=$line
    echo "<title>$name</title>" >> $outFile
done < $1

1. Create a new .sh (eg modifyFile.sh) file with the code above.
2. Grant it execute rights - chmod +x modifyFile.sh
3. Run it with the file which contains your file names or w/e - ./modifyFile.sh myInputFile.txt
4. You will get results in /tmp/myOutFile.out