Replacing all characters in a regular expression match with another character? [duplicate]

Question

I'm doing analysis on a file and I want to mask some characters (while retaining their original byte count) before moving this file down the pipeline. For example, given file.txt:

Hello there Cory Klein
Have fun
Hello there Samantha Rodgers

Writing a regular expression to match the names and substituting them with XXXXX is pretty easy with sed:

$ sed -e "s/\(Hello there \).*/\1XXXXX/" file.txt
Hello there XXXXX
Have fun
Hello there XXXXX

But I would like to replace each character in the name with a X character, like so:

Hello there XXXX XXXXX
Have fun
Hello there XXXXXXXX XXXXXXX

How do I replace all characters matching a regular expression with another character?

Any regular POSIX tool is OK sed, awk, perl, etc. I'm sure I could write a simple python script to accomplish this, but I'm curious whether this is possible with regex alone, which would likely be more succinct. If so, I'd love to learn how so I could likely apply the concept in other places in the future.

Answer 1

With sed you need to filter out lines that don't contain Hello there using an address:

/Hello there/{...}

Then replace whatever single non-whitespace character that comes after Hello there with one x:

s/(^.*Hello there *)?[^[:space:]]/\1x/g

We are leaving Hello there and its preceding characters intact using \1.

The whole command would be:

$ sed -r '/Hello there/{s/(^.*Hello there *)?[^[:space:]]/\1x/g}' file
Hello there xxxx xxxxx
Have fun
Hello there xxxxxxxx xxxxxxx