linux replace all occurrences of string in file code example

Example 1: ubuntu command line replace word in files

sed -i 's/original/new/g' file.txt

Example 2: sed replace

For using match in sed replacement, just border it with '\(' and '\)':
echo Before123 | sed 's/Before\([0-9]*\)/\1After/g'
123After	# number is matched withtin '\( \)' and replaced in '\1'
Example with 2 match replacements
echo a_b | sed 's/\(^.*\)_\(.*$\)/first is \1 and \2 is after/g'