Skip to main content

Sed

search block between patterns​

search between including pattern​

sample.txt
some thing lkdsfj slkjdf ljfdlsjf
some thing BEGIN lkdsfj slkjdf END ljfdlsjf
some thing lkdsfj slkjdf ljfdlsjf
END
some thing lkdsfj slkjdf ljfdlsjf

The command below returns the lines between first occurrence of BEGIN and first occurrence of END in the next following lines. As you can you can observe in the line containing BEGIN there is END, but this line is ignored. Hence the sed searches for the END in the next line.

sed -n -e '/BEGIN/,/END/ p' sample.txt
output
some thing BEGIN lkdsfj slkjdf END ljfdlsjf
some thing lkdsfj slkjdf ljfdlsjf
END

search between excluding BEGIN pattern in the output​

sed -n -e '/BEGIN/,/END/{ /BEGIN/! p}' sample.txt

replace with sed​

replace newline with comma​

Replace newline with comma and replaces the last comma at the end with newline

sed -z 's/\n/,/g;s/,$/\n/'

replace word in a file​

sed -i '' -e 's/oldword/newword/g' <file-name>

replace a word in all files in a directory​

grep -lr --exclude-dir=".git" -e "oldword" . | xargs sed -i '' -e 's/oldword/newword/g'