How to move all lines starting from string to the top of the text file in Bash

Question

I have a text file containing text data. I'd like to move all lines starting from XY: and OTP XY: to the top of the file. How can I do this in Bash using sed, awk, other?

Answer 1

With GNU ed (note: this modifies file in place; to output the edited file to standard output instead change wq to ,p q):

printf '%s\n' 'g/^\(\|OTP \)XY:/m0' wq | ed -s file

Answer 2

Using grep:

{ grep -E '^(OTP )?XY:' input_file; grep -E -v '^(OTP )?XY:' input_file; } > ouput_file

The 2 grep commands filter the input_file with an without (-v) the wanted pattern.

The { ... } > output_file stores both grepoutput to a file.

Answer 3

Another approach might be to use awk, for example:

awk -v re='^OTP XY:|^XY:'  'NR==FNR && $0 ~ re; NR!=FNR && $0 !~ re' file file

The above uses a single process, never stores more than a single line in memory, but does two passes on the file (it reads it twice).

If one of the two patterns is rather unlikely in the file, i.e. small enough to fit in ram, you could use something like this:

awk  '/^OTP XY:|^XY:/{print;next} {a = $0 ORS a};END{printf "%s", a}' file

This also uses a single process, this time a single pass of the file is made, but the memory usage could increase quite a bit depending on size/contents of the file.