Question

I have a file:

bla bla bla
bla bla bla
a rs1
a rs2
a rs3
b rs4
b rs5
b rs6
bla bla bla
bla bla bla

where rsN is random stuff.

I’m trying to get last line with a and first line with b:

a rs3
b rs4

with:

> grep -Po "(?s)^a.*?$.*?^b.*?$" test.regexp
a rs1
a rs2
a rs3
b rs4

but apparently I’m missing something.

thanks.

Answer 1

Here is something you can try with awk.

awk '$1=="a"{a=$0;next}$1=="b"{print a; print $0;exit}' file

Test:

[jaypal:~/Temp] cat file
bla bla bla
bla bla bla
a rs1
a rs2
a rs3
b rs4
b rs5
b rs6
bla bla bla
bla bla bla
[jaypal:~/Temp] awk '$1=="a"{a=$0;next}$1=="b"{print a; print $0;exit}' file
a rs3
b rs4