Grep patterns in selected lines

Question

How to keep rows containing pattern 1 only if the next line contains pattern 2 and starts with pattern 3 ?

Exemple: keep line starting with an A only if the next line starts with a B and contains the pattern B1

Input file

A   item1A
B   item1B2
A   item2A
A   item3A
B   item3B1
B   item3CB1
B   item3B2
A   item5A
B   item5B1
A   item6A

Expected output:

A   item3A
B   item3B1
B   item3CB1
A   item5A
B   item5B1

( I'm currently trying with grep -e ^A -e B1 but the side effect is that all A s lines are kept )

Please clarify why the third line of your expected output is there? The preceding line doesn't contain the letter A. — garethTheRed
– garethTheRed, Commented Aug 5, 2014 at 14:31
Because that condition is on line starting with an A ("keep line starting with A only if the next line starts with a B and contains the pattern B1"). The only condition for lines starting with B is that they must contain pattern B1... (I'm not sure my verbal explanation is clear enough...) — dovah
– dovah, Commented Aug 5, 2014 at 14:46

Stéphane Chazelas · Accepted Answer · 2014-08-05 14:37:07Z

2

With pcregrep:

pcregrep -M '^(A.*\n)?B.*B1'

With awk:

awk '
  /^B.*B1/ {if (p && last ~ /^A/) print last; print; p=0; next}
  {p=1; last=$0}'

answered Aug 5, 2014 at 14:31

585k96 gold badges1.1k silver badges1.7k bronze badges

1 Answer 1