CODEWITHSUNDEEP
sed
Drago
Drago

Reputation: 101

How to delete lines between the pattern but keep the line with the second pattern

I would like to delete the lines between two patterns but keep the line from the second pattern. For example for the file using MATER / 4401001 302 and / as a pattern:

$#         IDMAT   MATYP             RHO   ISINT    ISHG  ISTRAT   IFROZ
MATER /  4401001     302                       0       0       0        
$# BLANK                                                     QVM           IDMPD
                                                                               0
$#                                                                         TITLE
NAME PLINK Material                                                             
$#  SLFACM     FSNVL    DELTNL     STNOR     STTAN     IFLGC     BLANK    TLSTIF
       0.1                                                 0                    
$#   I3DOF    TOLCOR     IDRUP
         0        1.         0




$---+----1----+----2----+----3----+----4----+----5----+----6----+----7----+----8
$#         IDMAT   MATYP             RHO   ISINT    ISHG  ISTRAT   IFROZ
MATER /  4401005     103        2.753E-6       0       4

I would like to get:

$#         IDMAT   MATYP             RHO   ISINT    ISHG  ISTRAT   IFROZ
MATER /  4401005     103        2.753E-6       0       4

I tried to do it using the following piece of code: 

awk 'BEGIN{p=1} /MATER \/  4401001/ {p=0} /\// {p=1} p' llink1.inc > llink2.inc

But it doesn't work on my CentOS - the both files llink1.inc and llink2.inc are identical.

Could you help. Thank you

Upvotes: 2

Views: 574

Answers (3)

fedorqui
fedorqui

Reputation: 290105

If awk is possible, this can be a way:

$ awk 'BEGIN{p=1} /line 1/ {p=0} /line 2/ {p=1} p' file
bla bla
line 2 
bla bla 
bla bla

It is a matter of using the p print flag, unsetting it when line 1 is found and setting it again when line 2 appears.

To make sure the lines we are printing are the correct ones, say:

$ cat a
1bla bla
line 1
2bla bla  
3gov gov 
line 2 
4bla bla 
5bla bla

$ awk 'BEGIN{p=1} /line 1/ {p=0} /line 2/ {p=1} p' a
1bla bla
line 2 
4bla bla 
5bla bla

Given your new sample input, this works to me:

$ awk 'BEGIN{p=1} /MATER \/  4401001/ {p=0; next} /\// {p=1} p' file
$#         IDMAT   MATYP             RHO   ISINT    ISHG  ISTRAT   IFROZ
MATER /  4401005     103        2.753E-6       0       4

Upvotes: 2

Jotne
Jotne

Reputation: 41460

A short awk

awk '/line 1/ {f=1} /line 2/ {f=0} !f' file
bla bla
line 2
bla bla
bla bla

Even shorter, but not robust.

awk '/line/ {f=!f} !f' file

Upvotes: 0

Kent
Kent

Reputation: 195209

your sed line is close:

sed '/line 1/,/line 2/{/line 2/!d}' file

test

kent$ echo "bla bla
line 1
bla bla  
gov gov 
line 2 
bla bla 
bla bla"|sed '/line 1/,/line 2/{/line 2/!d}'   
bla bla
line 2 
bla bla 
bla bla

same idea with awk: (works for the example in question)

awk '/line 1/,/line 2/{if(!/line 2/)next}7' file

same example:

kent$ (master|✚9) echo "bla bla
line 1
bla bla  
gov gov 
line 2 
bla bla 
bla bla"|awk '/line 1/,/line 2/{if(!/line 2/)next}7'
bla bla
line 2 
bla bla 
bla bla

Upvotes: 5

Related Questions