AWK Script Issue insert newline for a regular expression match

Shell Programming and Scripting
21

Then try :

sed '/RXH/,/\}\;/d' input
22

Thanks Ctgsnb

Yeah I also tried like this and it worked perfectly fine with one file
but in another file I dont have

;

after the

}

, in that case what to do ?

 
pin (RXEDF) {
input 
output 
}
 pin (RXEHRR) {
input kdk
output kdkdk 
}
pint (RXHERT) {
input jdjd
output kkd
timing () {
   rise_constraint (setup_template_2x2) {
         index_1 ("1.000, 195.000") ;
   }
fall_constraint (setup_template_2x2) {
         index_1 ("1.000, 195.000") ;
   }
     }
timing () {
       related_pin : "TXHDCLK" ;
       timing_type : hold_rising ;
rise_constraint (hold_template_2x2) {
index_1 ("1.000, 195.000") ;
}
fall_constraint (hold_template_2x2) {
index_1 ("1.000, 195.000") ;
}
}
}
pin (RXCER) {
input ioti
output nfnf
}

Any suggestion for this ? no blank lines as well as no

;

Thanks
Jaita

23 
sed 's/^pin/#&/' input | tr '#' '\n' | sed '/RXH/,/^$/d'

note that this code will add an empty line before every lines beginning with 'pin'
you can then delete them again with a grep -v

note that instead of # you should choose a charactere that appear nowhere else in your code (�,�, ....) up to you to choose a right one

24

Try this:

awk '/RXH/{f=1}; f{c+=/}/?1:/{/?-1:0}; !f; f && !c{f=0}' file
2 Likes
25

Thanks Ctgsnb ,

It worked perfectly fine !! Good Thinking to modulate the code for our requirement

Hi Franklin ,

Can you explain ur code , Not having good expertise on the advanced Awk ,
Will be helpful for me .

Thanks
Jaita

26

@jaita

Note that this code assume your file has its { and } correctly formated :

If somewhere in your code you have

This kind of things :

timing () {   rise_constraint (setup_template_2x2) {
         index_1 ("1.000, 195.000") ;
   }
}

instead of

timing () {
   rise_constraint (setup_template_2x2) {
         index_1 ("1.000, 195.000") ;
   }
}

This may lead to unexpected results (this is just an example of bad formating)

If several { are on the same line whereas the corresponding } are on differents lines, the counter may be mistaken.

To secure Franklin code, you should count the number of { and } per line and increment and decrement the counter accordingly for every hit

27

Thanks Ctgsnb ,

I tried below code and it also worked well

sed -e 's/\(^ pin[^ ]\)/#\1/g' ro.lib | tr '#' '\n' > temp1

Now I am having issue with range of the regular patterns to be matched

For example having a file like this

 
File1
 
##############
 
pin RXA {
input 
output 
}
 
pin RXB {
input 
output 
}
 
pin RXC {
input 
output 
}
 
pin RXD {
input 
output 
}
 
pin RXE {
input 
output 
}
 
pin RXF {
input 
output 
}

Now I need to match range of the regular expression like RXB to RXE and delete their corresponding blocks

 
Output 
 
pin RXA {
input 
output 
}
 
pin RXF {
input 
output 
}

I tried to use below script

sed -e '/RXB/,/RXE/d'

but will not delete the whole content of the RXE Block till RXE Regular expression line it will match and input , output and } will remain as it is

One way is to Remove the black lines between blocks of RXB to RXE
like this

 
input
 
pin RXA {
input 
output 
}
 
pin RXB {
input 
output 
}
 pin RXC {
input 
output 
}
 pin RXD {
input 
output 
}
 pin RXE {
input 
output 
}
 
pin RXF {
input 
output 
}

and use the script

 
sed -e '/RXB/,/^$/d' input

but how to do that in both the situations for matching with range of regular expression and deleting the blank lines for the range of regular expression matches

28 
sed '/RX[BE]/,/^ *$/d' in

# sed '/RX[BE]/,/^ *$/d' in
pin RXA {
input
output
}

pin RXC {
input
output
}

pin RXD {
input
output
}

pin RXF {
input
output
}
#
1 Like
29

Hi Ctgsnb ,

Thanks for the reply

but I want to delete chunk of Blocks from a range of Regular expression RXB to RXE

Reframing the code again

 
pin RXA {
input 
output 
}

pin RXTRY {
input 
output 
}
pin TXCRY {
input 
output 
}
pin BXDRY {
input 
output 
}
pin MTRE {
input 
output 
}

pin RXF {
input 
output 
}

Matching the regular expression from range RXTRY to MTRE and delete the blocks

Output

 
pin RXA {
input 
output 
}

pin RXF {
input 
output 
}

Thanks
Jaita

30 
sed '/RX[B-E]/,/^ *$/d' in

# sed '/RX[B-E]/,/^ *$/d' in
pin RXA {
input
output
}

pin RXF {
input
output
}
31

Hi

I tried this one to the below code but here it is deleting everything

Code

 
pin RX_R_01 {
input 
output 
}
 
pin RX_R_02 {
input 
output 
}
 
pin RX_R_03 {
input 
output 
}
 
pin RX_R_04 {
input 
output 
}
 
pin RX_R_05 {
input 
output 
}
 
pin RX_R_06 {
input 
output 
}

Script

 
sed '/RX_[R_02-R_05]/,/^ *$/d' input

Wanted to delete the Block from RX_R_02 to RX_R_5

to give out the output

Output

 
 
pin RX_R_01 {
input 
output 
}

pin RX_R_06 {
input 
output 
}

but I am getting empty file , it is deleting everything

32 
sed '/RX_R_0[2-5]/,/^ *$/d' input
33

Franklin ,
Your code is not clear to me

 
awk '/RXH/{f=1}; f{c+=/}/?1:/{/?-1:0}; !f; f && !c{f=0}' file

Could not understand how the flags will be set , unset and counter will be incremented and decremented

Suppose I am having a file like this

 
 
FILE 1
 pin ykdh {
 ndn ndn 
 hdjhd jd 
 } 
 pin rtrx {
 in nmd
 mdkd kd
 }
 pin rtry {
 im dnd
 dnd mdmd
}
 pin hdh {
 ndn ndn 
 hdjhd jd 
 }
and I am applying the below script 
awk '/rxt/{f=1}; f{c+=/}/?1:/{/?-1:0}; !f; f && !c{f=0}' file
it is giving me a right output 
output 
pin ykdh {
 ndn ndn 
 hdjhd jd 
 } 
pin hdh {
 ndn ndn 
 hdjhd jd 
 }

But if I am changing a small parameter in the FILE1 { to ]

 
FILE2 
pin ykdh [
 ndn ndn 
 hdjhd jd 
 ] 
 pin rtrx [
 in nmd
 mdkd kd
 ]
 pin rtry [
 im dnd
 dnd mdmd
]
 pin hdh [
 ndn ndn 
 hdjhd jd 
 ]

and applying below script

 
awk '/rxt/{f=1}; f{c+=/]/?1:/[/?-1:0}; !f; f && !c{f=0}' file

it is not giving the output and script getting failed
Could not understand the script as such ,
after matching with the regular expression rxt f is getting set and assigned to value = 1
after that could not able to understand the code

f{c+=/]/?1:/[/?-1:0}; !f; f && !c{f=0}

' setting and unsetting of the flags , incrementing and decrementing
the counters and prining the relevant lines
Please make me understand this code , how it works

Thanks
Jaita

34 
# cat in3
pin RXA {
input
output
}

pin RXTRY {
input
output
}
pin TXCRY {
input
output
}
pin BXDRY {
input
output
}
pin MTRE {
input
output
}

pin RXF {
input
output
}
# awk '/RXA/{f=1}(f==1)&&/^$/{f=2}/RXF/{f=0;printf RS}(f!=2)' in3
pin RXA {
input
output
}

pin RXF {
input
output
}
# tr '{}' '[]' <in3 >in4
# awk '/RXA/{f=1}(f==1)&&/^$/{f=2}/RXF/{f=0;printf RS}(f!=2)' in4
pin RXA [
input
output
]

pin RXF [
input
output
]
#

If your file doesn't contains empty line so that the block are delimited by a line like } ... you can also use } instead of blank line :

# awk '/RXA/{f=1}(f==1)&&/^}$/{f=2}/RXF/{f=0;printf "}"RS}(f!=2)' in3
pin RXA {
input
output
}
pin RXF {
input
output
}

# awk '/RXA/{f=1}(f==1)&&/^]$/{f=2}/RXF/{f=0;printf "]"RS}(f!=2)' in4
pin RXA [
input
output
]
pin RXF [
input
output
]
#
1 Like
35

Thanks Ctgsnb for your reply !

Can u explain me the Frankiln's and ur code as well and
the query I raised for the Franklin's Code in my previous post

Thanks
Jaita

36 
awk '/RXTRY|MTRE/{_++} /}/ && !(_-2) {_=x;next}!_' file
37 
awk '/RXH/{f=1}; f{c+=/}/?1:/{/?-1:0}; !f; f && !c{f=0}' file

Explanation:

/RXH/{f=1} | 	If the line contains RXH, set the flag f
f{c+=/}/?1: | 	If f is set and the line contains a "}" increase counter c ..
/{/?-1:0} | 	..else if the line contains a "{" decrease c
!f | 	print the line if f is not set
f && !c{f=0} | 	If f is set and c == 0 unset the flag