CODEWITHSUNDEEP
linuxfilebashunixsed
iwan
iwan

Reputation: 7569

Edit file in unix using SED

I have file1.txt in UNIX as following

[Section A]
$param1=value1
$param2=value2

[Section B]
$param1=value1
$param2=value2
$param3=value3

I want to edit value2 in Section B to be new_value2 programmatically

[Section A]
$param1=value1
$param2=value2

[Section B]
$param1=value1
$param2=new_value2
$param3=value3

Any idea what should be the unix command to do this (using sed?)?

Thanks a lot.

Upvotes: 2

Views: 1721

Answers (5)

pizza
pizza

Reputation: 7640

Really simple solution.

ex file1.txt <<"INPUT"
/Section B
/param2
s/value2/new_value2/
:x
INPUT

Upvotes: 0

nosid
nosid

Reputation: 50124

sed -ie '/^\[Section B\]$/,/^$/s/^\$param2=value2$/$param2=new_value/' foo.txt

Edit: The above example is very strict regarding the old value and space characters. I add another example which is probably more suitable. The sed script consists of one command and is prefixed by the following address range:

/^\[Section B\]/,/^\[.*\]/

The address range consists of two regular expressions separated by a comma, and restricts the following command to the lines starting from where the first address matches, and continues until the second address matches (inclusively).

s/^\(\$param2[ \t]*=[ \t]*\).*$/\1new_value/

The substitution command does the actual replacement with the range. Everything together:

sed -ie '/^\[Section B\]/,/^\[.*\]/s/^\(\$param2[ \t]*=[ \t]*\).*$/\1new_value/' foo.txt

Upvotes: 3

Kaz
Kaz

Reputation: 58637

TXR program which parses file, performs edit, and reconstitutes:

@;
@; grab four comand line arguments
@;
@(next :args)
@(cases)
@file
@new_section
@new_param
@new_value
@(or)
@(throw "arguments needed: file section param value")
@(end)
@;
@; hash table mapping sections to assocation lists of values
@;
@(bind sec @(hash :equal-based))
@;
@; parse file, obtaining list of section names and filling in
@; section hash with an associ list of entries.
@;
@(next file)
@(collect)
[Section @secname]
@  (collect)
$@param=@val
@  (until)

@  (end)
@(do (set [sec secname] [mapcar cons param val]))
@(end)
@;
@; now edit
@;
@(do (let ((sec-entries [sec new_section]))
       (if (null sec-entries)
         (push new_section secname))
       (set [sec new_section] (acons-new new_param new_value sec-entries))))
@;
@; now regurgitate file
@;
@(do (each* ((s secname)
             (ent (mapcar (op sec) s)))
       (format t "[Section ~a]\n" s)
       (each ((e ent))
         (format t "$~a=~a\n" (car e) (cdr e)))
       (put-string "\n")))

Test runs:

# edit section B param2 to new_value2

$ txr config.txr config.txt B param2 new_value2
[Section A]
$param1=value1
$param2=value2

[Section B]
$param1=value1
$param2=new_value2
$param3=value3

# add new parameter x with value y to section A

$ txr config.txr config.txt A x y
[Section A]
$x=y
$param1=value1
$param2=value2

[Section B]
$param1=value1
$param2=value2
$param3=value3

# add new section with new parameter

$ txr config.txr config.txt foo bar xyzzy
[Section foo]
$bar=xyzzy

[Section A]
$param1=value1
$param2=value2

[Section B]
$param1=value1
$param2=value2
$param3=value3

Exercise for reader: implement deletion of a param/value pair.

Upvotes: 0

Vijay
Vijay

Reputation: 67301

In case you needa solution in awk :

nawk -F= '{if($0~/Section B/){print;getline;print;getline;gsub(/value2/,"value9",$2);print}else print}' file3

tested Below:

pearl.274> cat file3
[section A]
$param1=value1
$param2=value2

[Section B]
$param1=value1
$param2=value2
$param3=value3
pearl.275> nawk -F= '{if($0~/Section B/){print;getline;print;getline;gsub(/value2/,"new_value2",$2);print}else print}' file3
[section A]
$param1=value1
$param2=value2

[Section B]
$param1=value1
$param2 new_value2 
$param3=value3
pearl.276>

Upvotes: -1

codaddict
codaddict

Reputation: 455400

If Perl is fine with you, you can do:

perl -pe '$f=1 if(/\[Section B\]/);
          s/^\$param2=value2$/\$param2=new_value2/ if($f);' < file

See it

Upvotes: 0

Related Questions