Finding and replacing lines that begin with a pattern

Question

I have a text in a file file.txt like this

xxxxxxxxxxxxxxx
xxxxxxxxxxxxxxx
a    b   c // delimited by tab
xxxxxxxxxxxxxxx
xxxxxxxxxxxxxxx

I know using sed I can find and replace text in a file. If a line starts with a b(seperated by a tab) I need to replace it with d e f. So the above file will be

xxxxxxxxxxxxxxx
xxxxxxxxxxxxxxx
d    e   f // delimited by tab
xxxxxxxxxxxxxxx
xxxxxxxxxxxxxxx

I can do this to find and replace, I want only those instances where the line starts with a b and replace the whole line.

sed -i 's/a/\t/\b/\t\/c/REPLACED TEXT/g' file.TXT

jahroy · Accepted Answer · 2012-07-26T01:26:43.767

30

Use a ^ symbol to represent the beginning of a line:

sed -i 's/^a\tb.*$/REPLACED TEXT/g' file.TXT

Exmplanation:

^ means beginning of line/input
\t means tab symbol
. means any character
* means zero or more of the preceeding expression
$ means end of line/input

edited Jul 26 '12 at 01:26

answered Jul 25 '12 at 23:10

jahroy

22,322
9
59
108

score 1 · Answer 2 · answered Jul 25 '12 at 23:13

1

The following will replace the entire line if it begins with a<tab>b<tab>c. The .*$ makes the match include the entire line for replacement. Your example regex included c but the prose only mentioned a and b, so it wasn't quite clear if c is required. If not, then remove \tc from the regex.

s/^a\tb\tc.*$/REPLACED TEXT/g

answered Jul 25 '12 at 23:13

Mark Wilkins

40,729
5
57
110

yes we can remove \tc .. The line only have to start with a and b – Ank Jul 25 '12 at 23:16

score 1 · Answer 3 · answered Jul 26 '12 at 16:13

1

With awk

awk '/^a\tb/{$0="REPLACED TEXT"} 1' foo.txt

answered Jul 26 '12 at 16:13

slitvinov

5,693
20
31

score 0 · Answer 4 · answered Jul 26 '12 at 08:29

0

This might work for you (GNU sed):

sed -i '/^a\tb/c\replacement text' file

answered Jul 26 '12 at 08:29

potong

55,640
6
51
83

Finding and replacing lines that begin with a pattern

4 Answers4

Linked