CODEWITHSUNDEEP
regexshellunixsedsh
user331814
user331814

Reputation: 21

Replace only 1 leading zero, if it's more than 1 times then discard it

I want ot replace exactly 1 leading zero with @, if 2 or more leading Zeros found then discard it.

0123
00123
000123
123

Output:

@123
00123
000123
123

I have tried

sed 's/^0.\{1\}/@/' file

Note : Please conside above 4 records in each line of file.

Upvotes: 1

Views: 81

Answers (4)

stack0114106
stack0114106

Reputation: 8811

Using Perl

$ cat lead_zero.txt
0123
00123
000123
123
0
$  perl -pe ' s/(^0)(?!0+.*)/@/ ' lead_zero.txt
@123
00123
000123
123
@
$

If it is in second column of a pipe delimited file, then use this.

$ cat lead_zero.txt2
a|0123|x
b|00123|y
c|000123|xx
d|00000123|yy
e|123|zz
f|0|qq
$ perl -pe ' s/(.*?\|)0(?!0+.*)/$1@/ ' lead_zero.txt2
a|@123|x
b|00123|y
c|000123|xx
d|00000123|yy
e|123|zz
f|@|qq
$

Upvotes: 0

Ed Morton
Ed Morton

Reputation: 204731

You should have included a single 0 on it's own line in your sample input as that's an edge case that could be missed:

$ cat file
0123
00123
000123
123
0

$ sed -E 's/^0([^0]|$)/@\1/' file
@123
00123
000123
123
@

The above will work with any sed that has a -E option to use EREs, e.g. GNU sed and BSD/OSX sed.

Upvotes: 2

James Brown
James Brown

Reputation: 37464

One in awk:

$ awk '{sub(/^0/,(/^00/?0:"@"))}1' file
@123
00123
000123
123

Explained:

$ awk '{sub(/^0/,(/^00/?0:"@"))}1' file
        |    |       |  |  |
        |    |       |  |  otherwise with a @
        |    |       |  with a zero
        |    |       if followed by a zero
        |    leading zero
        replace

and if it is in the second | delimited field, use: 

$ awk 'BEGIN{FS=OFS="|"}{sub(/^0/,($2~/^00/?0:"@"),$2)}1' file
1|@123|1
2|00123|2
3|000123|3
4|123|4

Upvotes: 2

tripleee
tripleee

Reputation: 189948

Your regex looks for a zero, followed by any character. You want to check that it is not followed by a zero, but not replace the nonzero character.

You can capture and put back the character you want to keep:

sed 's/^0\([^0]\)/@\1/' file

Or you can take care to only replace the first character on lines which match this regex:

sed '/^0[^0]/s/^0/@/' file

As ever, specifying that something is repeated exactly once is pointless and redundant.

Upvotes: 1

Related Questions