CODEWITHSUNDEEP
shell
user1149499
user1149499

Reputation: 583

Change a tab to a semicolon?

I have a file that needs a semicolon delimiter after the first field, which is 4 or 5 characters in length. I tried using sed (modified a sed script I found) and it was an epic fail...it inserted semicolons between every character. I just need to add a semicolon to the end of the first field.

Here is some example text:

2A19    kRSUnicode  205.8
2FA1A   kRSUnicode  206.2
2FA1B   kRSUnicode  207.5

And the output I'm hoping for: 

2A19;   kRSUnicode  205.8
2FA1A;  kRSUnicode  206.2
2FA1B;  kRSUnicode  207.5

Anyone know how to do this? Thank you!

Upvotes: 2

Views: 9548

Answers (6)

galaxywatcher
galaxywatcher

Reputation: 831

Tabs and spaces are known as whitespace. This will replace the first character of whitespace with a semicolon:

sed 's/[[:space:]]/;/'
# or
sed 's/[[:space:]]/; /' # preserves the space char.

This will replace the first instance of all whitespace with a semicolon:

sed 's/[[:space:]]\{1,\}/;/'

This will replace all instances of whitespace with a semicolon:

sed 's/[[:space:]]\{1,\}/;/g'
# or
tr '\t ' ';'

Upvotes: 4

technosaurus
technosaurus

Reputation: 7812

Here is a shell only method: It wasn't 100% clear on whether the first delimiter space was a tab or a variable number of spaces, so I just used echo to squeeze them all into 1 space here

FILE="tmpfile"
OUTPUT=""
while read LINE || [ "$LINE" ]; do
    LINE=`echo $LINE`
OUTPUT="${OUTPUT}${LINE/ /; }
"
done < "${FILE}"
echo "${OUTPUT}" > "${FILE}"

or if it is in fact a single tab, simply:

...
while read LINE || [ "$LINE" ]; do
    OUTPUT="${OUTPUT}${LINE/    /; }
"
done < "${FILE}"
...

Upvotes: 0

jaypal singh
jaypal singh

Reputation: 77155

Update:

sed 's/\(\s\+\)/;\1/' file

Based on the Input Data, you can match the first tab and add ; before it.

[jaypal:~/Temp] cat file
2A19    kRSUnicode  205.8
2FA1A   kRSUnicode  206.2
2FA1B   kRSUnicode  207.5

[jaypal:~/Temp] sed 's/\(\s\+\)/;\1/' file
2A19;    kRSUnicode  205.8
2FA1A;   kRSUnicode  206.2
2FA1B;   kRSUnicode  207.5

Upvotes: 0

SiegeX
SiegeX

Reputation: 140427

Update

If you don't care about anything after the semi-colon, this is even more trivial with awk

$ awk '$0=$1";"' infile
2A19;
2FA1A;
2FA1B;

This is trivial with awk

awk '$1=$1";"' OFS='\t' infile

Output

$ awk '$1=$1";"' OFS='\t' infile
2A19;   kRSUnicode      205.8
2FA1A;  kRSUnicode      206.2
2FA1B;  kRSUnicode      207.5

Upvotes: 2

Balaswamy Vaddeman
Balaswamy Vaddeman

Reputation: 8540

you can loop below sed command for every line

echo "abcde I am a good boy" |sed -e 's/ /:/'

Upvotes: 0

shellter
shellter

Reputation: 37298

It really helps to include sample data and output in your question, but based on what you have said, the following will work

# replace the first space in a line with a semicolon.
printf    "sample data1, data2 data3\n" \
| sed 's/ /;/'

Produces the output 

sample;data1, data2 data3

you can process whatever file you have with

sed 's/ /;/' file > newFile

Or if your using Linux with GNU sed

sed -i 's/ /;/' file

I hope this helps.

Upvotes: 0

Related Questions