Hi,
I have gigabytes of text files that I need to search for "&" and replace with "&". Is there a way to do this efficiently (like sed command)?
Hope you could help.
Thanks.
give a try to
sed 's/&/&/g' infile
or
sed 's/[&]/&/g' infile
Are there any other uses of the & in the file? If so, we'll need something a bit more elaborate. Also, you probably need:
sed 's/&/&/g' infile >outfile
Ahem, ....
The "&" is a special character in sed regexps and means "the matched part completely here". For instance:
echo "huhu" | sed 's/hu/+&-/g' # will result in "+hu-+hu-"
You will have to escape the "&":
sed 's/\&/\&/g' infile >outfile
I hope this helps.
bakunin
Yes, but if not escaped, it refers to the literal '&' previously matched so in fact it gives the same result 
... but ok, i guess you pointed it out for education purpose...
1 Like
& is only special in the replacement text. It is an ordinary character in the regular expression. In the regular expression, the sequence \& yields undefined behavior. It will probably work as intended with most implementations, but it's not required to.
Regards,
Alister
I do.
vi ./file
:/%s/&/&/g
:wq!
Thanks guys! I was able to use this (placed in a script).
cat $INPUT/$file|sed -e"s/&/&/g" > temp
probably not the best for a gig+ sized file...
Because doing cat of a gig+ file will take a long time or may hand the terminal. Please avoid the cat
try this
sed -e"s/&/&/g" input_file > output_file
hope it will work
hi,
use this:
vi filename
esc
:%s/&/&/
wq
Hi,
well i have a question suppose i have file called testdata.dat,i want to replace the word Mr. wid Miss and write the changes to the same file i.e testdata.dat,what will be syntax?
sed 's/Mr./Miss./g' testdata.dat > testdata.dat
would the above command work?
If the -i option is available on your sed version, just :
sed -i 's/Mr\./Miss./g' testdata
Note the backslash before the first dot :
If you don't put the backslash, it will match any character (dot has a special meaning in regular expression)
for example in the case someone is called "Mr.Mrimba"
see what will happen with or without the backslash:
$ echo "Mr.Mrimba" | sed 's/Mr./Miss./g'
Miss.Miss.mba
$ echo "Mr.Mrimba" | sed 's/Mr\./Miss./g'
Miss.Mrimba
---------- Post updated at 01:29 PM ---------- Previous update was at 01:24 PM ----------
So if the backslash is mandatory if you want to match the literal dot, otherwise it will be interpreted as a regular expression which may lead to unexpected parsing.
This will also work
#sed -i 's/&/&/' file_name
Also can be done using vim.
for file in $(grep -l "PATTERN" `find . -name "*.php"`); do vim "+%s:PATTERN:REPLACE:g" "+wq" $file; done;
It can also be accomplished easily with :
perl -p -i -e 's/&/&/g' infile
@ubuntu:/tmp$ cat test
100
100
100
100
perl -pi -e 's/100/200/g' test
@ubuntu:/tmp$ cat test
200
200
200
200
NO!
Never ever use the same file at both sides of a pipe or redirection. It will be corrupted since process in concurrent. It does not finish and then pass the results as DOS did.
But yes, this is the usage of s/a/b/g construct.