I found your email in the site www.regularexpression.info

I need a regular expression to remove the duplicate lines in my code. Please help me.

example code :

Code:

edit(abcd); edit(abcd); edit(1234);

And it should be

Code:

edit(abcd); edit(1234);

I need a regular expression for this.


What language did you want to use?   Here's a possible solution in Perl

Code:

#!/usr/bin/perl open (FH,$ARGV[0]) or die ("No input file\n"); read (FH,$file,-s $ARGV[0]); $file =~ s/^(.*[\n\r]+)\1+/$1/mg; print $file;

Regular expressions are REALLY powerful - it's worth learning how to
write them yourself!

-- Technical note -- Nice example of using back references to previous matches (\1, etc) and also the "m" modifier to have ^ match at the embedded start of  any new line. .

I know you asked for a regular expression...but...Another way to do this would be to use a hash. (or associative array in other languages)

loop through your line entries as an array and put them in a hash as the key.
A hash can only have one unique key and will clobber any duplicates with itself.

Code:

my %uniques; foreach my $key(@array_of_lines){    $uniques{$key}++; }

The hash %uniques now holds only unique values as keys, the value will be a count of the amount of duplicates for that particular line.

I'm sure I'll be corrected but this should be faster than a regex?

jfp

Yes, that will indeed remove duplicate lines, but in a rather different way:

a) With a regular expression, duplicate lines are only removed if they follow one another directly; if a line occurs early in the file then happens to crop up again much later, the later copy will not be removed.

b) With a regular expression, the order of the lines is maintained.  Using a hash you can certainly output each unique line, but if they have to be in the same order as in the original input file, that's going to be an interesting task!

Looking back at the original question - I'm really not sure which the problem was - the one that's solved with a hash, or the one that's solved with a regular expression.   My original enquirer has written back "thanks for the reply", which doesn't give me any clue either; I guess we'll never know what the customer wanted  

---