Regex to remove \t \r \n from string

[morleyc@gmail.com]

Hi, i would like to remove a number of characters from my string (\t
\r \n which are throughout the string), i know regex can do this but i
have no idea how. Any pointers much appreciated.

Chris

[Nicholas Paldino [.NET/C# MVP]]

Re: Regex to remove \t \r \n from string

Chris,

Why not just use three calls to the Replace method on the String class?

string myString = input.Replace(" \t", "").Replace("\r ", "").Replace("\n ",
"");

You can use the character version here as well if you wish.

--
- Nicholas Paldino [.NET/C# MVP]
- mvp@spam.guard. caspershouse.co m

<morleyc@gmail. comwrote in message
news:1179238212 .282258.281370@ y80g2000hsf.goo glegroups.com.. .

Hi, i would like to remove a number of characters from my string (\t
\r \n which are throughout the string), i know regex can do this but i
have no idea how. Any pointers much appreciated.
>
Chris
>

[morleyc@gmail.com]

Re: Regex to remove \t \r \n from string

Why not just use three calls to the Replace method on the String class?

I am currently using the 3 replace calls :), however i have always
avoided reglular expressions before this seemed the ideal excuse to
learn them! I would also be interested in turning \r\n in a string to
just \n also. im sure it must be possible?

[Nicholas Paldino [.NET/C# MVP]]

Re: Regex to remove \t \r \n from string

Absolutely, just wondering why you wouldn't take the simpler, more
maintainable (depending on who is looking at it, at least from my point of
view) approach. =)

In this case, I believe you can have a regular expression of "[\t\r\n]"
and then call the Replace method, passing your input string and an empty
string (or whatever you want to replace any of the characters in that set
with) and it should work.


--
- Nicholas Paldino [.NET/C# MVP]
- mvp@spam.guard. caspershouse.co m

<morleyc@gmail. comwrote in message
news:1179243139 .701933.208180@ y80g2000hsf.goo glegroups.com.. .

> Why not just use three calls to the Replace method on the String
>class?

>
I am currently using the 3 replace calls :), however i have always
avoided reglular expressions before this seemed the ideal excuse to
learn them! I would also be interested in turning \r\n in a string to
just \n also. im sure it must be possible?
>
>
>
>

[tomisarobot@gmail.com]

Re: Regex to remove \t \r \n from string

it certainly is possible. you should create a little test project and
play with it. thing to remember about regex is to start small and
build up. its not hard really, but its horribly easy to assume that
things will behave differently than the reality.

been a while since ive done captures with PCRE, but for the simple
replace you are probably looking at something like this: [\r|\n|\t]

..net also has some context variable to make sure you have your
endlines localized correctly if thats all you are trying to do.

[Ben Voigt]

Re: Regex to remove \t \r \n from string


"Nicholas Paldino [.NET/C# MVP]" <mvp@spam.guard .caspershouse.c omwrote in
message news:49CF19FA-B96B-4CEF-A994-FCE96B26129A@mi crosoft.com...

Absolutely, just wondering why you wouldn't take the simpler, more
maintainable (depending on who is looking at it, at least from my point of
view) approach. =)

Because your simpler method involves three complete string copies instead of
one!


RegEx.Replace ought to do it.

[Jon Skeet [C# MVP]]

Re: Regex to remove \t \r \n from string

Ben Voigt <rbv@nospam.nos pamwrote:

"Nicholas Paldino [.NET/C# MVP]" <mvp@spam.guard .caspershouse.c omwrote in
message news:49CF19FA-B96B-4CEF-A994-FCE96B26129A@mi crosoft.com...

Absolutely, just wondering why you wouldn't take the simpler, more
maintainable (depending on who is looking at it, at least from my point of
view) approach. =)

>
Because your simpler method involves three complete string copies instead of
one!

Do we have any evidence that performance is an issue here? Further, do
we have evidence that regular expressions will actually make this
faster on the sample data?

Until both of those have been determined, I'd take a default course of
the simplest code which does the job.

RegEx.Replace ought to do it.

At what cost to readability though?

--
Jon Skeet - <skeet@pobox.co m>
http://www.pobox.com/~skeet Blog: http://www.msmvps.com/jon.skeet
If replying to the group, please do not mail me too

[Ben Voigt]

Re: Regex to remove \t \r \n from string


"Jon Skeet [C# MVP]" <skeet@pobox.co mwrote in message
news:MPG.20b43a 5f6978d6c8103@m snews.microsoft .com...

Ben Voigt <rbv@nospam.nos pamwrote:

>"Nicholas Paldino [.NET/C# MVP]" <mvp@spam.guard .caspershouse.c omwrote
>in
>message news:49CF19FA-B96B-4CEF-A994-FCE96B26129A@mi crosoft.com...

Absolutely, just wondering why you wouldn't take the simpler, more
maintainable (depending on who is looking at it, at least from my point
of
view) approach. =)

>>
>Because your simpler method involves three complete string copies instead
>of
>one!

>
Do we have any evidence that performance is an issue here? Further, do
we have evidence that regular expressions will actually make this
faster on the sample data?
>
Until both of those have been determined, I'd take a default course of
the simplest code which does the job.

Well, ok, but you asked why anyone would ever choose not to do it that way,
and I gave an example.

>

>RegEx.Replac e ought to do it.

>
At what cost to readability though?

Admittedly, a String.Replace( RegEx, String) method would be far more
readable, but set up a dependency from string on RegEx.

>
--
Jon Skeet - <skeet@pobox.co m>
http://www.pobox.com/~skeet Blog: http://www.msmvps.com/jon.skeet
If replying to the group, please do not mail me too

[Jon Skeet [C# MVP]]

Re: Regex to remove \t \r \n from string

Until both of those have been determined, I'd take a default course of

the simplest code which does the job.

>
Well, ok, but you asked why anyone would ever choose not to do it that way,
and I gave an example.

That's fair enough.

RegEx.Replace ought to do it.

At what cost to readability though?

>
Admittedly, a String.Replace( RegEx, String) method would be far more
readable, but set up a dependency from string on RegEx.

More importantly, it sets up a dependency on the reader understanding
regular expressions, which I've seen causing issues time and time again
in these newsgroups.

I'm all for regular expressions when their power is really needed, but
that tends to be pretty rare IME.

--
Jon Skeet - <skeet@pobox.co m>
http://www.pobox.com/~skeet Blog: http://www.msmvps.com/jon.skeet
If replying to the group, please do not mail me too

[=?ISO-8859-1?Q?Arne_Vajh=F8j?=]

Re: Regex to remove \t \r \n from string

Jon Skeet [C# MVP] wrote:

Ben Voigt <rbv@nospam.nos pamwrote:

>"Nicholas Paldino [.NET/C# MVP]" <mvp@spam.guard .caspershouse.c omwrote in
>message news:49CF19FA-B96B-4CEF-A994-FCE96B26129A@mi crosoft.com...

>> Absolutely, just wondering why you wouldn't take the simpler, more
>>maintainabl e (depending on who is looking at it, at least from my point of
>>view) approach. =)

>Because your simpler method involves three complete string copies instead of
>one!

>
Do we have any evidence that performance is an issue here? Further, do
we have evidence that regular expressions will actually make this
faster on the sample data?

A simple test seems to indicate that regex is slower.

String Replace : 15 -12 x 6666666 : 6,6875
StringBuilder Replace : 15 -12 x 6666666 : 6,546875
Regex Replace : 15 -12 x 6666666 : 27,1875
Regex Replace Optimized : 15 -12 x 6666666 : 15,828125
String Replace : 960 -768 x 104166 : 3,3125
StringBuilder Replace : 960 -768 x 104166 : 2,03125
Regex Replace : 960 -768 x 104166 : 17,421875
Regex Replace Optimized : 960 -768 x 104166 : 13,4375
String Replace : 1000 -1000 x 100000 : 1,15625
StringBuilder Replace : 1000 -1000 x 100000 : 2,4375
Regex Replace : 1000 -1000 x 100000 : 3,78125
Regex Replace Optimized : 1000 -1000 x 100000 : 2,703125

(see code below)

>RegEx.Replac e ought to do it.

>
At what cost to readability though?

Actually I think the regex code is more readable.

Arne

=============== =============== =============== =============

using System;
using System.Text;
using System.Text.Reg ularExpressions ;

namespace E
{
public class MainClass
{
private const int N = 100000000;
private const string FMT = "{0,-25} : {1} -{2} x {3} : {4}";
private static void TestStringRepla ce(string s)
{
int n = N / s.Length;
string s2 = null;
DateTime dt1 = DateTime.Now;
for(int i = 0; i < n; i++)
{
s2 = s.Replace("\r", "").Replace("\n ", "").Replace("\t ", "");
}
DateTime dt2 = DateTime.Now;
Console.WriteLi ne(String.Forma t(FMT, "String Replace", s.Length,
s2.Length, n, (dt2 - dt1).TotalSecon ds));
}
private static void TestStringBuild erReplace(strin g s)
{
int n = N / s.Length;
StringBuilder sb = new StringBuilder(s );
string s2 = null;
DateTime dt1 = DateTime.Now;
for(int i = 0; i < n; i++)
{
s2 = sb.Replace("\r" , "").Replace("\n ", "").Replace("\t ",
"").ToStrin g();
}
DateTime dt2 = DateTime.Now;
Console.WriteLi ne(String.Forma t(FMT, "StringBuil der Replace",
s.Length, s2.Length, n, (dt2 - dt1).TotalSecon ds));
}
private static void TestRegexReplac e(string s)
{
int n = N / s.Length;
string s2 = null;
DateTime dt1 = DateTime.Now;
for(int i = 0; i < n; i++)
{
s2 = Regex.Replace(s , "[\r\n\t]", "");
}
DateTime dt2 = DateTime.Now;
Console.WriteLi ne(String.Forma t(FMT, "Regex Replace", s.Length,
s2.Length, n, (dt2 - dt1).TotalSecon ds));
}
private static void TestRegexReplac eOptimized(stri ng s)
{
int n = N / s.Length;
Regex re = new Regex("[\r\n\t]", RegexOptions.Co mpiled);
string s2 = null;
DateTime dt1 = DateTime.Now;
for(int i = 0; i < n; i++)
{
s2 = re.Replace(s, "");
}
DateTime dt2 = DateTime.Now;
Console.WriteLi ne(String.Forma t(FMT, "Regex Replace Optimized",
s.Length, s2.Length, n, (dt2 - dt1).TotalSecon ds));
}
private static void Test(string s)
{
TestStringRepla ce(s);
TestStringBuild erReplace(s);
TestRegexReplac e(s);
TestRegexReplac eOptimized(s);
}
public static void Main(string[] args)
{
string shortstr = "aaa\rbbb\nccc\ tddd";
Test(shortstr);
string longstr = shortstr;
longstr += longstr;
longstr += longstr;
longstr += longstr;
longstr += longstr;
longstr += longstr;
longstr += longstr;
Test(longstr);
string nonestr = String.Empty.Pa dRight(1000, 'A');
Test(nonestr);
Console.ReadLin e();
}
}
}

[Jon Skeet [C# MVP]]

Re: Regex to remove \t \r \n from string

Arne Vajhøj <arne@vajhoej.d kwrote:

At what cost to readability though?

Actually I think the regex code is more readable.

Well, it's interesting that your regex is "[\r\n\t]". I'm actually
slightly surprised this even works, as the \r, \n and \t are being
taken literally by the regex engine rather than having been escaped in
the normal way. I'd have expected "[\\r\\n\\t]" or @"[\r\n\t]" to make
it clear to the regex engine that you really meant the carriage return
etc to be part of the regex, and not incidental or for the sake of
readability (splitting the regex over several lines, as shown in
Jesse's example in another thread).

That extra level of escaping which is required in *some* cases (but
clearly not all) as well as having to understand the basic language of
regex in the first place is what makes it less readable in my opinion.

--
Jon Skeet - <skeet@pobox.co m>
http://www.pobox.com/~skeet Blog: http://www.msmvps.com/jon.skeet
If replying to the group, please do not mail me too

[=?ISO-8859-1?Q?Arne_Vajh=F8j?=]

Re: Regex to remove \t \r \n from string

Jon Skeet [C# MVP] wrote:

Arne Vajhøj <arne@vajhoej.d kwrote:

>>At what cost to readability though?

>Actually I think the regex code is more readable.

>
Well, it's interesting that your regex is "[\r\n\t]". I'm actually
slightly surprised this even works, as the \r, \n and \t are being
taken literally by the regex engine rather than having been escaped in
the normal way. I'd have expected "[\\r\\n\\t]" or @"[\r\n\t]" to make
it clear to the regex engine that you really meant the carriage return
etc to be part of the regex, and not incidental or for the sake of
readability (splitting the regex over several lines, as shown in
Jesse's example in another thread).
>
That extra level of escaping which is required in *some* cases (but
clearly not all) as well as having to understand the basic language of
regex in the first place is what makes it less readable in my opinion.

I just used the regex provided by Nicholas.

And yes there are different rules inside and outside character
classes.

And I can not see the readability problem. The intent of the
code is obvious.

You are not sure that it works correctly. But that can be
verified.

The Substring/IndexOf combo could be less obvious to read
and would still need to be verified that it works.

Arne

[Jon Skeet [C# MVP]]

Re: Regex to remove \t \r \n from string

Arne Vajhøj <arne@vajhoej.d kwrote:

That extra level of escaping which is required in *some* cases (but
clearly not all) as well as having to understand the basic language of
regex in the first place is what makes it less readable in my opinion.

I just used the regex provided by Nicholas.

And yes there are different rules inside and outside character
classes.

And I can not see the readability problem. The intent of the
code is obvious.

To you, possibly. To me, even - I've done just enough regex to work out
what it means, although I wouldn't necessarily say it's obvious. To
every maintenance engineer? Not necessarily.

You are not sure that it works correctly. But that can be
verified.

There are lots of things that can be verified, but which are still less
obvious than writing things in a simpler way.

The Substring/IndexOf combo could be less obvious to read
and would still need to be verified that it works.

There's no Substring/IndexOf to be done - just three calls to Replace.
It's blindingly obvious what *they* do.

--
Jon Skeet - <skeet@pobox.co m>
http://www.pobox.com/~skeet Blog: http://www.msmvps.com/jon.skeet
If replying to the group, please do not mail me too

[=?ISO-8859-1?Q?Arne_Vajh=F8j?=]

Re: Regex to remove \t \r \n from string

Jon Skeet [C# MVP] wrote:

Arne Vajhøj <arne@vajhoej.d kwrote:

>And I can not see the readability problem. The intent of the
>code is obvious.

>
To you, possibly. To me, even - I've done just enough regex to work out
what it means, although I wouldn't necessarily say it's obvious. To
every maintenance engineer? Not necessarily.

It is a feature in .NET - it is a feature in most programming
environments today.

If they don't know, then they should learn.

>The Substring/IndexOf combo could be less obvious to read
>and would still need to be verified that it works.

>
There's no Substring/IndexOf to be done - just three calls to Replace.
It's blindingly obvious what *they* do.

No Substring/IndexOf in this case. But often regex is replaced
with some string manipulation code in the worst tradition of
C str functions.

Arne

[Jon Skeet [C# MVP]]

Re: Regex to remove \t \r \n from string

Arne Vajhøj <arne@vajhoej.d kwrote:

To you, possibly. To me, even - I've done just enough regex to work out
what it means, although I wouldn't necessarily say it's obvious. To
every maintenance engineer? Not necessarily.

It is a feature in .NET - it is a feature in most programming
environments today.
>
If they don't know, then they should learn.

I'd rather not have to check the ins and outs of regular expressions
when there's a *very* simple alternative. It's so easy to go wrong with
regular expressions - I only use them when they provide a clear
benefit, which I don't believe they do in this case.

Just because you *can* do something with a regex doesn't mean you
*should*. I'm happy to go back and be really careful with regular
expressions when there's a good reason to use them, like validating
something which is genuinely a *pattern*, but I've seen enough people
get confused by them to be wary of them myself.

The Substring/IndexOf combo could be less obvious to read
and would still need to be verified that it works.

There's no Substring/IndexOf to be done - just three calls to Replace.
It's blindingly obvious what *they* do.

No Substring/IndexOf in this case. But often regex is replaced
with some string manipulation code in the worst tradition of
C str functions.

And likewise simple string manipulation code is replaced with a regex
for no reason whatsoever, sometimes introducing bugs at the same time.

--
Jon Skeet - <skeet@pobox.co m>
http://www.pobox.com/~skeet Blog: http://www.msmvps.com/jon.skeet
If replying to the group, please do not mail me too