RegEx: Is there a way to say "string NOT found" ?

jdruin · 12-03-2008, 08:17 AM

For example:

I would like to make sure that the string "9999" is NOT found in the following list of four character strings:

0401|0417|2720|2801|2809|2904|4192|5206|5271|0901|2304|2402|3802

I tried:

^?!([0401|0417|2720|2801|2809|2904|4192|5206|5271|0901|2304|2402|3802])$
^([0401|0417|2720|2801|2809|2904|4192|5206|5271|0901|2304|2402|3802]){0}$
^(^[0401|0417|2720|2801|2809|2904|4192|5206|5271|0901|2304|2402|3802])$
^(^[0401|0417|2720|2801|2809|2904|4192|5206|5271|0901|2304|2402|3802])$

So far I have not been able to successfully verify that the string is not matched by any of the options in the list. Usually Google answers all my regex questions but no dice this time.

ghostdog74 · 12-03-2008, 08:29 AM

i couldn't understand what you want. Show a sample input file, and describe how you want your output to look like.

jdruin · 12-03-2008, 08:50 AM

Thanks for the reply.

Here is my input to the IsMatch function/method:

"9999"

My string to match is, for example this list:

0401|0417|2720

Since the string "9999" does not match "0401" or "0417" or "2720", I would like the function IsMatch() to return false.

For example, if I ran this code

Code:

boolAnswer = Regex.IsMatch("9999", ^?!([0401|0417|2720])$)

I would like the value of boolAnswer to be FALSE because "9999" does not match any of those 3 choices.

Hope this helps.

jan61 · 12-03-2008, 05:32 PM

Moin,

why not simply

Code:

boolAnswer = ! Regex.IsMatch("9999", ^([0401|0417|2720])$)

?

(or "not" - I don't know, which language you are using).

Jan

Telemachos · 12-03-2008, 06:28 PM

Quote:

Originally Posted by jan61

Moin,

why not simply

Code:

boolAnswer = ! Regex.IsMatch("9999", ^([0401|0417|2720])$)

?

(or "not" - I don't know, which language you are using).

My first thought was JavaScript, but Google seems to suggest .NET (?). This site has some discussion of how to use IsMatch in conditions: http://www.dotnetcoders.com/web/Lear...gexobject.aspx

sundialsvcs · 12-03-2008, 06:30 PM

It's quite simple: you're looking for the situation where a regex, which searches for "9999" (surrounded on either side by non-digits) anywhere in the string, does not match.

jdruin · 12-04-2008, 07:27 AM

Hello Telemachos,

The reason why is that the RegEx expressions are stored in a configuration file (an SQL Server database actually). The evaluation statement is in a compiled program. By altering the configuration file (table), we can then make the statement either positive (look for a match) or negative (look where NO match) without recompiling the code. The code is not static. It is meant to handle both positive and negative statement with equal dexterity. So rather than have the negation on the IsMatch() method, we want the neagation to be in the RegEx itself.

Hope this makes sense.

We have already considered an alternative. We will add one extra column to the configuration which will indicate whether we want a match or NO match. This would allow the RegEx to always be affirmative but still allow configurable negative statements. This was a little more work and complexity though, so before going down that path, I wanted to make sure there was no way to negate string matches in RegEx. If no one on LinuxQuestions can answer, I assume I have done due diligence.

jdruin · 12-04-2008, 07:29 AM

Hello sundialsvcs,

Thanks for the reply. Unfortunately I already know that part. What I really need know is how to accomplish this. I was looking for someone to provide an example.

jdruin · 12-04-2008, 07:31 AM

Hello jan61,

Please see my above response to tele. Good idea but we want configurability outside of the compilation. We are using VB.NET as the language, although as you implied, this is not an important detail.

Telemachos · 12-04-2008, 09:49 AM

The site I linked to above suggests something like this:

Code:

   bool IsMatch(string input); 

   if (rex1.IsMatch("My name is John") == true)  
       Console.WriteLine("Found a match.");

It should be easy enough to swith the if section to if (rex1.IsMatch("foobar") == false. But I'm uncertain about the details of this language.

jdruin · 12-04-2008, 12:41 PM

Well in VB.NET it would be more like this:

If Not Regex.IsMatch("9999", port_regex_pattern) Then
'do stuff here
End If

but I see what your saying. This wont work in my case because I want the RegEx expressions stored in a configuration file and I want the RegEx to carry the negation, not the VB code. I ended up going with the workaround described above because I am satisfied that the RegEx itself cannot be negated, so I am using one extra database column to indicate negativity. If that column is TRUE then the VB statement like you described executes. Otherwise an affirmative statement executes.

In VB this looks something like:

If _
(Not match_not_found_flag AndAlso Regex.IsMatch("9999", database_regex_pattern)) _
Or _
(match_not_found_flag AndAlso Not Regex.IsMatch("9999", database_regex_pattern)) _
Then
'do stuff
End If

The real statement contains variable search_pattern and regex_pattern. I put the 9999 in just to make it more readable.