We Keep Coding

Tamil language full-word search with .NET Regex

I have a Grid filled with Tamil words and a search string. I need to implement a full-word search through the Grid records. I'm using .NET Regex class for that approach. It sounds pretty simple, what I used to do is:
string pattern = #"\b" + searchText + #"\b".
It works as expected in Latin languages but for Tamil, this expression returns strange results. I have read about Unicode characters in regular expressions but that doesn't seem quite helpful to me. What I probably need is to determine where is the word boundary found and why.
As an example:
For the "\bஅம்மா\b" pattern Regex found matches in
அம்மாவிடம் and அம்மாக்கள் records but not in the original அம்மா record.

The last char in "அம்மா" word is ‎0BBE TAMIL VOWEL SIGN AA and it is a combining mark (in regex, it can be matched with \p{M}).
As \b only matches between start/end of string and a word char or between a word and a non-word char, it won't match after the char and a non-word char.
Use a usual workaround in this case.
var pattern = $#"(?<!\w){searchText}(?!\w)";
See this regex demo.
Here, (?<!\w) fails the match if there is a word char before searchText and (?!\w) fails the match if there is a word char after the text to find. Note you may also use Regex.Escape(searchText) if the text can contains special regex chars.
Or, if you want to avoid matching when inside base letters/diacritics, use
var pattern = $#"(?<![\p{{L}}\p{{M}}]){searchText}(?![\p{{L}}\p{{M}}])";
See this regex demo.
The (?<![\p{L}\p{M}]) and (?![\p{L}\p{M}]) lookarounds work similarly as the ones above, just they fails the match if there is a letter or a combining mark on either side of the search phrase.

Regular expression that works on dots

I have this regular expression :
string[] values = Regex
.Matches(mystring4, #"([\w-[\d]][\w\s-[\d]]+)|([0-9]+)")
.OfType<Match>()
.Select(match => match.Value.Trim())
.ToArray();
This regular expression turns this string :
MY LIMITED COMPANY (52100000 / 58447000)";
To these strings :
MY LIMITED COMPANY - 52100000 - 58447000
This also works on non-English characters.
But there is one problem, when I have this string : MY. LIMITED. COMPANY. , it splits that too. I don't want that. I don't want that regular expression to work on dots. How can I do that? Thanks.

You may add the dot after each \w in your pattern, and I also suggest removing unnecessary ( and ):
string[] values = Regex
.Matches("MY. LIMITED. COMPANY. (52100000 / 58447000)", #"[\w.-[\d]][\w.\s-[\d]]+|[0-9]+")
.OfType<Match>()
.Select(match => match.Value.Trim())
.ToArray();
foreach (var s in values)
Console.WriteLine(s);
See the C# demo
Pattern:
[\w.-[\d]] - one Unicode letter or underscore ([\w-[\d]]) or a dot (.)
[\w.\s-[\d]]+ - 1 or more (due to + quantifier at the end) characters that are either Unicode letters or underscore, ., or whitespace (\s)
| - or
[0-9]+ - one or more ASCII-only digits

I'd simplify the expression. What if the names in the front include numbers? Not that my solution doesn't exactly mimic the original expression. It will allow numbers in the name part.
Let's start from the beginning:
To match words all you need is a sequence of word characters:
\w+
This will match any alphanumerical characters including underscores (_).
Considering you want the possibility of the word ending with a dot, you can add it and make it optional (one or zero matches):
\w+\.?
Note the escape to make it an actual character rather than a character class "any character".
To match another potential word following, we now simply duplicate this match, add a white space before, and once again make it optional using the * quantifier:
\w+\.?(?:\w+\.?)*
In case you haven't seen a group starting with ?: is a non-matching group. In essence this works like a usual group, but won't save a matching group in your results.
And that's it already. This pattern will split your demo string as expected. Of course there could be other possible characters not being covered by this.
You can see the results of this matching online here and also play around with it.
To test your regular expressions (and to learn them), I'd really recommend you using a tool such as http://regex101.com
It has an input mask allowing you to provide your pattern and your target string. On the right hand side it will first explain the pattern to you (to see if it's indeed what you had in mind) and below it will show all the groups matched. Just keep in mind it actually uses slightly different flavors of regular expressions, but this shouldn't matter for such simple patterns. (I'm not affiliated with that site, just consider it really useful.)
As an alternative, to directly use C#'s regex parser, you can also try this Regex Tester. This works in a similar way, although doesn't include any explanations, which might be not as ideal for someone just getting started.

C# Regular Expression: Search the first 3 letters of each name

Does anyone know how to say I can get a regex (C#) search of the first 3 letters of a full name?
Without the use of (.*)
I used (.**)but it scrolls the text far beyond the requested name, or
if it finds the first condition and after 100 words find the second condition he return a text that is not the look, so I have to limit in number of words.
Example: \s*(?:\s+\S+){0,2}\s*
I would like to ignore names with less than 3 characters if they exist in name.
Search any name that contains the first 3 characters that start with:
'Mar Jac Rey' (regex that performs search)
Should match:
Marck Jacobs L. S. Reynolds
Marcus Jacobine Reys
Maroon Jacqueline by Reyils
Can anyone help me?

The zero or more quantifier (*) is 'greedy' by default—that is, it will consume as many characters as possible in order to finding the remainder of the pattern. This is why Mar.*Jac will match the first Mar in the input and the last Jac and everything in between.
One potential solution is just to make your pattern 'non-greedy' (*?). This will make it consume as few characters as possible in order to match the remainder of the pattern.
Mar.*?Jac.*?Rey
However, this is not a great solution because it would still match the various name parts regardless of what other text appears in between—e.g. Marcus Jacobine Should Not Match Reys would be a valid match.
To allow only whitespace or at most 2 consecutive non-whitespace characters to appear between each name part, you'd have to get more fancy:
\bMar\w*(\s+\S{0,2})*\s+Jac\w*(\s+\S{0,2})*\s+Rey\w*
The pattern (\s+\S{0,2})*\s+ will match any number of non-whitespace characters containing at most two characters, each surrounded by whitespace. The \w* after each name part ensures that the entire name is included in that part of the match (you might want to use \S* instead here, but that's not entirely clear from your question). And I threw in a word boundary (\b) at the beginning to ensure that the match does not start in the middle of a 'word' (e.g. OMar would not match).

I think what you want is this regular expression to check if it is true and is case insensitive
#"^[Mar|Jac|Rey]{3}"
Less specific:
#"^[\w]{3}"

If you want to capture the first three letters of every words of at least three characters words you could use something like :
((?<name>[\w]{3})\w+)+
And enable ExplicitCapture when initializing your Regex.
It will return you a serie of Match named "name", each one of them is a result.
Code sample :
Regex regex = new Regex(#"((?<name>[\w]{3})\w+)+", RegexOptions.ExplicitCapture | RegexOptions.IgnoreCase);
var match = regex.Matches("Marck Jacobs L. S. Reynolds");
If you want capture also 3 characters words, you can replace the last "\w" by a space. In this case think to handle the last word of the phrase.

Regex to find N (specified number) words before and after a specific word

Can someone help me doing following search using Regex in C#
I need a regex that will give me specific number of words before and after a specific word that will include the search word itself. Please note I want to be able to specify how many words before and after I want. If the search word is in the start of the text then it returns just the N number of words after the search word and vice-versa if the search word is in the end of the text. Also the search should be able to provide multiple matches from the text that I can process.
For instance I want to get "2" words around the search word "example" so that will return multiple results in following the multiline text.
"example one in this lesson is an example summary that we are using as an example.
How many words are in this example? We only want the first few examples."
Returns following results
"example one in" (note how first search has no first 2 words)
"is an example summary that"
"as an example. How many"
"in this example? We only"
Thanks.

given your example the following regex would work
^example \w+ \w+|\w+ \w+ example[\.\?,]? \w+ \w+
Example C# implementation
string example = #"example one in this lesson is an example summary that we are using as an example. How many words are in this example? We only want the first few examples.";
string pattern = #"^example \w+ \w+|\w+ \w+ example[\.\?,]? \w+ \w+";
foreach (Match m in Regex.Matches(example, pattern))
{
Response.Write(m.Value + "<br/>");
}
use a regex debugger (such as debuggex) or regex tester (such as regexpal) to do any fine tuning to the above as you try this on additional test cases.

Regex - Get all words that are not wrapped with a "/"

Im really trying to learn regex so here it goes.
I would really like to get all words in a string which do not have a "/" on either side.
For example, I need to do this to:
"Hello Great /World/"
I need to have the results:
"Hello"
"Great"
is this possible in regex, if so, how do I do it? I think i would like the results to be stored in a string array :)
Thank you

Just use this regular expression \b(?<!/)\w+(?!/)\b:
var str = "Hello Great /World/ /I/ am great too";
var words = Regex.Matches(str, #"\b(?<!/)\w+(?!/)\b")
.Cast<Match>()
.Select(m=>m.Value)
.ToArray();
This will get you:
Hello
Great
am
great
too

var newstr = Regex.Replace("Hello Great /World/", #"/(\w+?)/", "");
If you realy want an array of strings
var words = Regex.Matches(newstr, #"\w+")
.Cast<Match>()
.Select(m => m.Value)
.ToArray();

I would first split the string into the array, then filter out matching words. This solution might also be cleaner than a big regexp, because you can spot the requirements for "word" and the filter better.
The big regexp solution would be something like word boundary - not a slash - many no-whitespaces - not a slash - word boundary.

I would use a regex replace to replace all /[a-zA-Z]/ with '' (nothing) then get all words

Try this one : (Click here for a demo)
(\s(?<!/)([A-Za-z]+)(?!/))|((?<!/)([A-Za-z]+)(?!/)\s)

Using this example excerpt:
The /character/ "_" (underscore/under-strike) can be /used/ in /variable/ names /in/ many /programming/ /languages/, while the /character/ "/" (slash/stroke/solidus) is typically not allowed.
...this expression matches any string of letters, numbers, underscores, or apostrophes (fairly typical idea of a "word" in English) that does not have a / character both before and after it - wrapped with a "/"
\b([\w']+)\b(?<=(?<!/)\1|\1(?!/))
...and is the purest form, using only one character class to define "word" characters. It matches the example as follows:
Matched Not Matched
------------- -------------
The character
_ used
underscore variable
under in
strike programming
can languages
be character
in stroke
names
many
while
the
slash
solidus
is
typically
not
allowed
If excluding /stroke/, is not desired, then adding a bit to the end limitation will allow it, depending upon how you want to define the beginning of a "next" word:
\b([\w']+)\b(?<=(?<!/)\1|\1(?!/([^\w]))).
changes (?!/) to (?!/([^\w])), which allows /something/ if it does have a letter, number, or underscore immediately after it. This would move stroke from the "Not Matched" to the "Matched" list, above.
note: \w matches uppercase or lowercase letters, numbers and the underscore character
If you want to alter your concept for "word" from the above, simply exchange the characters and shorthand character classes contained in the [\w'] part of the expression to something like [a-zA-Z'] to exclude digits or [\w'-] to include hyphens, which would capture under-strike as a single match, rather than two separate matches:
\b([\w'-]+)\b(?<=(?<!/)\1|\1(?!/([^\w])))
IMPORTANT ALTERNATIVE!!! (I think)
I just thought of an alternative to Matching any words that are not wrapped with / symbols: simply consume all of these symbols and words that are surrounded in them (splitting). This has a few benefits: no lookaround means this could be used in more contexts (JavaScript does not support lookbehind and some flavors of regex don't support lookaround at all) while increasing efficiency; also, using a split expression means a direct result of a String array:
string input = "The /character/ "_" (underscore/under-strike) can be..."; //etc...
string[] resultsArray = Regex.Split(input, #"([^\w'-]+?(/[\w]+/)?)+");
voila!