正则表达式：匹配带有引号的SQL PRINT块

Question

I have the following text I am trying match using regular expressions:

PRINT CONVERT(NVARCHAR, CURRENT_TIMESTAMP, 111) + ' ' + CONVERT(NVARCHAR, CURRENT_TIMESTAMP, 108) + ' -Test Mode : ' + (CASE WHEN @turbo_mode_ind = 1 THEN 'some text ''test'' some more text.' ELSE 'and even more text ''temp'' when will it stop ?' END)

PRINT 'text don''t text'

PRINT 'text ''test2'' text'

What I want to match is:

PRINT CONVERT(NVARCHAR, CURRENT_TIMESTAMP, 111) + ' ' + CONVERT(NVARCHAR, CURRENT_TIMESTAMP, 108) + ' -Test Mode : ' + (CASE WHEN @turbo_mode_ind = 1 THEN 'some text ''test''

PRINT 'text ''test2''

So basically I want to match:

• starting at PRINT
• each char that comes after PRINT (.*)
• inclusive line-breaks (don't stop at line-breaks)
• with \\'{2}\\w+\\'{2} at the end of the match
• non-greedy (.*?)
• AND no empty line(s) between PRINT and \\'{2}\\w+\\'{2}

I have already compsed this, but it still matches empty line(s):

PRINT.*?\\'{2}\\w+\\'{2}(?!\\n\\s*\\n)

Answer 1

Edit after comment:

Looking at the requirements again I could not come up with a single regex solution quickly. In your comments you mention that you are using C#.

A possible solution would therfore be to first split the string at blank lines and then extracting the text.

Something like this:

string pattern = @"^$";

foreach (string result in Regex.Split(input, pattern, RegexOptions.Multiline) 
{
    Regex rxFindSql = Regex(@"PRINT.*?\'{2}\w+?\'{2}", RegexOptions.SingleLine)       
    MatchCollection matches = rxFindSql.Matches(result);
}  

This should do the trick but I did not test the code.

I hope this helps.