This morning, I was (and am still) having a problem getting some MULTILINE regular expression patterns to match properly. As such, I wanted to put a quick blog post together as a sanity check for myself. As I have blogged about before, when a Java regular expression is running in multiline mode (as denoted by the "?m" flag), the "^" and "$" expressions match the line start and line terminator (respectively) rather than the start and end of the source string. This allows us to do line-level pattern matching.
This is really useful; but, when we are running our regular expressions in multiline mode, we have to be aware that the new line and carriage return data is not matched inside of the ^ and $ expressions. As such, if we want to match a pattern across multiple lines in multiline mode, we have to define the new line and carriage return expressions explicitly in our pattern. To see this, take a look at the following demo:
- <!--- Store target data. --->
- <cfsavecontent variable="data">
- Create the Java pattern. Note that we are using the
- MULTILINE mode flag; this will allow ^ and $ to match the
- line delimiters.
- <cfset pattern = createObject( "java", "java.util.regex.Pattern" )
- "(?m)D.++$" &
- "(\r\n?|\n)" &
- "^.*+$" &
- "(\r\n?|\n)" &
- <!--- Get the matcher for our target text. --->
- <cfset matcher = pattern.matcher(
- javaCast( "string", trim( data ) )
- ) />
- <!--- Move to the first match. --->
- <cfset matcher.find() />
- <!--- Output the first match. --->
Here, we start out matching the letter "D" and then everything until the end of the line (remember that in Java, the "dot" does not match line terminators until DOTALL mode [aka. single-line mode] is turned on using ?s). We then add the line terminators to the pattern. We then match the entire next line and match its line terminators. We then match the line start followed by "G". When we run this code, we get the followingg output:
Anyway, nothing revolutionary going on here; like I said above, I mostly put this post together as a sanity check for myself to make sure that I really understood what was going on in Java's multiline mode.
Looking For A New Job?
- ColdFusion Developer at Clayton Homes
- Sr. Coldfusion GURU at eXcelaweb
- Senior Web Application Developer at Fidano
Did you look to see if it would return the end of line character if in single line mode? I have an idea how I would go about it and may try it later.
The way that expression is written, you know you'll have exactly one line between the DDD line and the line that starts with a G. Was that your intention? If you add a line or even a blank line no match is found. If you need more flexibility may I suggest something like:
a match is found even with:
Hmm, I think it would. When in single line mode, the dot will match line terminators, which means that .++ should possessively match ALL characters until it hits the end of the string.
Yes, I mean to only match a single line between the D and G (that was more along the lines of the use case I was trying to debug). But yes, the way you have it would be more flexible.
This is a good web log and i like it.
It will help people to learn ColdFusion.
I'm trying to match both of these: Single-line and Multi-line. Basically everything between the fail( and the );
Can you help me?
fail("MockFloorServiceDelegate timed out.");
fail( "Polygon failed verification: " +
errorCode + " lastVertex: " +
lastVertex.ordinal + ",
currentVertex: " +
I think "(?m)D(?s).*?^G" also works.