RegEx to Remove All Blank Lines

[JRL]

I need a way to remove all blank lines from a text list. The following seems to work but is not extremely elegant. Borrowed the Let>pattern=^[\s]+|[\s]+$ line from an adroege post. Figured out the rest from looking on the internet. I found many instances of code that supposedly removes all blank lines but could not find anything I could make work.

Anyone have a single line solution?

Code: Select all

Let>text=%crlf%lineone%crlf%linetwoo%crlf%%crlf%%crlf%linethreee%crlf%%crlf%Linefourrrr%crlf%

Let>pattern=^[\s]+|[\s]+$
RegEx>pattern,text,0,matches,num,1,,text

Let>pattern=[\r\n]+
RegEx>pattern,text,0,matches,num,1,_zKz_,text

StringReplace>text,_zKz_,crlf,text

MessageModal>text

[jpuziano]

Hi JRL,

A single line solution *AND* more elegant that what you have there huh...

Not asking for much eh?

How many points would something like that be worth to you I wonder?

Heh...

Oh alright alright... here's the magic one-liner:

RegEx>(?m)^$\x0D\x0A,text,0,matches,num,1,,text

And here it is in a demo macro that shows your text before and after:

Code: Select all

Let>text=%crlf%lineone%crlf%linetwoo%crlf%%crlf%%crlf%linethreee%crlf%%crlf%Linefourrrr%crlf%

MessageModal>text

//discard all blank lines except for one CRLF at the very end if it exists
Let>pattern=(?m)^$\x0D\x0A
RegEx>pattern,text,0,matches,num,1,,text

MessageModal>text

Enjoy!

[JRL]

Very nice... almost.

What about that last line still being blank? The last line is part of "all".

[jpuziano]

Hi JRL,

Hmm... all the other lines get to be terminated by CRLF... but you want that last line to be different huh...

Alrighty then, the pattern has to get a bit more complex, try this...

RegEx>(?m)(^$\x0D\x0A)|(?-m)(\x0D\x0A)+$,text,0,matches,num,1,,text

Here's the demo again...

Code: Select all

Let>text=%crlf%%crlf%%crlf%lineone%crlf%linetwo%crlf%%crlf%%crlf%linethree%crlf%%crlf%Linefour%crlf%%crlf%%crlf%

MDL>text

//discard all blank lines... including any CRLF chars at the very end! 
Let>pattern=(?m)(^$\x0D\x0A)|(?-m)(\x0D\x0A)+$
RegEx>pattern,text,0,matches,num,1,,text

MDL>text

You'll notice I threw in a few more blank lines in the data at the end and the beginning to torture test it...

I think you'll find now that all blank lines are removed... including the final CRLF terminating the last line of the text blob.

Enjoy!

[JRL]

Excellent!

You have now provided two fine examples of Regex. One that removes ALL blank lines including the last CRLF.

RegEx>(?m)(^$\x0D\x0A)|(?-m)(\x0D\x0A)+$,text,0,matches,num,1,,text

And one that removes blank lines but leaves the last CRLF.

RegEx>(?m)^$\x0D\x0A,text,0,matches,num,1,,text



I have one question. Is there an advantage using "\x0D\x0A" over using "\r\n" as a representation for CRLF?

[jpuziano]

Hi JRL,

Interesting question. I feel the RegEx is taking hold of you...

Well... \r\n is less characters so many would prefer it for that alone.

However I like representing CRLF as \x0D\x0A in a RegEx pattern because:

- Representing a control char using its hex value is more flexible

- Example: I copy some example code that uses \x0D\x0A

- Now I need to target other control chars, simple, just change the numbers

- Had I just used \r\n I may have forgotten the hex format and would have had to look that up again

- Final reason, in a big long ugly RegEx, \x0D\x0A just pops better, just a little faster/easier to understand when looking at it, in my opinion... but it all comes down to personal choice.

Use whatever works best for you and may the RegEx be with you!

[Phil Pendlebury]

Excellent stuff guys.