Community

I'm editing a text file - all lines end in 0D0A. Snippit of file is found below. I'm using Regular Expressions in TP 7.0.9.

I'm looking for lines with "m." followed by text and ";" in them. My find expression is ([^m\.]*)m\.([^;]*);(.*)\R (I've tried with and without "\R")

At line7, I expected the find to fail (no ";" before EOL) and move on - but it 'sees' the "m." and finds ";" in the next line! Obviously, I'm missing something. Can anyone identify?

sp:;Dorothy;b.Apr 1812;m.8 Sep 1828;d.4 May 1897)
2.;John;b.Abt 1831;d.Apr/Jun 1898)
sp:;Elizabeth;b.Abt 1830;m.10 Dec 1854;d.1907)
3.;Isabella;Wigham;b.Abt 1856)
4.;Margaret;Ann Hughes;b.Abt 1893)
4.;Harriett;Hughes;b.Abt 1897)
sp:;John;Lawson;m.Jan/Mar 1875)
4.;Elizabeth;Lawson;b.Abt 1876)
4.;Sarah;J. Lawson;b.Abt 1877)
3.;Sarah;Jane Wigham;b.Oct/Dec 1858;d.1941)
sp:;George;William Stoker;b.Abt 1857;m.Jan/Mar 1876;d.1891/1901)
4.;Elizabeth;Stoker;b.Abt 1878)

[^m\.] (or [^m.] ) matches any character, including a newline, that isn't m or . (dot). That's why the matches can extend over newlines.

If you really need to capture the text matching the parenthesised parts of your regex then
([^m.\r\n]*)m\.([^;\r\n]*);(.*)
or
([^m.\r\n]*)m\.(.*?);(.*)
might do the trick. Otherwise a simple
m\..*;
might find the lines you're interested in.

Thanks for the help, Ben!

Can I ask - why use both "\r" and "\n"? Both are listed as matching "any line ending character sequence " - is there something unwritten about their combination?

In a character set ([...]) \r means just a carriage return and \n means just a line feed. If your file has unix line endings you can leave out the \r.

This should be made clearer in the help.

In fact you can leave out the \r even if your file has Windows line endings, as a \r never occurs without a \n.