S, karthik (IE03x) am Donnerstag, 17. November 2005 09.54: > John and Jeff, > > Could you please explain the RE used. > > /^User:\s*(.+?)$(?=.*=([^[]+)\[)/sm; > > I understood that it looks for 'User:\s'. Then I am not able to figure > out the meaning of the remaining part of RE.
Hi karthik Let me try to explain (and maybe expect a better explanation from JWK): Because of $/ = "--\n"; which defines what is considered as record separator, the regex is every time applied to a whole data record consisting of 5 lines. /^User:\s* matches the beginning of the first record line and following space. (.+?)$ Then catches the user name into $1 (the $ matches the [real] line break). Now to (?=.*=([^[]+)\[) The (?=......) construct is a "zero-width positive look-ahead assertion". It matches nothing, just asserts that the construct content is following the position of the actual match position. (It's not 100% clear to me why JWK uses (?=) and not a simple (?:) - possibly he takes in account the case that there is no [2] data line) Now, left is, in (?=), .*=([^[]+)\[ which I split now: .* matches as much chars (including line breaks) as possible until to the... = last '=' char. This is found in the [2] data line. Left is now ([^[]+)\[ The () catches what's in it into $2. In the () is [^[]+ which is [^SOMETHING]+ . This matches one or more (due to +) char NOT (due to ^) containing SOMETHING. SOMETHING is '[' here. The match stops at \[ , that is, a literal '[' - in the first record example: [I:0 T:Str] The modifiers /m and /s are important for the interpretation of '$' and '.' in the regex. This and everything else is explained in detail in perldoc perlre === The regex won't do what you want if there are other lines after the [2] line. joe [...] > > > -- > > > User: honorhr40 > > > Got data from YY23 (md1=23,md2=24): > > > [0] xxxx=46[I:0 T:Int] > > > [1] pd=[I:0 T:Str] > > > [2] cosnm=normal[I:0 T:Str] > > > -- > > > User: lulu-918 > > > Got data from YY18 (md1=18,md2=19): > > > [0] xxxx=69[I:0 T:Int] > > > [1] pd=back.zzz.com[I:0 T:Str] > > > [2] cosnm=normal[I:0 T:Str] > > > > > > ...... -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] <http://learn.perl.org/> <http://learn.perl.org/first-response>
