I can't resist posting a regex solution. Have to admit I'm not entirely clear about what the OP wants to achieve, but I think it's to get hold of a random URL from an xml file so here goes. lcSQL_GetRegexMatches is a library function I use all the time to find all the matches of a regex in a string.
function getRandomURL pxml constant kRegex="https://www.domainB.com/show.php\?l=0&u=156&id=\d*"; --Need to escape the "?" local tMatches put lcSQL_GetRegexMatches(pxml,kRegex) into tMatches if tMatches is not empty then get the random of the number of lines of tMatches return char (item 1 of line it of tMatches) to (item 2 of line it of tMatches) of pxml else return empty end if end getRandomURL function lcSQL_GetRegexMatches ptext,pregex /* Finds all the matches of pregex in ptext and returns a list of the start,end char positions */ local tOffset,tStart,tEnd,tList,tLength --Make sure the whole regex string is a capture group if not (pregex begins with "(" and pregex ends with ")") then put "("before pregex put ")" after pregex end if --Make sure the regex is valid try get matchChunk(ptext,pregex,tStart,tEnd) catch e answer "Invalid regex expression:" return empty end try put 1 into tOffset put the length of ptext into tLength repeat if matchChunk(char tOffset to tLength of ptext,pregex,tStart,tEnd) then put tStart+tOffset-1, tEnd+tOffset-1 & return after tList add tEnd to tOffset else exit repeat end if end repeat return tList end lcSQL_GetRegexMatches Of course, if you knew in advance how may URLs were in the xml, you could do the whole thing in one call to MatchChunk with the appropriate number of start/end variables. Which reminds me that it would be so nice to have the matchChunkvariable be a list as in the above, or a numerically-keyed array with each key containing the comma-separated start and end positions, or just two variables with the first one holding a comma-separted list of the start positions and the second one holding a comma-separated list of the end positions. That would all cause backwards compatibility issues so I guess there'd have to be a new function, matchAllChunks maybe. Or you could ask Thierry when his new regex library will be ready. Pete lcSQL Software <http://www.lcsql.com> Home of lcStackBrowser <http://www.lcsql.com/lcstackbrowser.html> and SQLiteAdmin <http://www.lcsql.com/sqliteadmin.html> On Fri, Jan 16, 2015 at 6:28 PM, Kay C Lan <lan.kc.macm...@gmail.com> wrote: > If the above format is EXACTLY as the data appears, i.e. > > <![CDATA[ --this is exact > --random URL you want to capture > ]]> --this is exact > > then this is a very easy to parse. If the line above and below the random > URL is not as you've stated then there are other ways to skin the cat. > > on mouseUp > put URL tPathToXMLFile into tData > put false into tParse > put empty into tOutput > repeat for each line tLine in tData > switch > case (tLine = "]]>") > put false into tParse > break > case (tLine = "<