Convert SGML to XML with regex?

Question

I want to convert SGML to XML with regrex. Like:

convert:

<a><ab><abc>111<abc2>222</ab></a>

to:

<a><ab><abc>111</abc><abc2>222</abc2></ab></a>

And I write the following code to do the conversion:

String a = "<a><ab><abc>abc<abc2>abc2</ab></a>";
a = a.replaceAll("<([^<>]+?)>([^<>]+?)<(?!/\\$1>)", "<$1>$2</$1><");
System.out.println(a);

However the result is not the expected one:

<a><ab><abc>111</abc><abc2>222</ab></a>

My question, is it possible to do the conversion with regex? If yes, What's the issue in my code?

Highly relevant: http://stackoverflow.com/a/1732454/1451599 – dirkk Jan 09 '15 at 14:48 — dirkk, Jan 09 '15 at 14:48
@dirkk I got cancer after reading that post. – Sid Zhang Jan 09 '15 at 14:57 — Sid Zhang, Jan 09 '15 at 14:57

Avinash Raj · Accepted Answer · 2015-01-09T15:12:33.257

2

Use the below regex

<(([^<>]+?)>)([^<>]+?)(?=<(?!\1))

And then replace the match with

<$1$3</$2>

String s = "<a><ab><abc>111<abc2>222</ab></a>";
System.out.println(s.replaceAll("<(([^<>]+?)>)([^<>]+?)(?=<(?!\\1))", "<$1$3</$2>"));

Output:

<a><ab><abc>111</abc><abc2>222</abc2></ab></a>

edited Jan 09 '15 at 15:12

answered Jan 09 '15 at 14:59

Avinash Raj

`\\$` matches a literal `$` symbol. add a extra capturing group to capture the trailing `>` symbol. – Avinash Raj Jan 09 '15 at 15:52
I got trouble again when dealing with multilines. https://regex101.com/r/cD1nC8/4 – Sid Zhang Jan 09 '15 at 17:00

1 Answers1