Java Regex xml parsing

Question

I'm trying to find a tag from begin to end in xml and replace it with a blank. A sample xml is like this

<lins>
  <lin index="1"> ...<feature>Something</feature>... </lin>
  <lin index="2">...<feature>Something</feature>... </lin>
  <lin index="3">...<feature>Something</feature>....</lin>

  <lin index="1">...<feature>Icom</feature>... </lin>
  <lin index="2">...<feature>Icom</feature>... </lin>
<lins>

I need to remove <lin> to </lin> when ever I see Icom in between

<lin\s(.+?Icom.+?)+</lin> is removing all lin items since it matches the first begin <lin> tag and the last lin end tag. Greatly appreciated if you can suggest a way to do this. Also I can not use xml parsers in my situation.

Please re-write what do you need to remove. It is not clear. — AlexR
– AlexR, Commented Dec 21, 2011 at 14:21
I'm trying to find a tag from begin to end in xml and replace it with a blank. A sample xml is in the question. In the above example I need to find and remove <lin index="1">... <feature>Icom</feature>... </lin> <lin index="2">...<feature>Icom</feature>... </lin> The rule is when ever I see Icom remove <lin> to </lin> The regex I used is removing all lin tags. <lin (.+?Icom.+?)+</lin> — user1110005
– user1110005, Commented Dec 21, 2011 at 14:23
And what situation is that, virtually all platforms known to human have XML parsers. — ThomasRS
– ThomasRS, Commented Dec 21, 2011 at 14:35

Tim Pietzcker · Accepted Answer · 2011-12-21 14:27:47Z

4

String result = subject.replaceAll("(?s)<lin\\b(?:(?!</lin).)*Icom(?:(?!</lin).)*</lin>", "");

should do this, unless you have <lin> tags nested into each other (or inside comments/strings).

Explanation:

<lin\b              # Match <lin (but not link or linen)
(?:                 # Match...
 (?!</lin)          # as long as we're not at a closing tag
 .                  # any character
)*                  # any number of times.
Icom                # Match Icom
(?:(?!</lin).)*     # (as above:) Match any character except closing tag
</lin>              # Match closing tag

answered Dec 21, 2011 at 14:27

Tim Pietzcker

337k59 gold badges520 silver badges572 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

shift66 Over a year ago

it will not match if there's an inner <lin></lin> tag

Tim Pietzcker Over a year ago

@Ademiban: That's exactly what I wrote. Also, it will match, but it'll match an incorrect part of the string (which is probably even worse, which is why I wrote that <lin> tags must not be nested).

Chris Aldrich Over a year ago

@user1110005 would be great to accept this answer and put a check next to it.

shift66 · Accepted Answer · 2011-12-21 14:28:47Z

0

you cant do it with regexp.
For this example:

<tag>
    <tag> something </tag>
</tag>

<tag>
</tag>

If you use "<tag>(.*)</tag>" regexp, your group will be this:

    <tag> something </tag>
</tag>

<tag>

and if you use "<tag>(.*?)</tag>" regexp, your group will be this:

    <tag> something

You should use something like stack to get the ending of started tag.

answered Dec 21, 2011 at 14:28

shift66

12k13 gold badges54 silver badges83 bronze badges

1 Comment

Tim Pietzcker Over a year ago

Your points are valid (about nested strings), however your examples are both wrong.

ozoli · Accepted Answer · 2011-12-21 15:17:23Z

0

I think you need to add more groups to the regexp.

Add a group for the precondition to start checking for ex (

Then a group for the stuff inbetween, a group for Icom etc.

So off the top of my head my RegEx would look like:

(<lin\ index\=)(\w+Icom\w+)(\<\/lin>)

Note the escaping might be slightly off, but in essence you need more groups and some less eager matchers.

answered Dec 21, 2011 at 15:17

ozoli

1,4541 gold badge18 silver badges26 bronze badges

Collectives™ on Stack Overflow

Java Regex xml parsing

3 Answers 3

3 Comments

1 Comment

Comments

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

3 Comments

1 Comment

Comments

Related