Remove doctype containing entity from xml using java

nithin :

I'm trying to process an xml, before that i need to remove the doctype and entity declaration from the input xml.

I'm using the below code to remove the doctype and entity:

fileContent = fileContent.replaceAll("<!ENTITY ((.|\n|\r)*?)\">", "");
fileContent = fileContent.replaceAll("<!DOCTYPE((.|\n|\r)*?)>", "");

This removes the entity and then the doctype. This works fine if the xml contains below doctype declarations in the xml:

<!DOCTYPE ichicsr SYSTEM "http://www.w3.org/TR/html4/frameset.dtd">

<!DOCTYPE ichicsr SYSTEM "D:\UPGRADE\NTServices\Server\\Xml21.dtd"
[<!ENTITY % entitydoc SYSTEM "D:\UPGRADE\NTServices\Server\\latin-entities.dtd"> %entitydoc;]>

But if I have the doctype as given below, it doesn't work and the root tag in the xml get stripped off:

<!DOCTYPE ichicsr SYSTEM "D:\UPGRADE\NTServices\Server\\Xml21.dtd" 
[<!ENTITY % entitydoc SYSTEM 'D:\UPGRADE\NTServices\Server\\Xml21.dtd'>
]> 

Please let me know if the regular expression I'm using is incorrect or any other action needs to be taken.

Wiktor Stribiżew :

Your approach does not work because you have " required before final > in the ENTITIY regex. You may just replace \" with ['\"] there.

Besides, never use (.|\n|\r)*? in any regex since it is a performance killer. Instead, use .*? with Pattern.DOTALL (or inline (?s) variant), or at least [\s\S]*?.

However, there is a better way: merge the two regexps into one:

fileContent = fileContent.replaceAll("(?i)<!DOCTYPE[^<>]*(?:<!ENTITY[^<>]*>[^<>]*)?>", "");

See the regex demo.

Details

• (?i) - case insensitive Pattern.CASE_INSENSITIVE inline modifier
• <!DOCTYPE - literal text
• [^<>]* - 0+ chars other than < and >
• (?:<!ENTITY[^<>]*>[^<>]*)? - an optional occurrence of
  • <!ENTITY
  • [^<>]* - 0+ chars other than < and >
  • > - a > char
  • [^<>]* - 0+ chars other than < and >
• > - a > char.