Sorry if this has been asked, my search brought up many off topic posts.
I’m trying to convert wildcards from a user defined search string (wildcard is “*”) to PostgreSQL like wildcard “%”.
I’d like to handle escaping so that "%" => "\%" and "\*" => "*"
I know I could replace \* with something else prior to replacing * and then swap it back, but I’d prefer not to and instead only convert * using a pattern that selects it when not proceeded by \.
String convertWildcard(String like)
{
like = like.replaceAll("%", "\\%");
like = like.replaceAll("\\*", "%");
return like;
}
Assert.assertEquals("%", convertWildcard("*"));
Assert.assertEquals("\\%", convertWildcard("%"));
Assert.assertEquals("*", convertWildcard("\\*")); // FAIL
Assert.assertEquals("a%b", convertWildcard("a*b"));
Assert.assertEquals("a\\%b", convertWildcard("a%b"));
Assert.assertEquals("a*b", convertWildcard("a\\*b")); // FAIL
Ideas welcome.
EDIT
To clarify,
I want a method that makes 1 or more String.replaceAll calls to convert a string so that
- occurrences of
%become\%, - occurrences of
*become%and - occurrences of
\*become*.
The best answer will use the least calls to String.replaceAll.
You need what is called “Negative Lookbehind”. To select all
%not preceded by\:(?<!\\)%(pure regex expression)to use it in Java you need to add some masking:
string.replaceAll("(?<!\\\\)%", "*");