CODEWITHSUNDEEP
javaparsingstring
Richard Stokes
Richard Stokes

Reputation: 3552

Java String split not returning the right values

I'm trying to parse a txt file that represents a grammar to be used in a recursive descent parser. The txt file would look something like this:

SPRIME ::= Expr eof
Expr ::= Term Expr'
Expr' ::= + Term Expr' | - Term Expr' | e

To isolate the left hand side and split the right hand side into seperate production rules, I take each line and call: 

String[] firstSplit = line.split("::=");
String LHS = firstSplit[0];
String productionRules = firstSplit[1].split("|");

However, when I call the second split method, I am not returned an array of the Strings separated by the "|" character, but an array of each indiviudual character on the right hand side, including "|". So for instance, if I was parsing the Expr' rule and printed the productionRules array, it would look like this:

"+"
"Term"
"Expr'"
""
"|"

When what I really want should look like this:

Anyone have any ideas what I'm doing wrong?

Upvotes: 29

Views: 27081

Answers (4)

Alnitak
Alnitak

Reputation: 340055

The parameter to String.split() is a regular expression, and the vertical bar character is special.

Try escaping it with a backslash:

String productionRules = firstSplit[1].split("\\|");

NB: two backslashes are required, since the backslash character itself is special within string literals.

Upvotes: 71

Prince John Wesley
Prince John Wesley

Reputation: 63708

You need to escape pipe(|) symbol which is a regex OR operator . 

String productionRules = firstSplit[1].split("\\|");

or

String productionRules = firstSplit[1].split(Pattern.quote("|"));

Upvotes: 14

Jon Bright
Jon Bright

Reputation: 13748

The pipe character is the regex operator for "or". What you want is 

String productionRules = firstSplit[1].split("\\|");

which tells it to look for an actual pipe character.

Upvotes: 7

dcn
dcn

Reputation: 4469

Since split takes a regex as argument you have to escape all non-intended regex symbols.

Upvotes: 26

Related Questions