ANTLR function to retrieve text from a single node in a parse tree

Question

I want to retrieve text from a parse tree's node but when i use getText(), it retrieves all the text from that node's children. I have searched but i don't find any ANTLR function that can retrieve only the node's text or a function to delete a node's children (so that i can use getText() on that node once i have deleted his children). Any help is welcomed.

Answer 1

That would be because a “node” (and particularly, it’s text) includes it’s children. If you define it as the text that’s not part of a child, you’re left with just to tokens (leaf nodes) in that parser rule, and that’s probably a bit less than useful (some text with holes where other rules kick in).

Maybe you have a situation where something like this would be useful. If so, you might want to elaborate. We might, then propose something more specific.

---

To address your comment, here's the implementation of getText() for RuleContext

    public String getText() {
        if (getChildCount() == 0) {
            return "";
        }

        StringBuilder builder = new StringBuilder();
        for (int i = 0; i < getChildCount(); i++) {
            builder.append(getChild(i).getText());
        }

        return builder.toString();
    }

So, technically, the text of a node is nothing but the concatenation of the texts of it's children. By that definition, there'd be no text of a parent that's not child text.

Perhaps the definition of children nodes you mean is "children nodes that are RuleNodes" (as opposed to ErrorNodes or TerminalNodes). You could implement a getText() that ignored ParserRuleContext, but that would return a rather odd concatenation of tokens that could have occurred before, after, or between, your RuleNodes.

I don't think there's any generalized solution to what you're asking for. You'll need to pull out the specific text you want for parent nodes context by context.