JPA 中谓词的表达式解析器

Question

In a command line based application I want to offer basic support for filtering database results. For this, I need to parse user input. I want to support these tokens:

• boolean (logical) operators including and , or , not and parentheses
• variables, strings, dates and numbers
• comparison Operators: = , != , < , > , >= , <=

So an input string could look like this:

(age > 3 or name != 'tom') and city='london' and ((birthday=01.01.2010 or birthday >= 01.01.2013))

Basic stuff.

Question : Where do I get a (tokenizer and) parser for that?

• Are there libraries supporting these basics already?
• If not, I know that there are tools to generate Java classes for Tokenizers and Parsers. Do you have any recommendations?

I would then convert the results to JPA's Predicate s for the CriteriaQuery . First I thought, I could just offer JPQL in command line and let the user specify SQL based restrictions. But I don't want to let the user write conditions that might stop working eg when the database schema changes.

Answer 1

You are searching for some tokenizer and parser.

ANTLR is a lexer/parser generator, and the huge grammar repository that also, among others, contains SQL - even with dialects for some common databases at https://github.com/antlr/grammars-v4/tree/master/sql .

So if you still have that problem, you can generate the parser for free and just care about translating the abstract syntax tree into JPA criteria calls.