What Is A Parser? Definition, Types And Examples - TechTarget

What are the main types of parsers?

When a software language is created, its creators must specify a set of rules. These rules provide the grammar needed to construct valid statements in the language.

The following is a set of grammatical rules for a simple fictional language that only contains a few words:

<sentence> ::= <subject> <verb> <object> <subject> ::= <article> <noun> <article> ::= the | a <noun> ::= dog | cat | person <verb> ::= pets | fed <object> ::= <article> <noun>

In this language, a sentence must contain a subject, verb and noun in that order, and specific words are matched to the parts of speech. A subject is an article followed by a noun. A noun can be one of the following three words: dog, cat or person. And a verb can only be pets or fed.

Parsing checks a statement that a user provides as input against these rules to prove that the statement is valid. Different parsing algorithms check in different orders. There are two main types of parsers:

• Top-down parsers. These start with a rule at the top, such as <sentence> ::= <subject> <verb> <object>. Given the input string "The person fed a cat," the parser would look at the first rule, and work its way down all the rules checking to make sure they are correct. In this case, the first word is a <subject>, it follows the subject rule, and the parser will continue reading the sentence looking for a <verb>.
• Bottom-up parsers. These start with the rule at the bottom. In this case, the parser would look for an <object> first, then look for a <verb> next and so on.

More simply put, top-down parsers begin their work at the start symbol of the grammar at the top of the parse tree. They then work their way down from the rule to the sentence. Bottom-up parsers work their way up from the sentence to the rule.

Beyond these types, it's important to know the two types of derivation. Derivation is the order in which the grammar reconciles the input string. They are:

• LL parsers. These parse input from left to right using leftmost derivation to match the rules in the grammar to the input. This process derives a string that validates the input by expanding the leftmost element of the parse tree.
• LR parsers. These parse input from left to right using rightmost derivation. This process derives a string by expanding the rightmost element of the parse tree.

In addition, there are other types of parsers, including the following:

• Recursive descent parsers. Recursive descent parsers backtrack after each decision point to double-check accuracy. Recursive descent parsers use top-down parsing.
• Earley parsers. These parse all context-free grammars, unlike LL and LR parsers. Most real-world programming languages do not use context-free grammars.
• Shift-reduce parsers. These shift and reduce an input string. At each stage in the string, they reduce the word to a grammar rule. This approach reduces the string until it has been completely checked.