io.debezium.relational.ddl

Class DataTypeParser

java.lang.Object

io.debezium.relational.ddl.DataTypeParser

@ThreadSafe
public class DataTypeParser
extends Object

A parser of SQL data types. Callers set up a parser and register patterns that describe the possible lists of data type tokens, and then repeatedly parse streams of tokens looking for matches.

This is typically used within a LegacyDdlParser implementation to parse and identify data types appearing within a stream of DDL content.

Author:

Randall Hauch

See Also:

LegacyDdlParser

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
protected static class	DataTypeParser.ErrorCollector

Field Summary

Fields

Modifier and Type	Field and Description
private DataTypeGrammarParser	parser
private Map<String,Collection<DataTypeGrammarParser.DataTypePattern>>	patterns

Constructor Summary

Constructors

Constructor and Description
DataTypeParser() Create an empty data type parser with no data types registered.

Method Summary

Modifier and Type	Method and Description
DataType	parse(TokenStream stream, Consumer<Collection<ParsingException>> errorHandler) Examine the stream starting at its current position for a matching data type.
DataTypeParser	register(int jdbcType, String grammar) Register data type that may not contain a length/precision or scale.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

patterns

private final Map<String,Collection<DataTypeGrammarParser.DataTypePattern>> patterns

parser

private final DataTypeGrammarParser parser

Constructor Detail

DataTypeParser

public DataTypeParser()

Create an empty data type parser with no data types registered.

Method Detail

register

public DataTypeParser register(int jdbcType,
                               String grammar)

Register data type that may not contain a length/precision or scale.

The order that grammars are registered is the same order that they are evaluated by the resulting parser. However, when multiple grammars match the same input tokens, the grammar that successfully consumes the most input tokens will be selected.

Parameters:

jdbcType - the best JDBC type for the data type

grammar - the grammar the defines the data type format; may not be null

Returns:

this parser so callers can chain methods; never null

parse

public DataType parse(TokenStream stream,
                      Consumer<Collection<ParsingException>> errorHandler)

Examine the stream starting at its current position for a matching data type. If this method finds a matching data type, it will consume the stream of all tokens that make up the data type. However, if no data type is found, the stream is left unchanged.

This method looks for data types that match those registered patterns. This method also looks for multi-dimensional arrays, where any registered data type pattern is followed by one or more array dimensions of the form [n], where n is the integer dimension.

Sometimes, a data type matches one of the registered patterns but it contains a malformed length/precision, scale, and/or array dimensions. These parsing exceptions can be reported back to the caller via the supplied errorHandler, although these errors are only reported when no data type is found. This is often useful when the caller expects to find a data type, but no such data type can be found due to one or more parsing exceptions. When this happens, the method calls errorHandler with all of the ParsingExceptions and then returns null.

Parameters:

stream - the stream of tokens; may not be null

errorHandler - a function that should be called when no data type was found because at least one ParsingException occurred with the length/precision, scale, and/or array dimensions; may be null

Returns:

the data type if one was found, or null if none were found