com.continuent.tungsten.common.parsing.bytes
Class MySQLStatementTranslator

java.lang.Object
  com.continuent.tungsten.common.parsing.bytes.MySQLStatementTranslator

public class MySQLStatementTranslator
extends java.lang.Object

Utility class to translate MySQL statements from a native charset to Java Unicode strings accounting for introducers for binary and alternative character sets. Syntax for MySQL strings is described in MySQL online documentation. A typical example looks like the following:

 INSERT INTO `storage'_binary'` VALUES (25, 'col_binary', _binary'\0\0\0\0')
 

This string illustrates some of the potential ambiguities in string translation. To avoid confusion we implement a full tokenizer that ignores data embedded in normal strings or comments. We thus translate the preceding string into the following, where binary data are replaced by a hexadecimal string format:

 INSERT INTO `storage'_binary'` VALUES (25, 'col_binary', _binary x'00000000')
 

The translation is based on state machines according to the following principles.

• Binary and alternative charset strings denoted by introducers of the form _<introducer>'value' or _<introducer>"value" are converted to hex strings that can translate safely to Unicode.
• All bytes outside of such strings are translated using the character set assigned when creating the MySQLStatementTranslator instance.
• Values within ordinary strings starting with backtick (`), single, and double quotes excluded from parsing for introducers. The same applies to values within comments.

Performance is an important consideration in the translation algorithm as binary strings in particular are potentially quite large. The parsing and translation processing uses pointers into the byte string to minimize object creation. The translation values for byte strings are pre-computed strings. The performance overhead of parsing + translation is about 10% over unparsed string translation.

Finally, it should be noted that translation into the safe hex format doubles the size of binary strings. Users should expect to double memory allocations accordingly, including MySQL specific settings like max_packet_size, which sets the maximum size of a single client request.

Version:

1.0

Author:

Robert Hodges

Constructor Summary

MySQLStatementTranslator(java.lang.String charset)

Method Summary

java.lang.String

toJavaString(byte[] bytes, int offset, int length)

Methods inherited from class java.lang.Object

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

Constructor Detail

MySQLStatementTranslator

public MySQLStatementTranslator(java.lang.String charset)
                         throws java.io.UnsupportedEncodingException

Throws:

java.io.UnsupportedEncodingException

Method Detail

toJavaString

public java.lang.String toJavaString(byte[] bytes,
                                     int offset,
                                     int length)
                              throws java.io.UnsupportedEncodingException

Throws:

java.io.UnsupportedEncodingException