To write a full-text parser plugin, include the following header file in the plugin source file. Other MySQL or general header files might also be needed, depending on the plugin capabilities and requirements.

#include <mysql/plugin.h>

plugin.h defines the MYSQL_FTPARSER_PLUGIN server plugin type and the data structures needed to declare the plugin.

Set up the library descriptor for the plugin library file.

This descriptor contains the general plugin descriptor for the server plugin. For a full-text parser plugin, the type must be MYSQL_FTPARSER_PLUGIN. This is the value that identifies the plugin as being legal for use in a WITH PARSER clause when creating a FULLTEXT index. (No other plugin type is legal for this clause.)

For example, the library descriptor for a library that contains a single full-text parser plugin named simple_parser looks like this:

mysql_declare_plugin(ftexample)
{
  MYSQL_FTPARSER_PLUGIN,      /* type                            */
  &simple_parser_descriptor,  /* descriptor                      */
  "simple_parser",            /* name                            */
  "Oracle Corporation",       /* author                          */
  "Simple Full-Text Parser",  /* description                     */
  PLUGIN_LICENSE_GPL,         /* plugin license                  */
  simple_parser_plugin_init,  /* init function (when loaded)     */
  simple_parser_plugin_deinit,/* deinit function (when unloaded) */
  0x0001,                     /* version                         */
  simple_status,              /* status variables                */
  simple_system_variables,    /* system variables                */
  NULL,
  0
}
mysql_declare_plugin_end;

The name member (simple_parser) indicates the name to use for references to the plugin in statements such as INSTALL PLUGIN or UNINSTALL PLUGIN. This is also the name displayed by SHOW PLUGINS or INFORMATION_SCHEMA.PLUGINS.

For more information, see Section 24.2.4.2.1, “Server Plugin Library and Plugin Descriptors”.

Set up the type-specific plugin descriptor.

Each general plugin descriptor in the library descriptor points to a type-specific descriptor. For a full-text parser plugin, the type-specific descriptor is an instance of the st_mysql_ftparser structure in the plugin.h file:

struct st_mysql_ftparser
{
  int interface_version;
  int (*parse)(MYSQL_FTPARSER_PARAM *param);
  int (*init)(MYSQL_FTPARSER_PARAM *param);
  int (*deinit)(MYSQL_FTPARSER_PARAM *param);
};

As shown by the structure definition, the descriptor has an interface version number and contains pointers to three functions.

The interface version number is specified using a symbol, which is in the form: MYSQL_xxx_INTERFACE_VERSION. For full-text parser plugins, the symbol is MYSQL_FTPARSER_INTERFACE_VERSION. In the source code, you will find the actual interface version number for the full-text parser plugin defined in include/mysql/plugin_ftparser.h.

The init and deinit members should point to a function or be set to 0 if the function is not needed. The parse member must point to the function that performs the parsing.

In the simple_parser declaration, that descriptor is indicated by &simple_parser_descriptor. The descriptor specifies the version number for the full-text plugin interface (as given by MYSQL_FTPARSER_INTERFACE_VERSION), and the plugin's parsing, initialization, and deinitialization functions:

static struct st_mysql_ftparser simple_parser_descriptor=
{
  MYSQL_FTPARSER_INTERFACE_VERSION, /* interface version      */
  simple_parser_parse,              /* parsing function       */
  simple_parser_init,               /* parser init function   */
  simple_parser_deinit              /* parser deinit function */
};

A full-text parser plugin is used in two different contexts, indexing and searching. In both contexts, the server calls the initialization and deinitialization functions at the beginning and end of processing each SQL statement that causes the plugin to be invoked. However, during statement processing, the server calls the main parsing function in context-specific fashion:

The plugin declaration in the general plugin descriptor has init and deinit members that point initialization and deinitialization functions, and so does the type-specific plugin descriptor to which it points. However, these pairs of functions have different purposes and are invoked for different reasons:

Each interface function named in the plugin descriptor should return zero for success or nonzero for failure, and each of them receives an argument that points to a MYSQL_FTPARSER_PARAM structure containing the parsing context. The structure has this definition:

typedef struct st_mysql_ftparser_param
{
  int (*mysql_parse)(struct st_mysql_ftparser_param *,
                     char *doc, int doc_len);
  int (*mysql_add_word)(struct st_mysql_ftparser_param *,
                        char *word, int word_len,
                        MYSQL_FTPARSER_BOOLEAN_INFO *boolean_info);
  void *ftparser_state;
  void *mysql_ftparam;
  struct charset_info_st *cs;
  char *doc;
  int length;
  int flags;
  enum enum_ftparser_mode mode;
} MYSQL_FTPARSER_PARAM;

The structure members are used as follows:

If the parser is called in boolean mode, the param->mode value will be MYSQL_FTPARSER_FULL_BOOLEAN_INFO. The MYSQL_FTPARSER_BOOLEAN_INFO structure that the parser uses for passing token information to the server looks like this:

typedef struct st_mysql_ftparser_boolean_info
{
  enum enum_ft_token_type type;
  int yesno;
  int weight_adjust;
  char wasign;
  char trunc;
  /* These are parser state and must be removed. */
  char prev;
  char *quot;
} MYSQL_FTPARSER_BOOLEAN_INFO;

The parser should fill in the structure members as follows:

Plugins should not use the prev and quot members of the MYSQL_FTPARSER_BOOLEAN_INFO structure.

Set up the plugin interface functions.

The general plugin descriptor in the library descriptor names the initialization and deinitialization functions that the server should invoke when it loads and unloads the plugin. For simple_parser, these functions do nothing but return zero to indicate that they succeeded:

static int simple_parser_plugin_init(void *arg __attribute__((unused)))
{
  return(0);
}

static int simple_parser_plugin_deinit(void *arg __attribute__((unused)))
{
  return(0);
}

Because those functions do not actually do anything, you could omit them and specify 0 for each of them in the plugin declaration.

The type-specific plugin descriptor for simple_parser names the initialization, deinitialization, and parsing functions that the server invokes when the plugin is used. For simple_parser, the initialization and deinitialization functions do nothing:

static int simple_parser_init(MYSQL_FTPARSER_PARAM *param
                              __attribute__((unused)))
{
  return(0);
}

static int simple_parser_deinit(MYSQL_FTPARSER_PARAM *param
                                __attribute__((unused)))
{
  return(0);
}

Here too, because those functions do nothing, you could omit them and specify 0 for each of them in the plugin descriptor.

The main parsing function, simple_parser_parse(), acts as a replacement for the built-in full-text parser, so it needs to split text into words and pass each word to the server. The parsing function's first argument is a pointer to a structure that contains the parsing context. This structure has a doc member that points to the text to be parsed, and a length member that indicates how long the text is. The simple parsing done by the plugin considers nonempty runs of whitespace characters to be words, so it identifies words like this:

static int simple_parser_parse(MYSQL_FTPARSER_PARAM *param)
{
  char *end, *start, *docend= param->doc + param->length;

  for (end= start= param->doc;; end++)
  {
    if (end == docend)
    {
      if (end > start)
        add_word(param, start, end - start);
      break;
    }
    else if (isspace(*end))
    {
      if (end > start)
        add_word(param, start, end - start);
      start= end + 1;
    }
  }
  return(0);
}

As the parser finds each word, it invokes a function add_word() to pass the word to the server. add_word() is a helper function only; it is not part of the plugin interface. The parser passes the parsing context pointer to add_word(), as well as a pointer to the word and a length value:

static void add_word(MYSQL_FTPARSER_PARAM *param, char *word, size_t len)
{
  MYSQL_FTPARSER_BOOLEAN_INFO bool_info=
    { FT_TOKEN_WORD, 0, 0, 0, 0, ' ', 0 };

  param->mysql_add_word(param, word, len, &bool_info);
}

For boolean-mode parsing, add_word() fills in the members of the bool_info structure as described earlier in the discussion of the st_mysql_ftparser_boolean_info structure.

Set up the status variables. For the simple_parser plugin, the following status variable array sets up one status variable with a value that is static text, and another with a value that is stored in a long integer variable:

long number_of_calls= 0;

struct st_mysql_show_var simple_status[]=
{
  {"static", (char *)"just a static text", SHOW_CHAR},
  {"called", (char *)&number_of_calls,     SHOW_LONG},
  {0,0,0}
};

When the plugin is installed, the plugin name and the name value are joined with an underscore to form the name displayed by SHOW STATUS. For the array just shown, the resulting status variable names are simple_parser_static and simple_parser_called. This convention means that you can easily display the variables for a plugin using its name:

mysql> SHOW STATUS LIKE 'simple_parser%';
+----------------------+--------------------+
| Variable_name        | Value              |
+----------------------+--------------------+
| simple_parser_static | just a static text |
| simple_parser_called | 0                  |
+----------------------+--------------------+

To compile and install a plugin library file, use the instructions in Section 24.2.4.3, “Compiling and Installing Plugin Libraries”. To make the library file available for use, install it in the plugin directory (the directory named by the plugin_dir system variable). For the simple_parser plugin, it is compiled and installed when you build MySQL from source. It is also included in binary distributions. The build process produces a shared object library with a name of mypluglib.so (the .so suffix might differ depending on your platform).

To use the plugin, register it with the server. For example, to register the plugin at runtime, use this statement (adjust the .so suffix for your platform as necessary):

INSTALL PLUGIN simple_parser SONAME 'mypluglib.so';

For additional information about plugin loading, see Section 5.5.2, “Installing and Uninstalling Plugins”.

To verify plugin installation, examine the INFORMATION_SCHEMA.PLUGINS table or use the SHOW PLUGINS statement. See Section 5.5.3, “Obtaining Server Plugin Information”.

Test the plugin to verify that it works properly.

Create a table that contains a string column and associate the parser plugin with a FULLTEXT index on the column:

mysql> CREATE TABLE t (c VARCHAR(255),
    ->   FULLTEXT (c) WITH PARSER simple_parser
    -> ) ENGINE=MyISAM;
Query OK, 0 rows affected (0.01 sec)

Insert some text into the table and try some searches. These should verify that the parser plugin treats all nonwhitespace characters as word characters:

mysql> INSERT INTO t VALUES
    ->   ('latin1_general_cs is a case-sensitive collation'),
    ->   ('I\'d like a case of oranges'),
    ->   ('this is sensitive information'),
    ->   ('another row'),
    ->   ('yet another row');
Query OK, 5 rows affected (0.02 sec)
Records: 5  Duplicates: 0  Warnings: 0

mysql> SELECT c FROM t;
+-------------------------------------------------+
| c                                               |
+-------------------------------------------------+
| latin1_general_cs is a case-sensitive collation |
| I'd like a case of oranges                      |
| this is sensitive information                   |
| another row                                     |
| yet another row                                 |
+-------------------------------------------------+
5 rows in set (0.00 sec)

mysql> SELECT MATCH(c) AGAINST('case') FROM t;
+--------------------------+
| MATCH(c) AGAINST('case') |
+--------------------------+
|                        0 |
|          1.2968142032623 |
|                        0 |
|                        0 |
|                        0 |
+--------------------------+
5 rows in set (0.00 sec)

mysql> SELECT MATCH(c) AGAINST('sensitive') FROM t;
+-------------------------------+
| MATCH(c) AGAINST('sensitive') |
+-------------------------------+
|                             0 |
|                             0 |
|               1.3253291845322 |
|                             0 |
|                             0 |
+-------------------------------+
5 rows in set (0.01 sec)

mysql> SELECT MATCH(c) AGAINST('case-sensitive') FROM t;
+------------------------------------+
| MATCH(c) AGAINST('case-sensitive') |
+------------------------------------+
|                    1.3109166622162 |
|                                  0 |
|                                  0 |
|                                  0 |
|                                  0 |
+------------------------------------+
5 rows in set (0.01 sec)

mysql> SELECT MATCH(c) AGAINST('I\'d') FROM t;
+--------------------------+
| MATCH(c) AGAINST('I\'d') |
+--------------------------+
|                        0 |
|          1.2968142032623 |
|                        0 |
|                        0 |
|                        0 |
+--------------------------+
5 rows in set (0.01 sec)

Neither “case” nor “insensitive” match “case-insensitive” the way that they would for the built-in parser.