mirror/mariadb
1
0
Fork 0
mirror of https://github.com/MariaDB/server.git synced 2025-01-17 12:32:27 +01:00
Code Issues Projects Releases Packages Wiki Activity Actions
mariadb/storage/connect/ha_connect.cc
Olivier Bertrand 2f48842a0f Commit merged changes.
2013-02-19 12:07:11 +01:00

3879 lines
111 KiB
C++
Raw Blame History

/* Copyright (C) Olivier Bertrand 2004 - 2013
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; version 2 of the License.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */
/**
@file ha_connect.cc
@brief
The ha_connect engine is a stubbed storage engine that enables to create tables
based on external data. Principally they are based on plain files of many
different types, but also on collections of such files, collection of tables,
ODBC tables retrieving data from other DBMS having an ODBC server, and even
virtual tables.
@details
ha_connect will let you create/open/delete tables, the created table can be
done specifying an already existing file, the drop table command will just
suppress the table definition but not the eventual data file.
Indexes are not yet supported but data can be inserted, updated or deleted.
You can enable the CONNECT storage engine in your build by doing the
following during your build process:<br> ./configure
--with-connect-storage-engine (not implemented yet)
You can install the CONNECT handler as all other storage handlers.
Once this is done, MySQL will let you create tables with:<br>
CREATE TABLE <table name> (...) ENGINE=CONNECT;
The example storage engine does not use table locks. It
implements an example "SHARE" that is inserted into a hash by table
name. This is not used yet.
Please read the object definition in ha_connect.h before reading the rest
of this file.
@note
This MariaDB CONNECT handler is currently an adaptation of the XDB handler
that was written for MySQL version 4.1.2-alpha. Its overall design should
be enhanced in the future to meet MariaDB requirements.
@note
It was written also from the Brian's ha_example handler and contains parts
of it that are there but not currently used, such as table variables.
@note
When you create an CONNECT table, the MySQL Server creates a table .frm
(format) file in the database directory, using the table name as the file
name as is customary with MySQL. No other files are created. To get an idea
of what occurs, here is an example select that would do a scan of an entire
table:
@code
ha-connect::open
ha_connect::store_lock
ha_connect::external_lock
ha_connect::info
ha_connect::rnd_init
ha_connect::extra
ENUM HA_EXTRA_CACHE Cache record in HA_rrnd()
ha_connect::rnd_next
ha_connect::rnd_next
ha_connect::rnd_next
ha_connect::rnd_next
ha_connect::rnd_next
ha_connect::rnd_next
ha_connect::rnd_next
ha_connect::rnd_next
ha_connect::rnd_next
ha_connect::extra
ENUM HA_EXTRA_NO_CACHE End caching of records (def)
ha_connect::external_lock
ha_connect::extra
ENUM HA_EXTRA_RESET Reset database to after open
@endcode
Here you see that the connect storage engine has 9 rows called before
rnd_next signals that it has reached the end of its data. Calls to
ha_connect::extra() are hints as to what will be occuring to the request.
Happy use!<br>
-Olivier
*/
#ifdef USE_PRAGMA_IMPLEMENTATION
#pragma implementation // gcc: Class implementation
#endif
#define MYSQL_SERVER 1
#define DONT_DEFINE_VOID
//#include "sql_partition.h"
#include "sql_class.h"
#include "create_options.h"
#include "mysql_com.h"
#include "field.h"
#include "sql_parse.h"
#undef OFFSET
#define NOPARSE
#if defined(UNIX)
#include "osutil.h"
#endif // UNIX
#include "global.h"
#include "plgdbsem.h"
#if defined(ODBC_SUPPORT)
#include "odbccat.h"
#endif // ODBC_SUPPORT
#if defined(MYSQL_SUPPORT)
#include "myconn.h"
#endif // MYSQL_SUPPORT
#include "filamdbf.h"
#include "tabfmt.h"
#include "reldef.h"
#include "tabcol.h"
#include "xindex.h"
#if defined(WIN32)
#include "tabwmi.h"
#endif // WIN32
#include "connect.h"
#include "user_connect.h"
#include "ha_connect.h"
#include "mycat.h"
#include "myutil.h"
#define PLGXINI "plgcnx.ini" /* Configuration settings file */
#define my_strupr(p) my_caseup_str(default_charset_info, (p));
#define my_strlwr(p) my_casedn_str(default_charset_info, (p));
#define my_stricmp(a,b) my_strcasecmp(default_charset_info, (a), (b))
#ifdef LIBXML2_SUPPORT
void XmlInitParserLib(void);
void XmlCleanupParserLib(void);
#endif // LIBXML2_SUPPORT
/***********************************************************************/
/* DB static variables. */
/***********************************************************************/
extern "C" char plgxini[];
extern "C" char plgini[];
extern "C" char nmfile[];
extern "C" char pdebug[];
extern "C" {
char version[]= "Version 1.01.0001 February 08, 2013";
#if defined(XMSG)
char msglang[]; // Default message language
#endif
int trace= 0; // The general trace value
} // extern "C"
/****************************************************************************/
/* Initialize the ha_connect static members. */
/****************************************************************************/
#define CONNECT_INI "connect.ini"
char connectini[_MAX_PATH]= CONNECT_INI;
int xtrace= 0;
ulong ha_connect::num= 0;
//int DTVAL::Shift= 0;
static handler *connect_create_handler(handlerton *hton,
TABLE_SHARE *table,
MEM_ROOT *mem_root);
handlerton *connect_hton;
/* Variables for connect share methods */
/*
Hash used to track the number of open tables; variable for connect share
methods
*/
static HASH connect_open_tables;
/* The mutex used to init the hash; variable for example share methods */
mysql_mutex_t connect_mutex;
/**
structure for CREATE TABLE options (table options)
These can be specified in the CREATE TABLE:
CREATE TABLE ( ... ) {...here...}
*/
struct ha_table_option_struct {
const char *type;
const char *filename;
const char *optname;
const char *tabname;
const char *tablist;
const char *dbname;
const char *separator;
//const char *connect;
const char *qchar;
const char *module;
const char *subtype;
const char *catfunc;
const char *oplist;
const char *data_charset;
int lrecl;
int elements;
//int estimate;
int multiple;
int header;
int quoted;
int ending;
int compressed;
bool mapped;
bool huge;
bool split;
bool readonly;
};
#if defined(MARIADB)
ha_create_table_option connect_table_option_list[]=
{
// These option are for stand alone Connect tables
HA_TOPTION_STRING("TABLE_TYPE", type),
HA_TOPTION_STRING("FILE_NAME", filename),
HA_TOPTION_STRING("XFILE_NAME", optname),
//HA_TOPTION_STRING("CONNECT_STRING", connect),
HA_TOPTION_STRING("TABNAME", tabname),
HA_TOPTION_STRING("TABLE_LIST", tablist),
HA_TOPTION_STRING("DB_NAME", dbname),
HA_TOPTION_STRING("SEP_CHAR", separator),
HA_TOPTION_STRING("QCHAR", qchar),
HA_TOPTION_STRING("MODULE", module),
HA_TOPTION_STRING("SUBTYPE", subtype),
HA_TOPTION_STRING("CATFUNC", catfunc),
HA_TOPTION_STRING("OPTION_LIST", oplist),
HA_TOPTION_STRING("DATA_CHARSET", data_charset),
HA_TOPTION_NUMBER("LRECL", lrecl, 0, 0, INT_MAX32, 1),
HA_TOPTION_NUMBER("BLOCK_SIZE", elements, 0, 0, INT_MAX32, 1),
//HA_TOPTION_NUMBER("ESTIMATE", estimate, 0, 0, INT_MAX32, 1),
HA_TOPTION_NUMBER("MULTIPLE", multiple, 0, 0, 2, 1),
HA_TOPTION_NUMBER("HEADER", header, 0, 0, 3, 1),
HA_TOPTION_NUMBER("QUOTED", quoted, -1, 0, 3, 1),
HA_TOPTION_NUMBER("ENDING", ending, -1, 0, INT_MAX32, 1),
HA_TOPTION_NUMBER("COMPRESS", compressed, 0, 0, 2, 1),
//HA_TOPTION_BOOL("COMPRESS", compressed, 0),
HA_TOPTION_BOOL("MAPPED", mapped, 0),
HA_TOPTION_BOOL("HUGE", huge, 0),
HA_TOPTION_BOOL("SPLIT", split, 0),
HA_TOPTION_BOOL("READONLY", readonly, 0),
HA_TOPTION_END
};
#endif // MARIADB
/**
structure for CREATE TABLE options (field options)
These can be specified in the CREATE TABLE per field:
CREATE TABLE ( field ... {...here...}, ... )
*/
struct ha_field_option_struct
{
int offset;
int freq; // Not used by this version
int opt; // Not used by this version
int buflen;
const char *dateformat;
const char *fieldformat;
char *special;
};
#if defined(MARIADB)
ha_create_table_option connect_field_option_list[]=
{
HA_FOPTION_NUMBER("FLAG", offset, -1, 0, INT_MAX32, 1),
HA_FOPTION_NUMBER("FREQUENCY", freq, 0, 0, INT_MAX32, 1), // not used
HA_FOPTION_NUMBER("OPT_VALUE", opt, 0, 0, 2, 1), // used for indexing
HA_FOPTION_NUMBER("BUF_LENGTH", buflen, 0, 0, INT_MAX32, 1),
HA_FOPTION_STRING("DATE_FORMAT", dateformat),
HA_FOPTION_STRING("FIELD_FORMAT", fieldformat),
HA_FOPTION_STRING("SPECIAL", special),
HA_FOPTION_END
};
#endif // MARIADB
/***********************************************************************/
/* Push G->Message as a MySQL warning. */
/***********************************************************************/
bool PushWarning(PGLOBAL g, PTDBASE tdbp)
{
PHC phc;
THD *thd;
MYCAT *cat= (MYCAT*)tdbp->GetDef()->GetCat();
if (!cat || !(phc= cat->GetHandler()) || !phc->GetTable() ||
!(thd= (phc->GetTable())->in_use))
return true;
push_warning(thd, MYSQL_ERROR::WARN_LEVEL_WARN, 0, g->Message);
return false;
} // end of PushWarning
/**
@brief
Function we use in the creation of our hash to get key.
*/
static uchar* connect_get_key(CONNECT_SHARE *share, size_t *length,
my_bool not_used __attribute__((unused)))
{
*length=share->table_name_length;
return (uchar*) share->table_name;
}
#ifdef HAVE_PSI_INTERFACE
static PSI_mutex_key ex_key_mutex_connect, ex_key_mutex_CONNECT_SHARE_mutex;
static PSI_mutex_info all_connect_mutexes[]=
{
{ &ex_key_mutex_connect, "connect", PSI_FLAG_GLOBAL},
{ &ex_key_mutex_CONNECT_SHARE_mutex, "CONNECT_SHARE::mutex", 0}
};
static void init_connect_psi_keys()
{
const char* category= "connect";
int count;
if (PSI_server == NULL)
return;
count= array_elements(all_connect_mutexes);
PSI_server->register_mutex(category, all_connect_mutexes, count);
}
#endif
/**
@brief
Plugin initialization
*/
static int connect_init_func(void *p)
{
DBUG_ENTER("connect_init_func");
char dir[_MAX_PATH - sizeof(CONNECT_INI) - 1];
#ifdef LIBXML2_SUPPORT
XmlInitParserLib();
#endif // LIBXML2_SUPPORT
/* Build connect.ini file name */
my_getwd(dir, sizeof(dir) - 1, MYF(0));
snprintf(connectini, sizeof(connectini), "%s%s", dir, CONNECT_INI);
sql_print_information("CONNECT: %s=%s", CONNECT_INI, connectini);
if ((xtrace= GetPrivateProfileInt("CONNECT", "Trace", 0, connectini)))
{
sql_print_information("CONNECT: xtrace=%d", xtrace);
sql_print_information("CONNECT: plgini=%s", plgini);
sql_print_information("CONNECT: plgxini=%s", plgxini);
sql_print_information("CONNECT: nmfile=%s", nmfile);
sql_print_information("CONNECT: pdebug=%s", pdebug);
sql_print_information("CONNECT: version=%s", version);
trace= xtrace;
} // endif xtrace
#ifdef HAVE_PSI_INTERFACE
init_connect_psi_keys();
#endif
connect_hton= (handlerton *)p;
mysql_mutex_init(ex_key_mutex_connect, &connect_mutex, MY_MUTEX_INIT_FAST);
//VOID(mysql_mutex_init(&connect_mutex, MY_MUTEX_INIT_FAST));
(void) my_hash_init(&connect_open_tables, system_charset_info, 32, 0, 0,
(my_hash_get_key) connect_get_key, 0, 0);
//connect_hton->name= "CONNECT";
connect_hton->state= SHOW_OPTION_YES;
//connect_hton->comment= "CONNECT handler";
connect_hton->create= connect_create_handler;
connect_hton->flags= HTON_TEMPORARY_NOT_SUPPORTED | HTON_NO_PARTITION;
#if defined(MARIADB)
connect_hton->db_type= DB_TYPE_AUTOASSIGN;
connect_hton->table_options= connect_table_option_list;
connect_hton->field_options= connect_field_option_list;
#else // !MARIADB
//connect_hton->system_database= connect_system_database;
//connect_hton->is_supported_system_table= connect_is_supported_system_table;
#endif // !MARIADB
if (xtrace)
sql_print_information("connect_init: hton=%p", p);
DTVAL::SetTimeShift(); // Initialize time zone shift once for all
DBUG_RETURN(0);
}
/**
@brief
Plugin clean up
*/
static int connect_done_func(void *p)
{
int error= 0;
PCONNECT pc, pn;
DBUG_ENTER("connect_done_func");
#ifdef LIBXML2_SUPPORT
XmlCleanupParserLib();
#endif // LIBXML2_SUPPORT
if (connect_open_tables.records)
error= 1;
for (pc= user_connect::to_users; pc; pc= pn) {
if (pc->g)
PlugCleanup(pc->g, true);
pn= pc->next;
delete pc;
} // endfor pc
my_hash_free(&connect_open_tables);
mysql_mutex_destroy(&connect_mutex);
DBUG_RETURN(error);
}
/**
@brief
Example of simple lock controls. The "share" it creates is a
structure we will pass to each example handler. Do you have to have
one of these? Well, you have pieces that are used for locking, and
they are needed to function.
*/
static CONNECT_SHARE *get_share(const char *table_name, TABLE *table)
{
CONNECT_SHARE *share;
uint length;
char *tmp_name;
mysql_mutex_lock(&connect_mutex);
length=(uint) strlen(table_name);
if (!(share=(CONNECT_SHARE*)my_hash_search(&connect_open_tables,
(uchar*) table_name, length))) {
if (!(share=(CONNECT_SHARE *)my_multi_malloc(MYF(MY_WME | MY_ZEROFILL),
&share, sizeof(*share), &tmp_name, length+1, NullS))) {
mysql_mutex_unlock(&connect_mutex);
return NULL;
} // endif share
share->use_count=0;
share->table_name_length=length;
share->table_name=tmp_name;
strmov(share->table_name, table_name);
if (my_hash_insert(&connect_open_tables, (uchar*) share))
goto error;
thr_lock_init(&share->lock);
mysql_mutex_init(ex_key_mutex_CONNECT_SHARE_mutex,
&share->mutex, MY_MUTEX_INIT_FAST);
} // endif share
share->use_count++;
mysql_mutex_unlock(&connect_mutex);
return share;
error:
mysql_mutex_destroy(&share->mutex);
my_free(share);
return NULL;
}
/**
@brief
Free lock controls. We call this whenever we close a table. If the table had
the last reference to the share, then we free memory associated with it.
*/
static int free_share(CONNECT_SHARE *share)
{
mysql_mutex_lock(&connect_mutex);
if (!--share->use_count) {
my_hash_delete(&connect_open_tables, (uchar*) share);
thr_lock_delete(&share->lock);
mysql_mutex_destroy(&share->mutex);
#if !defined(MARIADB)
my_free(share->table_options);
my_free(share->field_options);
#endif // !MARIADB
my_free(share);
} // endif share
mysql_mutex_unlock(&connect_mutex);
return 0;
}
static handler* connect_create_handler(handlerton *hton,
TABLE_SHARE *table,
MEM_ROOT *mem_root)
{
handler *h= new (mem_root) ha_connect(hton, table);
if (xtrace)
printf("New CONNECT %p, table: %s\n",
h, table ? table->table_name.str : "<null>");
return h;
} // end of connect_create_handler
/****************************************************************************/
/* ha_connect constructor. */
/****************************************************************************/
ha_connect::ha_connect(handlerton *hton, TABLE_SHARE *table_arg)
:handler(hton, table_arg)
{
hnum= ++num;
xp= NULL; // Tested in next call
xp= (table) ? GetUser(table->in_use) : NULL;
tdbp= NULL;
sdval= NULL;
xmod= MODE_ANY;
istable= false;
//*tname= '\0';
bzero((char*) &xinfo, sizeof(XINFO));
valid_info= false;
valid_query_id= 0;
creat_query_id= (table && table->in_use) ? table->in_use->query_id : 0;
stop= false;
//hascond= false;
indexing= -1;
data_file_name= NULL;
index_file_name= NULL;
enable_activate_all_index= 0;
int_table_flags= (HA_NO_TRANSACTIONS | HA_NO_PREFIX_CHAR_KEYS);
ref_length= sizeof(int);
#if !defined(MARIADB)
share= NULL;
table_options= NULL;
field_options= NULL;
#endif // !MARIADB
tshp= NULL;
} // end of ha_connect constructor
/****************************************************************************/
/* ha_connect destructor. */
/****************************************************************************/
ha_connect::~ha_connect(void)
{
if (xp) {
PCONNECT p;
xp->count--;
for (p= user_connect::to_users; p; p= p->next)
if (p == xp)
break;
if (p && !p->count) {
if (p->next)
p->next->previous= p->previous;
if (p->previous)
p->previous->next= p->next;
else
user_connect::to_users= p->next;
} // endif p
if (!xp->count) {
PlugCleanup(xp->g, true);
delete xp;
} // endif count
} // endif xp
#if !defined(MARIADB)
my_free(table_options);
my_free(field_options);
#endif // !MARIADB
} // end of ha_connect destructor
/****************************************************************************/
/* Get a pointer to the user of this handler. */
/****************************************************************************/
PCONNECT ha_connect::GetUser(THD *thd)
{
const char *dbn= NULL;
if (!thd)
return NULL;
if (xp && thd == xp->thdp)
return xp;
for (xp= user_connect::to_users; xp; xp= xp->next)
if (thd == xp->thdp)
break;
if (!xp) {
xp= new user_connect(thd, dbn);
if (xp->user_init(this)) {
delete xp;
xp= NULL;
} // endif user_init
} else
xp->count++;
return xp;
} // end of GetUser
/****************************************************************************/
/* Get the global pointer of the user of this handler. */
/****************************************************************************/
PGLOBAL ha_connect::GetPlug(THD *thd)
{
PCONNECT lxp= GetUser(thd);
return (lxp) ? lxp->g : NULL;
} // end of GetPlug
/****************************************************************************/
/* Return the value of an option specified in the option list. */
/****************************************************************************/
char *ha_connect::GetListOption(const char *opname,
const char *oplist,
const char *def)
{
char key[16], val[256];
char *pk, *pv, *pn;
char *opval= (char *) def;
int n;
for (pk= (char*)oplist; pk; pk= ++pn) {
pn= strchr(pk, ',');
pv= strchr(pk, '=');
if (pv && (!pn || pv < pn)) {
n= pv - pk;
memcpy(key, pk, n);
key[n]= 0;
pv++;
if (pn) {
n= pn - pv;
memcpy(val, pv, n);
val[n]= 0;
} else
strcpy(val, pv);
} else {
if (pn) {
n= pn - pk;
memcpy(key, pk, n);
key[n]= 0;
} else
strcpy(key, pk);
val[0]= 0;
} // endif pv
if (!stricmp(opname, key)) {
opval= (char*)PlugSubAlloc(xp->g, NULL, strlen(val) + 1);
strcpy(opval, val);
break;
} else if (!pn)
break;
} // endfor pk
return opval;
} // end of GetListOption
/****************************************************************************/
/* Return the table option structure. */
/****************************************************************************/
PTOS ha_connect::GetTableOptionStruct(TABLE *tab)
{
#if defined(MARIADB)
return (tshp) ? tshp->option_struct : tab->s->option_struct;
#else // !MARIADB
if (share && share->table_options)
return share->table_options;
else if (table_options)
return table_options;
char *pk, *pv, *pn, *val;
size_t len= sizeof(ha_table_option_struct) + tab->s->comment.length + 1;
PTOS top= (PTOS)my_malloc(len, MYF(MY_FAE | MY_ZEROFILL));
top->quoted= -1; // Default value
top->ending= -1; // Default value
pk= (char *)top + sizeof(ha_table_option_struct);
memcpy(pk, tab->s->comment.str, tab->s->comment.length);
for (; pk; pk= ++pn) {
pn= strchr(pk, ',');
pv= strchr(pk, '=');
if (pn) *pn= 0;
if (pv) *pv= 0;
val= (pv && (!pn || pv < pn)) ? pv + 1 : "";
if (!stricmp(pk, "type") || !stricmp(pk, "Table_Type")) {
top->type= val;
} else if (!stricmp(pk, "fn") || !stricmp(pk, "filename")
|| !stricmp(pk, "File_Name")) {
top->filename= val;
} else if (!stricmp(pk, "optfn") || !stricmp(pk, "optname")
|| !stricmp(pk, "Xfile_Name")) {
top->optname= val;
} else if (!stricmp(pk, "name") || !stricmp(pk, "tabname")) {
top->tabname= val;
} else if (!stricmp(pk, "tablist") || !stricmp(pk, "tablelist")
|| !stricmp(pk, "Table_list")) {
top->tablist= val;
} else if (!stricmp(pk, "sep") || !stricmp(pk, "separator")
|| !stricmp(pk, "Sep_Char")) {
top->separator= val;
} else if (!stricmp(pk, "db") || !stricmp(pk, "database")
|| !stricmp(pk, "DB_Name")) {
top->dbname= val;
} else if (!stricmp(pk, "qchar")) {
top->qchar= val;
} else if (!stricmp(pk, "module")) {
top->module= val;
} else if (!stricmp(pk, "subtype")) {
top->subtype= val;
} else if (!stricmp(pk, "lrecl")) {
top->lrecl= atoi(val);
} else if (!stricmp(pk, "elements")) {
top->elements= atoi(val);
} else if (!stricmp(pk, "multiple")) {
top->multiple= atoi(val);
} else if (!stricmp(pk, "header")) {
top->header= atoi(val);
} else if (!stricmp(pk, "quoted")) {
top->quoted= atoi(val);
} else if (!stricmp(pk, "ending")) {
top->ending= atoi(val);
} else if (!stricmp(pk, "compressed")) {
top->compressed= atoi(val);
} else if (!stricmp(pk, "mapped")) {
top->mapped= (!*val || *val == 'y' || *val == 'Y' || atoi(val) != 0);
} else if (!stricmp(pk, "huge")) {
top->huge= (!*val || *val == 'y' || *val == 'Y' || atoi(val) != 0);
} else if (!stricmp(pk, "split")) {
top->split= (!*val || *val == 'y' || *val == 'Y' || atoi(val) != 0);
} else if (!stricmp(pk, "readonly") || !stricmp(pk, "protected")) {
top->readonly= (!*val || *val == 'y' || *val == 'Y' || atoi(val) != 0);
} // endif's
if (!pn)
break;
} // endfor pk
// This to get all other options
top->oplist= tab->s->comment.str;
if (share)
share->table_options= top;
else
table_options= top;
return top;
#endif // !MARIADB
} // end of GetTableOptionStruct
/****************************************************************************/
/* Return the value of a string option or NULL if not specified. */
/****************************************************************************/
char *ha_connect::GetStringOption(char *opname, char *sdef)
{
char *opval= NULL;
PTOS options= GetTableOptionStruct(table);
if (!options)
;
else if (!stricmp(opname, "Type"))
opval= (char*)options->type;
else if (!stricmp(opname, "Filename"))
opval= (char*)options->filename;
else if (!stricmp(opname, "Optname"))
opval= (char*)options->optname;
else if (!stricmp(opname, "Tabname"))
opval= (char*)options->tabname;
else if (!stricmp(opname, "Tablist"))
opval= (char*)options->tablist;
else if (!stricmp(opname, "Database"))
opval= (char*)options->dbname;
else if (!stricmp(opname, "Separator"))
opval= (char*)options->separator;
else if (!stricmp(opname, "Connect"))
// opval= (char*)options->connect;
opval= table->s->connect_string.str;
else if (!stricmp(opname, "Qchar"))
opval= (char*)options->qchar;
else if (!stricmp(opname, "Module"))
opval= (char*)options->module;
else if (!stricmp(opname, "Subtype"))
opval= (char*)options->subtype;
else if (!stricmp(opname, "Catfunc"))
opval= (char*)options->catfunc;
else if (!stricmp(opname, "Data_charset"))
opval= (char*)options->data_charset;
if (!opval && options->oplist)
opval= GetListOption(opname, options->oplist);
if (!opval) {
if (sdef && !strcmp(sdef, "*")) {
// Return the handler default value
if (!stricmp(opname, "Database"))
opval= (char*)GetDBName(NULL); // Current database
} else
opval= sdef; // Caller default
} // endif !opval
return opval;
} // end of GetStringOption
/****************************************************************************/
/* Return the value of a Boolean option or bdef if not specified. */
/****************************************************************************/
bool ha_connect::GetBooleanOption(char *opname, bool bdef)
{
bool opval= bdef;
char *pv;
PTOS options= GetTableOptionStruct(table);
if (!options)
;
else if (!stricmp(opname, "Mapped"))
opval= options->mapped;
else if (!stricmp(opname, "Huge"))
opval= options->huge;
//else if (!stricmp(opname, "Compressed"))
// opval= options->compressed;
else if (!stricmp(opname, "Split"))
opval= options->split;
else if (!stricmp(opname, "Readonly"))
opval= options->readonly;
else if (options->oplist)
if ((pv= GetListOption(opname, options->oplist)))
opval= (!*pv || *pv == 'y' || *pv == 'Y' || atoi(pv) != 0);
return opval;
} // end of GetBooleanOption
/****************************************************************************/
/* Return the value of an integer option or NO_IVAL if not specified. */
/****************************************************************************/
int ha_connect::GetIntegerOption(char *opname)
{
int opval= NO_IVAL;
char *pv;
PTOS options= GetTableOptionStruct(table);
if (!options)
;
else if (!stricmp(opname, "Lrecl"))
opval= options->lrecl;
else if (!stricmp(opname, "Elements"))
opval= options->elements;
else if (!stricmp(opname, "Estimate"))
// opval= options->estimate;
opval= (int)table->s->max_rows;
else if (!stricmp(opname, "Avglen"))
opval= (int)table->s->avg_row_length;
else if (!stricmp(opname, "Multiple"))
opval= options->multiple;
else if (!stricmp(opname, "Header"))
opval= options->header;
else if (!stricmp(opname, "Quoted"))
opval= options->quoted;
else if (!stricmp(opname, "Ending"))
opval= options->ending;
else if (!stricmp(opname, "Compressed"))
opval= (options->compressed);
if (opval == NO_IVAL && options->oplist)
if ((pv= GetListOption(opname, options->oplist)))
opval= atoi(pv);
return opval;
} // end of GetIntegerOption
/****************************************************************************/
/* Set the value of the opname option (does not work for oplist options) */
/* Currently used only to set the Lrecl value. */
/****************************************************************************/
bool ha_connect::SetIntegerOption(char *opname, int n)
{
PTOS options= GetTableOptionStruct(table);
if (!options)
return true;
if (!stricmp(opname, "Lrecl"))
options->lrecl= n;
else if (!stricmp(opname, "Elements"))
options->elements= n;
//else if (!stricmp(opname, "Estimate"))
// options->estimate= n;
else if (!stricmp(opname, "Multiple"))
options->multiple= n;
else if (!stricmp(opname, "Header"))
options->header= n;
else if (!stricmp(opname, "Quoted"))
options->quoted= n;
else if (!stricmp(opname, "Ending"))
options->ending= n;
else if (!stricmp(opname, "Compressed"))
options->compressed= n;
else
return true;
//else if (options->oplist)
// SetListOption(opname, options->oplist, n);
return false;
} // end of SetIntegerOption
/****************************************************************************/
/* Return a field option structure. */
/****************************************************************************/
PFOS ha_connect::GetFieldOptionStruct(Field *fdp)
{
#if defined(MARIADB)
return fdp->option_struct;
#else // !MARIADB
if (share && share->field_options)
return &share->field_options[fdp->field_index];
else if (field_options)
return &field_options[fdp->field_index];
char *pc, *pk, *pv, *pn, *val;
int i, k, n= table->s->fields;
size_t len= n + n * sizeof(ha_field_option_struct);
PFOS fp, fop;
for (i= 0; i < n; i++)
len+= table->s->field[i]->comment.length;
fop= (PFOS)my_malloc(len, MYF(MY_FAE | MY_ZEROFILL));
pc= (char*)fop + n * sizeof(ha_field_option_struct);
for (i= k= 0; i < n; i++) {
fp= &fop[i];
fp->offset= -1; // Default value
if (!table->s->field[i]->comment.length)
continue;
memcpy(pc, table->s->field[i]->comment.str,
table->s->field[i]->comment.length);
for (pk= pc; pk; pk= ++pn) {
if ((pn= strchr(pk, ','))) *pn= 0;
if ((pv= strchr(pk, '='))) *pv= 0;
val= (pv && (!pn || pv < pn)) ? pv + 1 : "";
if (!stricmp(pk, "datefmt") || !stricmp(pk, "date_format")) {
fp->dateformat= val;
} else if (!stricmp(pk, "fieldfmt") || !stricmp(pk, "field_format")) {
fp->fieldformat= val;
} else if (!stricmp(pk, "special")) {
fp->special= val;
} else if (!stricmp(pk, "offset") || !stricmp(pk, "flag")) {
fp->offset= atoi(val);
} else if (!stricmp(pk, "freq")) {
fp->freq= atoi(val);
} else if (!stricmp(pk, "opt")) {
fp->opt= atoi(val);
} else if (!stricmp(pk, "buflen")) {
fp->buflen= atoi(val);
} // endif's
if (!pn)
break;
} // endfor pk
pc+= table->s->field[i]->comment.length + 1;
} // endfor i
if (share)
share->field_options= fop;
else
field_options= fop;
return &fop[fdp->field_index];
#endif // !MARIADB
} // end of GetFildOptionStruct
/****************************************************************************/
/* Returns the column description structure used to make the column. */
/****************************************************************************/
void *ha_connect::GetColumnOption(void *field, PCOLINFO pcf)
{
const char *cp;
int len;
ha_field_option_struct *fop;
Field* fp;
Field* *fldp;
// Double test to be on the safe side
if (!table)
return NULL;
// Find the column to describe
if (field) {
fldp= (Field**)field;
fldp++;
} else
fldp= (tshp) ? tshp->field : table->field;
if (!(fp= *fldp))
return NULL;
// Get the CONNECT field options structure
fop= GetFieldOptionStruct(fp);
pcf->Flags= 0;
// Now get column information
if (fop && fop->special) {
pcf->Name= "*";
return fldp;
} else
pcf->Name= (char*)fp->field_name;
pcf->Prec= 0;
pcf->Opt= (fop) ? fop->opt : 0;
if ((pcf->Length= fp->field_length) < 0)
pcf->Length= 256; // BLOB?
switch (fp->type()) {
case MYSQL_TYPE_BLOB:
case MYSQL_TYPE_VARCHAR:
pcf->Flags |= U_VAR;
case MYSQL_TYPE_STRING:
pcf->Type= TYPE_STRING;
// Do something for case
cp= fp->charset()->name;
// Find if collation name ends by _ci
if (!strcmp(cp + strlen(cp) - 3, "_ci")) {
pcf->Prec= 1; // Case insensitive
pcf->Opt= 0; // Prevent index opt until it is safe
} // endif ci
break;
case MYSQL_TYPE_LONG:
pcf->Type= TYPE_INT;
break;
case MYSQL_TYPE_SHORT:
pcf->Type= TYPE_SHORT;
break;
case MYSQL_TYPE_DOUBLE:
case MYSQL_TYPE_FLOAT:
pcf->Type= TYPE_FLOAT;
pcf->Prec= max(min(fp->decimals(), ((unsigned)pcf->Length - 2)), 0);
break;
case MYSQL_TYPE_DATE:
case MYSQL_TYPE_TIME:
case MYSQL_TYPE_DATETIME:
case MYSQL_TYPE_TIMESTAMP:
pcf->Type= TYPE_DATE;
break;
case MYSQL_TYPE_LONGLONG:
pcf->Type= TYPE_BIGINT;
break;
default:
pcf->Type=TYPE_ERROR;
} // endswitch type
// This is used to skip null bit
if (fp->real_maybe_null())
pcf->Flags |= U_NULLS;
#if defined(MARIADB)
// Mark virtual columns as such
if (fp->vcol_info && !fp->stored_in_db)
pcf->Flags |= U_VIRTUAL;
#endif // MARIADB
pcf->Key= 0; // Not used when called from MySQL
pcf->Remark= fp->comment.str;
if (fop) {
pcf->Offset= fop->offset;
// pcf->Freq= fop->freq;
pcf->Datefmt= (char*)fop->dateformat;
pcf->Fieldfmt= (char*)fop->fieldformat;
// This is useful in particular for date columns
if ((len= fop->buflen) > pcf->Length)
pcf->Length= len;
} else {
pcf->Offset= -1;
// pcf->Freq= 0;
pcf->Datefmt= NULL;
pcf->Fieldfmt= NULL;
} // endif fop
return fldp;
} // end of GetColumnOption
/****************************************************************************/
/* Returns the index description structure used to make the index. */
/****************************************************************************/
PIXDEF ha_connect::GetIndexInfo(int n)
{
char *name, *pn;
bool unique;
PIXDEF xdp= NULL;
PKPDEF kpp, pkp= NULL;
PGLOBAL& g= xp->g;
KEY kp;
// Find the index to describe
if ((unsigned)n < table->s->keynames.count)
// kp= table->key_info[n]; which one ???
kp= table->s->key_info[n];
else
return NULL;
// Now get index information
pn= (char*)table->s->keynames.type_names[n];
name= (char*)PlugSubAlloc(g, NULL, strlen(pn) + 1);
strcpy(name, pn); // This is probably unuseful
unique= (kp.flags & 1) != 0;
// Allocate the index description block
xdp= new(g) INDEXDEF(name, unique, n);
// Get the the key parts info
for (int k= 0; (unsigned)k < kp.key_parts; k++) {
pn= (char*)kp.key_part[k].field->field_name;
name= (char*)PlugSubAlloc(g, NULL, strlen(pn) + 1);
strcpy(name, pn); // This is probably unuseful
// Allocate the key part description block
kpp= new(g) KPARTDEF(name, k + 1);
kpp->SetKlen(kp.key_part[k].length);
// Index on auto increment column is an XXROW index
if (kp.key_part[k].field->flags & AUTO_INCREMENT_FLAG && kp.key_parts == 1)
xdp->SetAuto(true);
if (pkp)
pkp->SetNext(kpp);
else
xdp->SetToKeyParts(kpp);
pkp= kpp;
} // endfor k
xdp->SetNParts(kp.key_parts);
return xdp;
} // end of GetIndexInfo
const char *ha_connect::GetDBName(const char* name)
{
return (name) ? name : table->s->db.str;
} // end of GetDBName
const char *ha_connect::GetTableName(void)
{
return table->s->table_name.str;
} // end of GetTableName
/****************************************************************************/
/* Returns the column real or special name length of a field. */
/****************************************************************************/
int ha_connect::GetColNameLen(Field *fp)
{
int n;
PFOS fop= GetFieldOptionStruct(fp);
// Now get the column name length
if (fop && fop->special)
n= strlen(fop->special) + 1;
else
n= strlen(fp->field_name);
return n;
} // end of GetColNameLen
/****************************************************************************/
/* Returns the column real or special name of a field. */
/****************************************************************************/
char *ha_connect::GetColName(Field *fp)
{
PFOS fop= GetFieldOptionStruct(fp);
return (fop && fop->special) ? fop->special : (char*)fp->field_name;
} // end of GetColName
/****************************************************************************/
/* Adds the column real or special name of a field to a string. */
/****************************************************************************/
void ha_connect::AddColName(char *cp, Field *fp)
{
PFOS fop= GetFieldOptionStruct(fp);
// Now add the column name
if (fop && fop->special)
// The prefix * mark the column as "special"
strcat(strcpy(cp, "*"), strupr(fop->special));
else
strcpy(cp, (char*)fp->field_name);
} // end of AddColName
/****************************************************************************/
/* Get the table description block of a CONNECT table. */
/****************************************************************************/
PTDB ha_connect::GetTDB(PGLOBAL g)
{
const char *table_name;
PTDB tp;
// Double test to be on the safe side
if (!g || !table)
return NULL;
table_name= GetTableName();
if (tdbp && !stricmp(tdbp->GetName(), table_name)
&& tdbp->GetMode() == xmod && !xp->CheckQuery(valid_query_id)) {
tp= tdbp;
tp->SetMode(xmod);
} else if ((tp= CntGetTDB(g, table_name, xmod, this)))
valid_query_id= xp->last_query_id;
else
printf("GetTDB: %s\n", g->Message);
return tp;
} // end of GetTDB
/****************************************************************************/
/* Open a CONNECT table, restricting column list if cols is true. */
/****************************************************************************/
bool ha_connect::OpenTable(PGLOBAL g, bool del)
{
bool rc= false;
char *c1= NULL, *c2=NULL;
// Double test to be on the safe side
if (!g || !table) {
printf("OpenTable logical error; g=%p table=%p\n", g, table);
return true;
} // endif g
if (!(tdbp= GetTDB(g)))
return true;
else if (tdbp->IsReadOnly())
switch (xmod) {
case MODE_WRITE:
case MODE_INSERT:
case MODE_UPDATE:
case MODE_DELETE:
strcpy(g->Message, MSG(READ_ONLY));
return true;
default:
break;
} // endswitch xmode
// Get the list of used fields (columns)
char *p;
unsigned int k1, k2, n1, n2;
Field* *field;
MY_BITMAP *map= (xmod != MODE_INSERT) ? table->read_set : table->write_set;
MY_BITMAP *ump= (xmod == MODE_UPDATE) ? table->write_set : NULL;
k1= k2= 0;
n1= n2= 1; // 1 is space for final null character
for (field= table->field; *field; field++) {
if (bitmap_is_set(map, (*field)->field_index)) {
n1+= (GetColNameLen(*field) + 1);
k1++;
} // endif
if (ump && bitmap_is_set(ump, (*field)->field_index)) {
n2+= (GetColNameLen(*field) + 1);
k2++;
} // endif
} // endfor field
if (k1) {
p= c1= (char*)PlugSubAlloc(g, NULL, n1);
for (field= table->field; *field; field++)
if (bitmap_is_set(map, (*field)->field_index)) {
AddColName(p, *field);
p+= (strlen(p) + 1);
} // endif used field
*p= '\0'; // mark end of list
} // endif k1
if (k2) {
p= c2= (char*)PlugSubAlloc(g, NULL, n2);
for (field= table->field; *field; field++)
if (bitmap_is_set(ump, (*field)->field_index)) {
AddColName(p, *field);
p+= (strlen(p) + 1);
} // endif used field
*p= '\0'; // mark end of list
} // endif k2
// Open the table
if (!(rc= CntOpenTable(g, tdbp, xmod, c1, c2, del, this))) {
istable= true;
// strmake(tname, table_name, sizeof(tname)-1);
if (xmod == MODE_ANY && stop && *tdbp->GetName() != '#') {
// We are in a create index query
if (!((PTDBASE)tdbp)->GetDef()->Indexable()) {
sprintf(g->Message, "Table %s cannot be indexed", tdbp->GetName());
rc= true;
} else if (xp) // xp can be null when called from create
xp->tabp= (PTDBDOS)tdbp; // The table on which the index is created
} // endif xmod
// tdbp->SetOrig((PTBX)table); // used by CheckCond
} else
printf("OpenTable: %s\n", g->Message);
if (rc) {
tdbp= NULL;
valid_info= false;
} // endif rc
return rc;
} // end of OpenTable
/****************************************************************************/
/* IsOpened: returns true if the table is already opened. */
/****************************************************************************/
bool ha_connect::IsOpened(void)
{
return (!xp->CheckQuery(valid_query_id) && tdbp
&& tdbp->GetUse() == USE_OPEN);
} // end of IsOpened
/****************************************************************************/
/* Close a CONNECT table. */
/****************************************************************************/
int ha_connect::CloseTable(PGLOBAL g)
{
int rc= CntCloseTable(g, tdbp);
tdbp= NULL;
sdval=NULL;
valid_info= false;
indexing= -1;
return rc;
} // end of CloseTable
/***********************************************************************/
/* Make a pseudo record from current row values. Specific to MySQL. */
/***********************************************************************/
int ha_connect::MakeRecord(char *buf)
{
char *p, *fmt, val[32];
int rc= 0;
Field* *field;
Field *fp;
my_bitmap_map *org_bitmap;
CHARSET_INFO *charset= tdbp->data_charset();
const MY_BITMAP *map;
PVAL value;
PCOL colp= NULL;
DBUG_ENTER("ha_connect::MakeRecord");
if (xtrace > 1)
#if defined(MARIADB)
printf("Maps: read=%08X write=%08X vcol=%08X defr=%08X defw=%08X\n",
*table->read_set->bitmap, *table->write_set->bitmap,
*table->vcol_set->bitmap,
*table->def_read_set.bitmap, *table->def_write_set.bitmap);
#else // !MARIADB
printf("Maps: read=%p write=%p defr=%p defw=%p\n",
*table->read_set->bitmap, *table->write_set->bitmap,
*table->def_read_set.bitmap, *table->def_write_set.bitmap);
#endif // !MARIADB
// Avoid asserts in field::store() for columns that are not updated
org_bitmap= dbug_tmp_use_all_columns(table, table->write_set);
// This is for variable_length rows
memset(buf, 0, table->s->null_bytes);
// When sorting read_set selects all columns, so we use def_read_set
map= (const MY_BITMAP *)&table->def_read_set;
// Make the pseudo record from field values
for (field= table->field; *field && !rc; field++) {
fp= *field;
#if defined(MARIADB)
if (fp->vcol_info && !fp->stored_in_db)
continue; // This is a virtual column
#endif // MARIADB
if (bitmap_is_set(map, fp->field_index)) {
// This is a used field, fill the buffer with value
for (colp= tdbp->GetColumns(); colp; colp= colp->GetNext())
if (!stricmp(colp->GetName(), GetColName(fp)))
break;
if (!colp) {
printf("Column %s not found\n", fp->field_name);
dbug_tmp_restore_column_map(table->write_set, org_bitmap);
DBUG_RETURN(HA_ERR_WRONG_IN_RECORD);
} // endif colp
value= colp->GetValue();
// All this could be better optimized
switch (value->GetType()) {
case TYPE_DATE:
if (!sdval)
sdval= AllocateValue(xp->g, TYPE_STRING, 20);
switch (fp->type()) {
case MYSQL_TYPE_DATE:
fmt= "%Y-%m-%d";
break;
case MYSQL_TYPE_TIME:
fmt= "%H:%M:%S";
break;
default:
fmt= "%Y-%m-%d %H:%M:%S";
} // endswitch type
// Get date in the format required by MySQL fields
value->FormatValue(sdval, fmt);
p= sdval->GetCharValue();
break;
case TYPE_FLOAT:
p= NULL;
break;
case TYPE_STRING:
// Passthru
default:
p= value->GetCharString(val);
} // endswitch Type
if (p) {
if (fp->store(p, strlen(p), charset, CHECK_FIELD_WARN)) {
// Avoid "error" on null fields
if (value->GetIntValue())
rc= HA_ERR_WRONG_IN_RECORD;
DBUG_PRINT("MakeRecord", (p));
} // endif store
} else
if (fp->store(value->GetFloatValue())) {
rc= HA_ERR_WRONG_IN_RECORD;
DBUG_PRINT("MakeRecord", (value->GetCharString(val)));
} // endif store
} // endif bitmap
} // endfor field
// This is copied from ha_tina and is necessary to avoid asserts
dbug_tmp_restore_column_map(table->write_set, org_bitmap);
DBUG_RETURN(rc);
} // end of MakeRecord
/***********************************************************************/
/* Set row values from a MySQL pseudo record. Specific to MySQL. */
/***********************************************************************/
int ha_connect::ScanRecord(PGLOBAL g, uchar *buf)
{
char attr_buffer[1024];
char data_buffer[1024];
int rc= 0;
PCOL colp;
PVAL value;
Field *fp;
PTDBASE tp= (PTDBASE)tdbp;
String attribute(attr_buffer, sizeof(attr_buffer),
table->s->table_charset);
my_bitmap_map *bmap= dbug_tmp_use_all_columns(table, table->read_set);
const CHARSET_INFO *charset= tdbp->data_charset();
String data_charset_value(data_buffer, sizeof(data_buffer), charset);
// Scan the pseudo record for field values and set column values
for (Field **field=table->field ; *field ; field++) {
fp= *field;
#if defined(MARIADB)
if ((fp->vcol_info && !fp->stored_in_db) ||
fp->option_struct->special)
continue; // Is a virtual column possible here ???
#endif // MARIADB
if (bitmap_is_set(table->write_set, fp->field_index)) {
for (colp= tp->GetSetCols(); colp; colp= colp->GetNext())
if (!stricmp(colp->GetName(), fp->field_name))
break;
if (!colp) {
printf("Column %s not found\n", fp->field_name);
rc= HA_ERR_WRONG_IN_RECORD;
goto err;
} else
value= colp->GetValue();
// This is a used field, fill the value from the row buffer
// All this could be better optimized
switch (value->GetType()) {
case TYPE_FLOAT:
value->SetValue(fp->val_real());
break;
case TYPE_DATE:
if (!sdval) {
sdval= (DTVAL*)AllocateValue(xp->g, TYPE_DATE, 19);
// Get date in the format produced by MySQL fields
((DTVAL*)sdval)->SetFormat(g, "YYYY-MM-DD hh:mm:ss", 19);
} // endif sdval
fp->val_str(&attribute);
sdval->SetValue_psz(attribute.c_ptr());
value->SetValue_pval(sdval);
break;
default:
fp->val_str(&attribute);
if (charset == &my_charset_bin)
{
value->SetValue_psz(attribute.c_ptr());
}
else
{
// Convert from SQL field charset to DATA_CHARSET
uint cnv_errors;
data_charset_value.copy(attribute.ptr(), attribute.length(),
attribute.charset(), charset, &cnv_errors);
value->SetValue_psz(data_charset_value.c_ptr());
}
} // endswitch Type
#ifdef NEWCHANGE
} else if (xmod == MODE_UPDATE) {
PCOL cp;
for (cp= tp->GetColumns(); cp; cp= cp->GetNext())
if (!stricmp(colp->GetName(), cp->GetName()))
break;
if (!cp) {
rc= HA_ERR_WRONG_IN_RECORD;
goto err;
} // endif cp
value->SetValue_pval(cp->GetValue());
} else // mode Insert
value->Reset();
#else
} // endif bitmap_is_set
#endif
} // endfor field
err:
dbug_tmp_restore_column_map(table->read_set, bmap);
return rc;
} // end of ScanRecord
/***********************************************************************/
/* Check change in index column. Specific to MySQL. */
/* Should be elaborated to check for real changes. */
/***********************************************************************/
int ha_connect::CheckRecord(PGLOBAL g, const uchar *oldbuf, uchar *newbuf)
{
return ScanRecord(g, newbuf);
} // end of dummy CheckRecord
/***********************************************************************/
/* Return the string representing an operator. */
/***********************************************************************/
const char *ha_connect::GetValStr(OPVAL vop, bool neg)
{
const char *val;
switch (vop) {
case OP_EQ:
val= " = ";
break;
case OP_NE:
val= " <> ";
break;
case OP_GT:
val= " > ";
break;
case OP_GE:
val= " >= ";
break;
case OP_LT:
val= " < ";
break;
case OP_LE:
val= " <= ";
break;
case OP_IN:
val= (neg) ? " NOT IN (" : " IN (";
break;
case OP_NULL:
val= " IS NULL";
break;
case OP_LIKE:
val= " LIKE ";
break;
case OP_XX:
val= " BETWEEN ";
break;
case OP_EXIST:
val= " EXISTS ";
break;
case OP_AND:
val= " AND ";
break;
case OP_OR:
val= " OR ";
break;
case OP_NOT:
val= " NOT ";
break;
case OP_CNC:
val= " || ";
break;
case OP_ADD:
val= " + ";
break;
case OP_SUB:
val= " - ";
break;
case OP_MULT:
val= " * ";
break;
case OP_DIV:
val= " / ";
break;
default:
val= " ? ";
} /* endswitch */
return val;
} // end of GetValStr
/***********************************************************************/
/* Check the WHERE condition and return an ODBC/WQL filter. */
/***********************************************************************/
PFIL ha_connect::CheckCond(PGLOBAL g, PFIL filp, AMT tty, Item *cond)
{
unsigned int i;
bool ismul= false;
PPARM pfirst= NULL, pprec= NULL, pp[2]= {NULL, NULL};
OPVAL vop= OP_XX;
if (!cond)
return NULL;
if (xtrace > 1)
printf("Cond type=%d\n", cond->type());
if (cond->type() == COND::COND_ITEM) {
char *p1, *p2;
Item_cond *cond_item= (Item_cond *)cond;
if (xtrace > 1)
printf("Cond: Ftype=%d name=%s\n", cond_item->functype(),
cond_item->func_name());
switch (cond_item->functype()) {
case Item_func::COND_AND_FUNC: vop= OP_AND; break;
case Item_func::COND_OR_FUNC: vop= OP_OR; break;
default: return NULL;
} // endswitch functype
List<Item>* arglist= cond_item->argument_list();
List_iterator<Item> li(*arglist);
Item *subitem;
p1= filp + strlen(filp);
strcpy(p1, "(");
p2= p1 + 1;
for (i= 0; i < arglist->elements; i++)
if ((subitem= li++)) {
if (!CheckCond(g, filp, tty, subitem)) {
if (vop == OP_OR)
return NULL;
else
*p2= 0;
} else {
p1= p2 + strlen(p2);
strcpy(p1, GetValStr(vop, FALSE));
p2= p1 + strlen(p1);
} // endif CheckCond
} else
return NULL;
if (*p1 != '(')
strcpy(p1, ")");
else
return NULL;
} else if (cond->type() == COND::FUNC_ITEM) {
unsigned int i;
// int n;
bool iscol, neg= FALSE;
Item_func *condf= (Item_func *)cond;
Item* *args= condf->arguments();
if (xtrace > 1)
printf("Func type=%d argnum=%d\n", condf->functype(),
condf->argument_count());
// neg= condf->
switch (condf->functype()) {
case Item_func::EQUAL_FUNC:
case Item_func::EQ_FUNC: vop= OP_EQ; break;
case Item_func::NE_FUNC: vop= OP_NE; break;
case Item_func::LT_FUNC: vop= OP_LT; break;
case Item_func::LE_FUNC: vop= OP_LE; break;
case Item_func::GE_FUNC: vop= OP_GE; break;
case Item_func::GT_FUNC: vop= OP_GT; break;
case Item_func::IN_FUNC: vop= OP_IN;
neg= ((Item_func_opt_neg *)condf)->negated;
case Item_func::BETWEEN: ismul= true; break;
default: return NULL;
} // endswitch functype
if (condf->argument_count() < 2)
return NULL;
else if (ismul && tty == TYPE_AM_WMI)
return NULL; // Not supported by WQL
for (i= 0; i < condf->argument_count(); i++) {
if (xtrace > 1)
printf("Argtype(%d)=%d\n", i, args[i]->type());
if (i >= 2 && !ismul) {
if (xtrace > 1)
printf("Unexpected arg for vop=%d\n", vop);
continue;
} // endif i
if ((iscol= args[i]->type() == COND::FIELD_ITEM)) {
const char *fnm;
ha_field_option_struct *fop;
Item_field *pField= (Item_field *)args[i];
if (pField->field->table != table)
return NULL; // Field does not belong to this table
else
fop= GetFieldOptionStruct(pField->field);
if (fop && fop->special) {
if (tty == TYPE_AM_TBL && !stricmp(fop->special, "TABID"))
fnm= "TABID";
else
return NULL;
} else if (tty == TYPE_AM_TBL)
return NULL;
else
fnm= pField->field->field_name;
if (xtrace > 1) {
printf("Field index=%d\n", pField->field->field_index);
printf("Field name=%s\n", pField->field->field_name);
} // endif xtrace
// IN and BETWEEN clauses should be col VOP list
if (i && ismul)
return NULL;
strcat(filp, fnm);
} else {
char buff[256];
String *res, tmp(buff,sizeof(buff), &my_charset_bin);
Item_basic_constant *pval= (Item_basic_constant *)args[i];
if ((res= pval->val_str(&tmp)) == NULL)
return NULL; // To be clarified
if (xtrace > 1)
printf("Value=%.*s\n", res->length(), res->ptr());
// IN and BETWEEN clauses should be col VOP list
if (!i && ismul)
return NULL;
// Append the value to the filter
if (args[i]->type() == COND::STRING_ITEM)
strcat(strcat(strcat(filp, "'"), res->ptr()), "'");
else
strncat(filp, res->ptr(), res->length());
} // endif
if (!i)
strcat(filp, GetValStr(vop, neg));
else if (vop == OP_XX && i == 1)
strcat(filp, " AND ");
else if (vop == OP_IN)
strcat(filp, (i == condf->argument_count() - 1) ? ")" : ",");
} // endfor i
} else {
if (xtrace > 1)
printf("Unsupported condition\n");
return NULL;
} // endif's type
return filp;
} // end of CheckCond
/**
Push condition down to the table handler.
@param cond Condition to be pushed. The condition tree must not be
modified by the caller.
@return
The 'remainder' condition that caller must use to filter out records.
NULL means the handler will not return rows that do not match the
passed condition.
@note
CONNECT handles the filtering only for table types that construct
an SQL or WQL query, but still leaves it to MySQL because only some
parts of the filter may be relevant.
The first suballocate finds the position where the string will be
constructed in the sarea. The second one does make the suballocation
with the proper length.
*/
const COND *ha_connect::cond_push(const COND *cond)
{
DBUG_ENTER("ha_connect::cond_push");
if (tdbp) {
AMT tty= tdbp->GetAmType();
if (tty == TYPE_AM_WMI || tty == TYPE_AM_ODBC ||
tty == TYPE_AM_TBL || tty == TYPE_AM_MYSQL) {
PGLOBAL& g= xp->g;
PFIL filp= (PFIL)PlugSubAlloc(g, NULL, 0);
*filp= 0;
if (CheckCond(g, filp, tty, (Item *)cond)) {
if (xtrace)
puts(filp);
tdbp->SetFilter(filp);
// cond= NULL; // This does not work anyway
PlugSubAlloc(g, NULL, strlen(filp) + 1);
} // endif filp
} // endif tty
} // endif tdbp
// Let MySQL do the filtering
DBUG_RETURN(cond);
} // end of cond_push
/**
Number of rows in table. It will only be called if
(table_flags() & (HA_HAS_RECORDS | HA_STATS_RECORDS_IS_EXACT)) != 0
*/
ha_rows ha_connect::records()
{
if (!valid_info)
info(HA_STATUS_VARIABLE);
if (tdbp && tdbp->Cardinality(NULL))
return stats.records;
else
return HA_POS_ERROR;
} // end of records
/**
Return an error message specific to this handler.
@param error error code previously returned by handler
@param buf pointer to String where to add error message
@return
Returns true if this is a temporary error
*/
bool ha_connect::get_error_message(int error, String* buf)
{
DBUG_ENTER("ha_connect::get_error_message");
if (xp && xp->g)
buf->copy(xp->g->Message, (uint)strlen(xp->g->Message),
system_charset_info);
DBUG_RETURN(false);
} // end of get_error_message
/**
@brief
If frm_error() is called then we will use this to determine
the file extensions that exist for the storage engine. This is also
used by the default rename_table and delete_table method in
handler.cc.
For engines that have two file name extentions (separate meta/index file
and data file), the order of elements is relevant. First element of engine
file name extentions array should be meta/index file extention. Second
element - data file extention. This order is assumed by
prepare_for_repair() when REPAIR TABLE ... USE_FRM is issued.
@note: PlugDB will handle all file creation/deletion. When dropping
a CONNECT table, we don't want the PlugDB table to be dropped or erased.
Therefore we provide a void list of extensions.
@see
rename_table method in handler.cc and
delete_table method in handler.cc
*/
static const char *ha_connect_exts[]= {
NullS
};
const char **ha_connect::bas_ext() const
{
return ha_connect_exts; // a null list, see @note above
} // end of bas_ext
/**
@brief
Used for opening tables. The name will be the name of the file.
@details
A table is opened when it needs to be opened; e.g. when a request comes in
for a SELECT on the table (tables are not open and closed for each request,
they are cached).
Called from handler.cc by handler::ha_open(). The server opens all tables by
calling ha_open() which then calls the handler specific open().
@note
For CONNECT no open can be done here because field information is not yet
updated. >>>>> TO BE CHECKED <<<<<
(Thread information could be get by using 'ha_thd')
@see
handler::ha_open() in handler.cc
*/
int ha_connect::open(const char *name, int mode, uint test_if_locked)
{
int rc= 0;
DBUG_ENTER("ha_connect::open");
if (xtrace)
printf("open: name=%s mode=%d test=%ud\n", name, mode, test_if_locked);
if (!(share= get_share(name, table)))
DBUG_RETURN(1);
thr_lock_data_init(&share->lock,&lock,NULL);
// Try to get the user if possible
if (table && table->in_use) {
PGLOBAL g= GetPlug(table->in_use);
// Try to set the database environment
if (g)
rc= (CntCheckDB(g, this, name)) ? (-2) : 0;
} // endif table
DBUG_RETURN(rc);
} // end of open
/**
@brief
Make the indexes for this table
*/
int ha_connect::optimize(THD* thd, HA_CHECK_OPT* check_opt)
{
int rc= 0;
PGLOBAL& g= xp->g;
PDBUSER dup= PlgGetUser(g);
// Ignore error on the opt file
dup->Check &= ~CHK_OPT;
tdbp= GetTDB(g);
dup->Check |= CHK_OPT;
if (tdbp || (tdbp= GetTDB(g))) {
if (!((PTDBASE)tdbp)->GetDef()->Indexable()) {
sprintf(g->Message, "Table %s is not indexable", tdbp->GetName());
rc= RC_INFO;
} else
rc= ((PTDBASE)tdbp)->ResetTableOpt(g, true);
} else
rc= HA_ERR_INTERNAL_ERROR;
return rc;
} // end of optimize
/**
@brief
Closes a table. We call the free_share() function to free any resources
that we have allocated in the "shared" structure.
@details
Called from sql_base.cc, sql_select.cc, and table.cc. In sql_select.cc it is
only used to close up temporary tables or during the process where a
temporary table is converted over to being a myisam table.
For sql_base.cc look at close_data_tables().
@see
sql_base.cc, sql_select.cc and table.cc
*/
int ha_connect::close(void)
{
int rc= 0;
DBUG_ENTER("ha_connect::close");
// If this is called by a later query, the table may have
// been already closed and the tdbp is not valid anymore.
if (tdbp && xp->last_query_id == valid_query_id)
rc= CloseTable(xp->g);
DBUG_RETURN(free_share(share) || rc);
} // end of close
/**
@brief
write_row() inserts a row. No extra() hint is given currently if a bulk load
is happening. buf() is a byte array of data. You can use the field
information to extract the data from the native byte array type.
@details
Example of this would be:
@code
for (Field **field=table->field ; *field ; field++)
{
...
}
@endcode
See ha_tina.cc for an example of extracting all of the data as strings.
ha_berekly.cc has an example of how to store it intact by "packing" it
for ha_berkeley's own native storage type.
See the note for update_row() on auto_increments and timestamps. This
case also applies to write_row().
Called from item_sum.cc, item_sum.cc, sql_acl.cc, sql_insert.cc,
sql_insert.cc, sql_select.cc, sql_table.cc, sql_udf.cc, and sql_update.cc.
@see
item_sum.cc, item_sum.cc, sql_acl.cc, sql_insert.cc,
sql_insert.cc, sql_select.cc, sql_table.cc, sql_udf.cc and sql_update.cc
*/
int ha_connect::write_row(uchar *buf)
{
int rc= 0;
PGLOBAL& g= xp->g;
DBUG_ENTER("ha_connect::write_row");
// Open the table if it was not opened yet (possible ???)
if (!IsOpened())
if (OpenTable(g)) {
if (strstr(g->Message, "read only"))
rc= HA_ERR_TABLE_READONLY;
else
rc= HA_ERR_INITIALIZATION;
DBUG_RETURN(rc);
} // endif tdbp
if (tdbp->GetMode() == MODE_ANY)
DBUG_RETURN(0);
// Set column values from the passed pseudo record
if ((rc= ScanRecord(g, buf)))
DBUG_RETURN(rc);
// Return result code from write operation
if (CntWriteRow(g, tdbp)) {
DBUG_PRINT("write_row", (g->Message));
printf("write_row: %s\n", g->Message);
rc= HA_ERR_INTERNAL_ERROR;
} // endif RC
DBUG_RETURN(rc);
} // end of write_row
/**
@brief
Yes, update_row() does what you expect, it updates a row. old_data will have
the previous row record in it, while new_data will have the newest data in it.
Keep in mind that the server can do updates based on ordering if an ORDER BY
clause was used. Consecutive ordering is not guaranteed.
@details
Currently new_data will not have an updated auto_increament record, or
and updated timestamp field. You can do these for example by doing:
@code
if (table->timestamp_field_type & TIMESTAMP_AUTO_SET_ON_UPDATE)
table->timestamp_field->set_time();
if (table->next_number_field && record == table->record[0])
update_auto_increment();
@endcode
Called from sql_select.cc, sql_acl.cc, sql_update.cc, and sql_insert.cc.
@see
sql_select.cc, sql_acl.cc, sql_update.cc and sql_insert.cc
*/
int ha_connect::update_row(const uchar *old_data, uchar *new_data)
{
int rc= 0;
PGLOBAL& g= xp->g;
DBUG_ENTER("ha_connect::update_row");
if (xtrace > 1)
printf("update_row: old=%s new=%s\n", old_data, new_data);
// Check values for possible change in indexed column
if ((rc= CheckRecord(g, old_data, new_data)))
return rc;
if (CntUpdateRow(g, tdbp)) {
DBUG_PRINT("update_row", (g->Message));
printf("update_row CONNECT: %s\n", g->Message);
rc= HA_ERR_INTERNAL_ERROR;
} // endif RC
DBUG_RETURN(rc);
} // end of update_row
/**
@brief
This will delete a row. buf will contain a copy of the row to be deleted.
The server will call this right after the current row has been called (from
either a previous rnd_nexT() or index call).
@details
If you keep a pointer to the last row or can access a primary key it will
make doing the deletion quite a bit easier. Keep in mind that the server does
not guarantee consecutive deletions. ORDER BY clauses can be used.
Called in sql_acl.cc and sql_udf.cc to manage internal table
information. Called in sql_delete.cc, sql_insert.cc, and
sql_select.cc. In sql_select it is used for removing duplicates
while in insert it is used for REPLACE calls.
@see
sql_acl.cc, sql_udf.cc, sql_delete.cc, sql_insert.cc and sql_select.cc
*/
int ha_connect::delete_row(const uchar *buf)
{
int rc= 0;
DBUG_ENTER("ha_connect::delete_row");
if (CntDeleteRow(xp->g, tdbp, false)) {
rc= HA_ERR_INTERNAL_ERROR;
printf("delete_row CONNECT: %s\n", xp->g->Message);
} // endif DeleteRow
DBUG_RETURN(rc);
} // end of delete_row
/****************************************************************************/
/* We seem to come here at the begining of an index use. */
/****************************************************************************/
int ha_connect::index_init(uint idx, bool sorted)
{
int rc;
PGLOBAL& g= xp->g;
DBUG_ENTER("index_init");
if ((rc= rnd_init(0)))
return rc;
indexing= CntIndexInit(g, tdbp, (signed)idx);
if (indexing <= 0) {
DBUG_PRINT("index_init", (g->Message));
printf("index_init CONNECT: %s\n", g->Message);
active_index= (uint)-1;
rc= -1;
} else {
if (((PTDBDOX)tdbp)->To_Kindex->GetNum_K()) {
if (((PTDBASE)tdbp)->GetFtype() != RECFM_NAF)
((PTDBDOX)tdbp)->GetTxfp()->ResetBuffer(g);
active_index= idx;
} else // Void table
indexing= 0;
rc= 0;
} // endif indexing
DBUG_RETURN(rc);
} // end of index_init
/****************************************************************************/
/* We seem to come here at the end of an index use. */
/****************************************************************************/
int ha_connect::index_end()
{
DBUG_ENTER("index_end");
active_index= -1;
DBUG_RETURN(rnd_end());
} // end of index_end
/****************************************************************************/
/* This is internally called by all indexed reading functions. */
/****************************************************************************/
int ha_connect::ReadIndexed(uchar *buf, OPVAL op, const uchar *key, uint key_len)
{
int rc;
//statistic_increment(ha_read_key_count, &LOCK_status);
switch (CntIndexRead(xp->g, tdbp, op, key, (int)key_len)) {
case RC_OK:
xp->fnd++;
rc= MakeRecord((char*)buf);
break;
case RC_EF: // End of file
rc= HA_ERR_END_OF_FILE;
break;
case RC_NF: // Not found
xp->nfd++;
rc= (op == OP_SAME) ? HA_ERR_END_OF_FILE : HA_ERR_KEY_NOT_FOUND;
break;
default: // Read error
DBUG_PRINT("ReadIndexed", (xp->g->Message));
printf("ReadIndexed: %s\n", xp->g->Message);
rc= HA_ERR_INTERNAL_ERROR;
} // endswitch RC
if (xtrace > 1)
printf("ReadIndexed: op=%d rc=%d\n", op, rc);
table->status= (rc == RC_OK) ? 0 : STATUS_NOT_FOUND;
return rc;
} // end of ReadIndexed
#ifdef NOT_USED
/**
@brief
Positions an index cursor to the index specified in the handle. Fetches the
row if available. If the key value is null, begin at the first key of the
index.
*/
int ha_connect::index_read_map(uchar *buf, const uchar *key,
key_part_map keypart_map __attribute__((unused)),
enum ha_rkey_function find_flag
__attribute__((unused)))
{
DBUG_ENTER("ha_connect::index_read");
DBUG_RETURN(HA_ERR_WRONG_COMMAND);
}
#endif // NOT_USED
/****************************************************************************/
/* This is called by handler::index_read_map. */
/****************************************************************************/
int ha_connect::index_read(uchar * buf, const uchar * key, uint key_len,
enum ha_rkey_function find_flag)
{
int rc;
OPVAL op= OP_XX;
DBUG_ENTER("ha_connect::index_read");
switch(find_flag) {
case HA_READ_KEY_EXACT: op= OP_EQ; break;
case HA_READ_AFTER_KEY: op= OP_GT; break;
case HA_READ_KEY_OR_NEXT: op= OP_GE; break;
default: DBUG_RETURN(-1);
} // endswitch find_flag
if (xtrace > 1)
printf("%p index_read: op=%d\n", this, op);
if (indexing > 0)
rc= ReadIndexed(buf, op, key, key_len);
else
rc= -1;
DBUG_RETURN(rc);
} // end of index_read
/**
@brief
Used to read forward through the index.
*/
int ha_connect::index_next(uchar *buf)
{
int rc;
DBUG_ENTER("ha_connect::index_next");
//statistic_increment(ha_read_next_count, &LOCK_status);
if (indexing > 0)
rc= ReadIndexed(buf, OP_NEXT);
else if (!indexing)
rc= rnd_next(buf);
else
rc= -1;
DBUG_RETURN(rc);
} // end of index_next
#ifdef NOT_USED
/**
@brief
Used to read backwards through the index.
*/
int ha_connect::index_prev(uchar *buf)
{
DBUG_ENTER("ha_connect::index_prev");
DBUG_RETURN(HA_ERR_WRONG_COMMAND);
}
#endif // NOT_USED
/**
@brief
index_first() asks for the first key in the index.
@details
Called from opt_range.cc, opt_sum.cc, sql_handler.cc, and sql_select.cc.
@see
opt_range.cc, opt_sum.cc, sql_handler.cc and sql_select.cc
*/
int ha_connect::index_first(uchar *buf)
{
int rc;
DBUG_ENTER("ha_connect::index_first");
if (indexing > 0)
rc= ReadIndexed(buf, OP_FIRST);
else if (indexing < 0)
rc= -1;
else if (CntRewindTable(xp->g, tdbp)) {
table->status= STATUS_NOT_FOUND;
rc= -1;
} else
rc= rnd_next(buf);
DBUG_RETURN(rc);
} // end of index_first
#ifdef NOT_USED
/**
@brief
index_last() asks for the last key in the index.
@details
Called from opt_range.cc, opt_sum.cc, sql_handler.cc, and sql_select.cc.
@see
opt_range.cc, opt_sum.cc, sql_handler.cc and sql_select.cc
*/
int ha_connect::index_last(uchar *buf)
{
DBUG_ENTER("ha_connect::index_last");
DBUG_RETURN(HA_ERR_WRONG_COMMAND);
}
#endif // NOT_USED
/****************************************************************************/
/* This is called to get more rows having the same index value. */
/****************************************************************************/
int ha_connect::index_next_same(uchar *buf, const uchar *key, uint keylen)
{
int rc;
DBUG_ENTER("ha_connect::index_next_same");
//statistic_increment(ha_read_next_count, &LOCK_status);
if (!indexing)
rc= rnd_next(buf);
else if (indexing > 0)
rc= ReadIndexed(buf, OP_SAME);
else
rc= -1;
DBUG_RETURN(rc);
} // end of index_next_same
/**
@brief
rnd_init() is called when the system wants the storage engine to do a table
scan. See the example in the introduction at the top of this file to see when
rnd_init() is called.
@details
Called from filesort.cc, records.cc, sql_handler.cc, sql_select.cc, sql_table.cc,
and sql_update.cc.
@note
We always call open and extern_lock/start_stmt before comming here.
@see
filesort.cc, records.cc, sql_handler.cc, sql_select.cc, sql_table.cc and sql_update.cc
*/
int ha_connect::rnd_init(bool scan)
{
PGLOBAL g= ((table && table->in_use) ? GetPlug(table->in_use) :
(xp) ? xp->g : NULL);
DBUG_ENTER("ha_connect::rnd_init");
if (xtrace)
printf("%p in rnd_init: scan=%d\n", this, scan);
if (g) {
// Open the table if it was not opened yet (possible ???)
if (!IsOpened()) {
if (!table || xmod == MODE_INSERT)
DBUG_RETURN(HA_ERR_INITIALIZATION);
if (OpenTable(g, xmod == MODE_DELETE))
#if defined(MARIADB)
DBUG_RETURN(HA_ERR_INITIALIZATION);
#else // !MARIADB
DBUG_RETURN(HA_ERR_INTERNAL_ERROR);
#endif // !MARIADB
} else
void(CntRewindTable(g, tdbp)); // Read from beginning
} // endif g
xp->nrd= xp->fnd= xp->nfd= 0;
xp->tb1= my_interval_timer();
DBUG_RETURN(0);
} // end of rnd_init
/**
@brief
Not described.
@note
The previous version said:
Stop scanning of table. Note that this may be called several times during
execution of a sub select.
=====> This has been moved to external lock to avoid closing subselect tables.
*/
int ha_connect::rnd_end()
{
int rc= 0;
DBUG_ENTER("ha_connect::rnd_end");
// If this is called by a later query, the table may have
// been already closed and the tdbp is not valid anymore.
// if (tdbp && xp->last_query_id == valid_query_id)
// rc= CloseTable(xp->g);
DBUG_RETURN(rc);
} // end of rnd_end
/**
@brief
This is called for each row of the table scan. When you run out of records
you should return HA_ERR_END_OF_FILE. Fill buff up with the row information.
The Field structure for the table is the key to getting data into buf
in a manner that will allow the server to understand it.
@details
Called from filesort.cc, records.cc, sql_handler.cc, sql_select.cc, sql_table.cc,
and sql_update.cc.
@see
filesort.cc, records.cc, sql_handler.cc, sql_select.cc, sql_table.cc and sql_update.cc
*/
int ha_connect::rnd_next(uchar *buf)
{
int rc;
DBUG_ENTER("ha_connect::rnd_next");
//statistic_increment(ha_read_rnd_next_count, &LOCK_status);
#if !defined(MARIADB)
if (!tdbp) // MySQL ignores error from rnd_init
DBUG_RETURN(HA_ERR_INTERNAL_ERROR);
#endif // !MARIADB
if (tdbp->GetMode() == MODE_ANY) {
// We will stop on next read
if (!stop) {
stop= true;
DBUG_RETURN(RC_OK);
} else
DBUG_RETURN(HA_ERR_END_OF_FILE);
} // endif Mode
switch (CntReadNext(xp->g, tdbp)) {
case RC_OK:
rc= MakeRecord((char*)buf);
break;
case RC_EF: // End of file
rc= HA_ERR_END_OF_FILE;
break;
case RC_NF: // Not found
rc= HA_ERR_RECORD_DELETED;
break;
default: // Read error
printf("rnd_next CONNECT: %s\n", xp->g->Message);
rc= (records()) ? HA_ERR_INTERNAL_ERROR : HA_ERR_END_OF_FILE;
break;
} // endswitch RC
#ifndef DBUG_OFF
if (rc || !(xp->nrd++ % 16384)) {
ulonglong tb2= my_interval_timer();
double elapsed= (double) (tb2 - xp->tb1) / 1000000000ULL;
DBUG_PRINT("rnd_next", ("rc=%d nrd=%u fnd=%u nfd=%u sec=%.3lf\n",
rc, (uint)xp->nrd, (uint)xp->fnd,
(uint)xp->nfd, elapsed));
xp->tb1= tb2;
xp->fnd= xp->nfd= 0;
} // endif nrd
#endif
table->status= (!rc) ? 0 : STATUS_NOT_FOUND;
DBUG_RETURN(rc);
} // end of rnd_next
/**
@brief
position() is called after each call to rnd_next() if the data needs
to be ordered. You can do something like the following to store
the position:
@code
my_store_ptr(ref, ref_length, current_position);
@endcode
@details
The server uses ref to store data. ref_length in the above case is
the size needed to store current_position. ref is just a byte array
that the server will maintain. If you are using offsets to mark rows, then
current_position should be the offset. If it is a primary key like in
BDB, then it needs to be a primary key.
Called from filesort.cc, sql_select.cc, sql_delete.cc, and sql_update.cc.
@see
filesort.cc, sql_select.cc, sql_delete.cc and sql_update.cc
*/
void ha_connect::position(const uchar *record)
{
DBUG_ENTER("ha_connect::position");
if (((PTDBASE)tdbp)->GetDef()->Indexable())
my_store_ptr(ref, ref_length, (my_off_t)((PTDBASE)tdbp)->GetRecpos());
DBUG_VOID_RETURN;
} // end of position
/**
@brief
This is like rnd_next, but you are given a position to use
to determine the row. The position will be of the type that you stored in
ref. You can use my_get_ptr(pos,ref_length) to retrieve whatever key
or position you saved when position() was called.
@details
Called from filesort.cc, records.cc, sql_insert.cc, sql_select.cc, and sql_update.cc.
@note
Is this really useful? It was never called even when sorting.
@see
filesort.cc, records.cc, sql_insert.cc, sql_select.cc and sql_update.cc
*/
int ha_connect::rnd_pos(uchar *buf, uchar *pos)
{
int rc;
PTDBASE tp= (PTDBASE)tdbp;
DBUG_ENTER("ha_connect::rnd_pos");
if (!tp->SetRecpos(xp->g, (int)my_get_ptr(pos, ref_length)))
rc= rnd_next(buf);
else
rc= HA_ERR_KEY_NOT_FOUND;
DBUG_RETURN(rc);
} // end of rnd_pos
/**
@brief
::info() is used to return information to the optimizer. See my_base.h for
the complete description.
@details
Currently this table handler doesn't implement most of the fields really needed.
SHOW also makes use of this data.
You will probably want to have the following in your code:
@code
if (records < 2)
records= 2;
@endcode
The reason is that the server will optimize for cases of only a single
record. If, in a table scan, you don't know the number of records, it
will probably be better to set records to two so you can return as many
records as you need. Along with records, a few more variables you may wish
to set are:
records
deleted
data_file_length
index_file_length
delete_length
check_time
Take a look at the public variables in handler.h for more information.
Called in filesort.cc, ha_heap.cc, item_sum.cc, opt_sum.cc, sql_delete.cc,
sql_delete.cc, sql_derived.cc, sql_select.cc, sql_select.cc, sql_select.cc,
sql_select.cc, sql_select.cc, sql_show.cc, sql_show.cc, sql_show.cc, sql_show.cc,
sql_table.cc, sql_union.cc, and sql_update.cc.
@see
filesort.cc, ha_heap.cc, item_sum.cc, opt_sum.cc, sql_delete.cc, sql_delete.cc,
sql_derived.cc, sql_select.cc, sql_select.cc, sql_select.cc, sql_select.cc,
sql_select.cc, sql_show.cc, sql_show.cc, sql_show.cc, sql_show.cc, sql_table.cc,
sql_union.cc and sql_update.cc
*/
int ha_connect::info(uint flag)
{
bool pure= false;
PGLOBAL g= GetPlug((table) ? table->in_use : NULL);
DBUG_ENTER("ha_connect::info");
if (xtrace)
printf("%p In info: flag=%u valid_info=%d\n", this, flag, valid_info);
if (!valid_info) {
// tdbp must be available to get updated info
if (!tdbp || xp->CheckQuery(valid_query_id)) {
if (xmod == MODE_ANY) { // Pure info, not a query
pure= true;
// xmod= MODE_READ;
} // endif xmod
// tdbp= OpenTable(g, xmod == MODE_DELETE);
tdbp= GetTDB(g);
} // endif tdbp
valid_info= CntInfo(g, tdbp, &xinfo);
} // endif valid_info
if (flag & HA_STATUS_VARIABLE) {
stats.records= xinfo.records;
stats.deleted= 0;
stats.data_file_length= xinfo.data_file_length;
stats.index_file_length= 0;
stats.delete_length= 0;
stats.check_time= 0;
stats.mean_rec_length= xinfo.mean_rec_length;
} // endif HA_STATUS_VARIABLE
if (flag & HA_STATUS_CONST) {
// This is imported from the previous handler and must be reconsidered
stats.max_data_file_length= LL(4294967295);
stats.max_index_file_length= LL(4398046510080);
stats.create_time= 0;
data_file_name= xinfo.data_file_name;
index_file_name= NULL;
// sortkey= (uint) - 1; // Table is not sorted
ref_length= sizeof(int); // Pointer size to row
table->s->db_options_in_use= 03;
stats.block_size= 1024;
table->s->keys_in_use.set_prefix(table->s->keys);
table->s->keys_for_keyread= table->s->keys_in_use;
// table->s->keys_for_keyread.subtract(table->s->read_only_keys);
table->s->db_record_offset= 0;
} // endif HA_STATUS_CONST
if (flag & HA_STATUS_ERRKEY) {
errkey= 0;
} // endif HA_STATUS_ERRKEY
if (flag & HA_STATUS_TIME)
stats.update_time= 0;
if (flag & HA_STATUS_AUTO)
stats.auto_increment_value= 1;
if (tdbp && pure)
CloseTable(g); // Not used anymore
DBUG_RETURN(0);
} // end of info
/**
@brief
extra() is called whenever the server wishes to send a hint to
the storage engine. The myisam engine implements the most hints.
ha_innodb.cc has the most exhaustive list of these hints.
@note
This is not yet implemented for CONNECT.
@see
ha_innodb.cc
*/
int ha_connect::extra(enum ha_extra_function operation)
{
DBUG_ENTER("ha_connect::extra");
DBUG_RETURN(0);
} // end of extra
/**
@brief
Used to delete all rows in a table, including cases of truncate and cases where
the optimizer realizes that all rows will be removed as a result of an SQL statement.
@details
Called from item_sum.cc by Item_func_group_concat::clear(),
Item_sum_count_distinct::clear(), and Item_func_group_concat::clear().
Called from sql_delete.cc by mysql_delete().
Called from sql_select.cc by JOIN::reinit().
Called from sql_union.cc by st_select_lex_unit::exec().
@see
Item_func_group_concat::clear(), Item_sum_count_distinct::clear() and
Item_func_group_concat::clear() in item_sum.cc;
mysql_delete() in sql_delete.cc;
JOIN::reinit() in sql_select.cc and
st_select_lex_unit::exec() in sql_union.cc.
*/
int ha_connect::delete_all_rows()
{
int rc= 0;
PGLOBAL g= xp->g;
DBUG_ENTER("ha_connect::delete_all_rows");
// Close and reopen the table so it will be deleted
rc= CloseTable(g);
if (!(OpenTable(g))) {
if (CntDeleteRow(g, tdbp, true)) {
printf("%s\n", g->Message);
rc= HA_ERR_INTERNAL_ERROR;
} // endif
} else
rc= HA_ERR_INITIALIZATION;
DBUG_RETURN(rc);
} // end of delete_all_rows
/**
@brief
This create a lock on the table. If you are implementing a storage engine
that can handle transacations look at ha_berkely.cc to see how you will
want to go about doing this. Otherwise you should consider calling flock()
here. Hint: Read the section "locking functions for mysql" in lock.cc to understand
this.
@details
Called from lock.cc by lock_external() and unlock_external(). Also called
from sql_table.cc by copy_data_between_tables().
@note
Following what we did in the MySQL XDB handler, we use this call to actually
physically open the table. This could be reconsider when finalizing this handler
design, which means we have a better understanding of what MariaDB does.
@see
lock.cc by lock_external() and unlock_external() in lock.cc;
the section "locking functions for mysql" in lock.cc;
copy_data_between_tables() in sql_table.cc.
*/
int ha_connect::external_lock(THD *thd, int lock_type)
{
int rc= 0;
bool del= false;
MODE newmode;
PGLOBAL g= GetPlug(thd);
DBUG_ENTER("ha_connect::external_lock");
if (xtrace)
printf("%p external_lock: lock_type=%d\n", this, lock_type);
if (!g)
DBUG_RETURN(-99);
// Action will depend on lock_type
switch (lock_type) {
case F_WRLCK:
newmode= MODE_WRITE;
break;
case F_RDLCK:
newmode= MODE_READ;
break;
case F_UNLCK:
default:
newmode= MODE_ANY;
} // endswitch mode
if (newmode == MODE_ANY) {
// This is unlocking, do it by closing the table
if (xp->CheckQueryID())
rc= 2; // Logical error ???
else if (tdbp) {
if (tdbp->GetMode() == MODE_ANY && *tdbp->GetName() == '#'
&& xp->tabp) {
PDOSDEF defp1= (PDOSDEF)((PTDBASE)tdbp)->GetDef();
PDOSDEF defp2= (PDOSDEF)xp->tabp->GetDef();
PIXDEF xp1, xp2, sxp;
// Look for new created indexes
for (xp1= defp1->GetIndx(); xp1; xp1= xp1->GetNext()) {
for (xp2= defp2->GetIndx(); xp2; xp2= xp2->GetNext())
if (!stricmp(xp1->GetName(), xp2->GetName()))
break; // Index already made
if (!xp2) {
// Here we do make the index on tabp
sxp= xp1->GetNext();
xp1->SetNext(NULL);
xp->tabp->MakeIndex(g, xp1, true);
xp1->SetNext(sxp);
} // endif xp2
} // endfor xp1
// Look for dropped indexes
for (xp2= defp2->GetIndx(); xp2; xp2= xp2->GetNext()) {
for (xp1= defp1->GetIndx(); xp1; xp1= xp1->GetNext())
if (!stricmp(xp1->GetName(), xp2->GetName()))
break; // Index not to drop
if (!xp1) {
// Here we erase the index file
sxp= xp2->GetNext();
xp2->SetNext(NULL);
defp2->DeleteIndexFile(g, xp2);
xp2->SetNext(sxp);
} // endif xp1
} // endfor xp2
} // endif Mode
rc= CloseTable(g);
} // endif tdbp
DBUG_RETURN(rc);
} // endif MODE_ANY
if (xtrace) {
printf("%p external_lock: cmdtype=%d\n", this, thd->lex->sql_command);
printf("Cmd=%.*s\n", (int) thd->query_string.length(),
thd->query_string.str());
} // endif xtrace
// Next code is temporarily replaced until sql_command is set
stop= false;
if (newmode == MODE_WRITE) {
switch (thd->lex->sql_command) {
case SQLCOM_INSERT:
case SQLCOM_CREATE_TABLE:
case SQLCOM_LOAD:
case SQLCOM_INSERT_SELECT:
newmode= MODE_INSERT;
break;
// case SQLCOM_REPLACE:
// case SQLCOM_REPLACE_SELECT:
// newmode= MODE_UPDATE; // To be checked
// break;
case SQLCOM_DELETE:
case SQLCOM_DELETE_MULTI:
del= true;
case SQLCOM_TRUNCATE:
newmode= MODE_DELETE;
break;
case SQLCOM_UPDATE:
case SQLCOM_UPDATE_MULTI:
newmode= MODE_UPDATE;
break;
case SQLCOM_SELECT:
case SQLCOM_OPTIMIZE:
newmode= MODE_READ;
break;
case SQLCOM_DROP_TABLE:
case SQLCOM_RENAME_TABLE:
case SQLCOM_ALTER_TABLE:
newmode= MODE_ANY;
break;
case SQLCOM_DROP_INDEX:
case SQLCOM_CREATE_INDEX:
newmode= MODE_ANY;
stop= true;
break;
default:
printf("Unsupported sql_command=%d", thd->lex->sql_command);
sprintf(g->Message, "Unsupported sql_command=%d", thd->lex->sql_command);
DBUG_RETURN(HA_ERR_INTERNAL_ERROR);
} // endswitch newmode
} else if (newmode == MODE_READ) {
switch (thd->lex->sql_command) {
case SQLCOM_INSERT:
case SQLCOM_CREATE_TABLE:
case SQLCOM_LOAD:
case SQLCOM_INSERT_SELECT:
// case SQLCOM_REPLACE:
// case SQLCOM_REPLACE_SELECT:
case SQLCOM_DELETE:
case SQLCOM_DELETE_MULTI:
case SQLCOM_TRUNCATE:
case SQLCOM_UPDATE:
case SQLCOM_UPDATE_MULTI:
case SQLCOM_SELECT:
case SQLCOM_OPTIMIZE:
break;
case SQLCOM_DROP_INDEX:
case SQLCOM_CREATE_INDEX:
stop= true;
case SQLCOM_DROP_TABLE:
case SQLCOM_RENAME_TABLE:
case SQLCOM_ALTER_TABLE:
newmode= MODE_ANY;
break;
default:
printf("Unsupported sql_command=%d", thd->lex->sql_command);
sprintf(g->Message, "Unsupported sql_command=%d", thd->lex->sql_command);
DBUG_RETURN(HA_ERR_INTERNAL_ERROR);
} // endswitch newmode
} // endif's newmode
if (xtrace)
printf("New mode=%d\n", newmode);
// If this is the start of a new query, cleanup the previous one
if (xp->CheckCleanup()) {
tdbp= NULL;
valid_info= false;
} // endif CheckCleanup
if (xtrace)
printf("Calling CntCheckDB db=%s\n", GetDBName(NULL));
// Set or reset the good database environment
if (CntCheckDB(g, this, GetDBName(NULL))) {
printf("%p external_lock: %s\n", this, g->Message);
rc= HA_ERR_INTERNAL_ERROR;
// This can NOT be called without open called first, but
// the table can have been closed since then
} else if (!tdbp || xp->CheckQuery(valid_query_id) || xmod != newmode) {
if (tdbp)
CloseTable(g);
xmod= newmode;
if (!table)
rc= 3; // Logical error
// Delay open until used fields are known
} // endif tdbp
if (xtrace)
printf("external_lock: rc=%d\n", rc);
DBUG_RETURN(rc);
} // end of external_lock
/**
@brief
The idea with handler::store_lock() is: The statement decides which locks
should be needed for the table. For updates/deletes/inserts we get WRITE
locks, for SELECT... we get read locks.
@details
Before adding the lock into the table lock handler (see thr_lock.c),
mysqld calls store lock with the requested locks. Store lock can now
modify a write lock to a read lock (or some other lock), ignore the
lock (if we don't want to use MySQL table locks at all), or add locks
for many tables (like we do when we are using a MERGE handler).
Berkeley DB, for example, changes all WRITE locks to TL_WRITE_ALLOW_WRITE
(which signals that we are doing WRITES, but are still allowing other
readers and writers).
When releasing locks, store_lock() is also called. In this case one
usually doesn't have to do anything.
In some exceptional cases MySQL may send a request for a TL_IGNORE;
This means that we are requesting the same lock as last time and this
should also be ignored. (This may happen when someone does a flush
table when we have opened a part of the tables, in which case mysqld
closes and reopens the tables and tries to get the same locks at last
time). In the future we will probably try to remove this.
Called from lock.cc by get_lock_data().
@note
In this method one should NEVER rely on table->in_use, it may, in fact,
refer to a different thread! (this happens if get_lock_data() is called
from mysql_lock_abort_for_thread() function)
@see
get_lock_data() in lock.cc
*/
THR_LOCK_DATA **ha_connect::store_lock(THD *thd,
THR_LOCK_DATA **to,
enum thr_lock_type lock_type)
{
if (lock_type != TL_IGNORE && lock.type == TL_UNLOCK)
lock.type=lock_type;
*to++ = &lock;
return to;
}
/**
@brief
Used to delete a table. By the time delete_table() has been called all
opened references to this table will have been closed (and your globally
shared references released). The variable name will just be the name of
the table. You will need to remove any files you have created at this point.
@details
If you do not implement this, the default delete_table() is called from
handler.cc and it will delete all files with the file extensions returned
by bas_ext().
Called from handler.cc by delete_table and ha_create_table(). Only used
during create if the table_flag HA_DROP_BEFORE_CREATE was specified for
the storage engine.
@see
delete_table and ha_create_table() in handler.cc
*/
int ha_connect::delete_table(const char *name)
{
DBUG_ENTER("ha_connect::delete_table");
/* This is not implemented but we want someone to be able that it works. */
DBUG_RETURN(0);
}
/**
@brief
Given a starting key and an ending key, estimate the number of rows that
will exist between the two keys.
@details
end_key may be empty, in which case determine if start_key matches any rows.
Called from opt_range.cc by check_quick_keys().
@see
check_quick_keys() in opt_range.cc
*/
ha_rows ha_connect::records_in_range(uint inx, key_range *min_key,
key_range *max_key)
{
ha_rows rows;
DBUG_ENTER("ha_connect::records_in_range");
if (indexing < 0 || inx != active_index)
index_init(inx, false);
if (xtrace)
printf("records_in_range: inx=%d indexing=%d\n", inx, indexing);
if (indexing > 0) {
int nval;
uint len[2];
const uchar *key[2];
bool incl[2];
key_part_map kmap[2];
key[0]= (min_key) ? min_key->key : NULL;
key[1]= (max_key) ? max_key->key : NULL;
len[0]= (min_key) ? min_key->length : 0;
len[1]= (max_key) ? max_key->length : 0;
incl[0]= (min_key) ? (min_key->flag == HA_READ_KEY_EXACT) : false;
incl[1]= (max_key) ? (max_key->flag == HA_READ_AFTER_KEY) : false;
kmap[0]= (min_key) ? min_key->keypart_map : 0;
kmap[1]= (max_key) ? max_key->keypart_map : 0;
if ((nval= CntIndexRange(xp->g, tdbp, key, len, incl, kmap)) < 0)
rows= HA_POS_ERROR;
else
rows= (ha_rows)nval;
} else if (indexing < 0)
rows= HA_POS_ERROR;
else
rows= 100000000; // Don't use missing index
DBUG_RETURN(rows);
} // end of records_in_range
/**
Convert an ISO-8859-1 column name to UTF-8
*/
char *ha_connect::encode(PGLOBAL g, char *cnm)
{
char *buf= (char*)PlugSubAlloc(g, NULL, strlen(cnm) * 3);
uint dummy_errors;
uint32 len= copy_and_convert(buf, strlen(cnm) * 3,
&my_charset_utf8_general_ci,
cnm, strlen(cnm),
&my_charset_latin1,
&dummy_errors);
buf[len]= '\0';
return buf;
} // end of Encode
/**
Store field definition for create.
@return
Return 0 if ok
*/
bool ha_connect::add_fields(THD *thd, void *alt_info,
LEX_STRING *field_name,
enum_field_types type,
char *length, char *decimals,
uint type_modifier,
// Item *default_value, Item *on_update_value,
LEX_STRING *comment,
// char *change,
// List<String> *interval_list,
CHARSET_INFO *cs,
// uint uint_geom_type,
void *vcolinfo,
engine_option_value *create_options)
{
register Create_field *new_field;
Alter_info *alter_info= (Alter_info*)alt_info;
Virtual_column_info *vcol_info= (Virtual_column_info *)vcolinfo;
DBUG_ENTER("ha_connect::add_fields");
if (check_string_char_length(field_name, "", NAME_CHAR_LEN,
system_charset_info, 1))
{
my_error(ER_TOO_LONG_IDENT, MYF(0), field_name->str); /* purecov: inspected */
DBUG_RETURN(1); /* purecov: inspected */
}
#if 0
if (type_modifier & PRI_KEY_FLAG)
{
Key *key;
lex->col_list.push_back(new Key_part_spec(*field_name, 0));
key= new Key(Key::PRIMARY, null_lex_str,
&default_key_create_info,
0, lex->col_list, NULL);
alter_info->key_list.push_back(key);
lex->col_list.empty();
}
if (type_modifier & (UNIQUE_FLAG | UNIQUE_KEY_FLAG))
{
Key *key;
lex->col_list.push_back(new Key_part_spec(*field_name, 0));
key= new Key(Key::UNIQUE, null_lex_str,
&default_key_create_info, 0,
lex->col_list, NULL);
alter_info->key_list.push_back(key);
lex->col_list.empty();
}
if (default_value)
{
/*
Default value should be literal => basic constants =>
no need fix_fields()
We allow only one function as part of default value -
NOW() as default for TIMESTAMP type.
*/
if (default_value->type() == Item::FUNC_ITEM &&
!(((Item_func*)default_value)->functype() == Item_func::NOW_FUNC &&
type == MYSQL_TYPE_TIMESTAMP))
{
my_error(ER_INVALID_DEFAULT, MYF(0), field_name->str);
DBUG_RETURN(1);
}
else if (default_value->type() == Item::NULL_ITEM)
{
default_value= 0;
if ((type_modifier & (NOT_NULL_FLAG | AUTO_INCREMENT_FLAG)) ==
NOT_NULL_FLAG)
{
my_error(ER_INVALID_DEFAULT, MYF(0), field_name->str);
DBUG_RETURN(1);
}
}
else if (type_modifier & AUTO_INCREMENT_FLAG)
{
my_error(ER_INVALID_DEFAULT, MYF(0), field_name->str);
DBUG_RETURN(1);
}
}
if (on_update_value && type != MYSQL_TYPE_TIMESTAMP)
{
my_error(ER_INVALID_ON_UPDATE, MYF(0), field_name->str);
DBUG_RETURN(1);
}
#endif // 0
if (!(new_field= new Create_field()) ||
new_field->init(thd, field_name->str, type, length, decimals, type_modifier,
NULL, NULL, comment, NULL,
NULL, cs, 0, vcol_info,
create_options))
DBUG_RETURN(1);
alter_info->create_list.push_back(new_field);
//lex->last_field=new_field;
DBUG_RETURN(0);
}
/**
@brief
pre_create() is called when creating a table with no columns.
@details
When pre_create() is called the .frm file have not already been
created. You can overwrite some definitions at this point but the
main purpose of it is to define the columns for some table types.
@note
Not really implemented yet.
*/
bool ha_connect::pre_create(THD *thd, HA_CREATE_INFO *create_info,
void *alt_info)
{
char spc= ',', qch= 0;
const char *typn= "?";
const char *fncn= "?";
const char *user;
char *fn, *dsn, *tab, *db, *host, *pwd, *prt, *sep; // *csn;
#if defined(WIN32)
char *nsp= NULL, *cls= NULL;
#endif // WIN32
int port= MYSQL_PORT, hdr= 0, mxr= 0;
uint tm, fnc= FNC_NO, supfnc= (FNC_NO | FNC_COL);
bool b= false, ok= false, dbf= false;
TABTYPE ttp= TAB_UNDEF;
LEX_STRING *comment, *name, *val;
MEM_ROOT *mem= thd->mem_root;
CHARSET_INFO *cs;
Alter_info *alter_info= (Alter_info*)alt_info;
engine_option_value *pov, *start= create_info->option_list, *end= NULL;
PQRYRES qrp;
PCOLRES crp;
PGLOBAL g= GetPlug(thd);
if (!g)
return true;
fn= dsn= tab= host= pwd= prt= sep= NULL;
user= NULL;
db= thd->db; // Default value
// Get the useful create options
for (pov= start; pov; pov= pov->next) {
if (!stricmp(pov->name.str, "table_type")) {
typn= pov->value.str;
ttp= GetTypeID(typn);
} else if (!stricmp(pov->name.str, "file_name")) {
fn= pov->value.str;
} else if (!stricmp(pov->name.str, "tabname")) {
tab= pov->value.str;
} else if (!stricmp(pov->name.str, "db_name")) {
db= pov->value.str;
} else if (!stricmp(pov->name.str, "catfunc")) {
fncn= pov->value.str;
fnc= GetFuncID(fncn);
} else if (!stricmp(pov->name.str, "sep_char")) {
sep= pov->value.str;
spc= (!strcmp(sep, "\\t")) ? '\t' : *sep;
} else if (!stricmp(pov->name.str, "qchar")) {
qch= *pov->value.str;
} else if (!stricmp(pov->name.str, "quoted")) {
if (!qch)
qch= '"';
} else if (!stricmp(pov->name.str, "header")) {
hdr= atoi(pov->value.str);
} else if (!stricmp(pov->name.str, "option_list")) {
host= GetListOption("host", pov->value.str, "localhost");
user= GetListOption("user", pov->value.str, "root");
pwd= GetListOption("password", pov->value.str);
prt= GetListOption("port", pov->value.str);
port= (prt) ? atoi(prt) : MYSQL_PORT;
#if defined(WIN32)
nsp= GetListOption("namespace", pov->value.str);
cls= GetListOption("class", pov->value.str);
#endif // WIN32
mxr= atoi(GetListOption("maxerr", pov->value.str, "0"));
} // endelse option_list
end= pov;
} // endfor pov
// Check table type
if (ttp == TAB_UNDEF || ttp == TAB_NIY) {
sprintf(g->Message, "Unknown Table_type '%s'", typn);
push_warning(thd, MYSQL_ERROR::WARN_LEVEL_WARN, 0, g->Message);
strcpy(g->Message, "Using Table_type DOS");
push_warning(thd, MYSQL_ERROR::WARN_LEVEL_WARN, 0, g->Message);
ttp= TAB_DOS;
typn= "DOS";
name= thd->make_lex_string(NULL, "table_type", 10, true);
val= thd->make_lex_string(NULL, typn, strlen(typn), true);
pov= new(mem) engine_option_value(*name, *val, false, &start, &end);
} // endif ttp
// Test whether columns must be specified
if (alter_info->create_list.elements)
return false;
dbf= ttp == TAB_DBF;
if (!tab && !(fnc & (FNC_TABLE | FNC_COL)))
tab= (char*)create_info->alias;
switch (ttp) {
#if defined(ODBC_SUPPORT)
case TAB_ODBC:
if (!(dsn= create_info->connect_string.str)
&& !(fnc & (FNC_DSN | FNC_DRIVER)))
sprintf(g->Message, "Missing %s connection string", typn);
else
ok= true;
supfnc |= (FNC_TABLE | FNC_DSN | FNC_DRIVER);
break;
#endif // ODBC_SUPPORT
case TAB_DBF:
dbf= true;
// Passthru
case TAB_CSV:
if (!fn)
sprintf(g->Message, "Missing %s file name", typn);
else
ok= true;
break;
#if defined(MYSQL_SUPPORT)
case TAB_MYSQL:
if (!user)
user= "root"; // Avoid crash
ok= true;
break;
#endif // MYSQL_SUPPORT
#if defined(WIN32)
case TAB_WMI:
ok= true;
break;
#endif // WIN32
default:
sprintf(g->Message, "Cannot get column info for table type %s", typn);
} // endif ttp
// Check for supported catalog function
if (ok && !(supfnc & fnc)) {
sprintf(g->Message, "Unsupported catalog function %s for table type %s",
fncn, typn);
ok= false;
} // endif supfnc
if (ok) {
char *length, *decimals, *cnm, *rem;
int i, len, dec, typ;
enum_field_types type;
PDBUSER dup= PlgGetUser(g);
PCATLG cat= (dup) ? dup->Catalog : NULL;
if (cat)
cat->SetDataPath(g, thd->db);
else
return true; // Should never happen
switch (ttp) {
case TAB_DBF:
qrp= DBFColumns(g, fn, fnc == FNC_COL);
break;
#if defined(ODBC_SUPPORT)
case TAB_ODBC:
switch (fnc) {
case FNC_NO:
case FNC_COL:
qrp= ODBCColumns(g, dsn, tab, NULL, fnc == FNC_COL);
break;
case FNC_TABLE:
qrp= ODBCTables(g, dsn, tab, true);
break;
case FNC_DSN:
qrp= ODBCDataSources(g, true);
break;
case FNC_DRIVER:
qrp= ODBCDrivers(g, true);
break;
default:
sprintf(g->Message, "invalid catfunc %s", fncn);
} // endswitch info
break;
#endif // ODBC_SUPPORT
#if defined(MYSQL_SUPPORT)
case TAB_MYSQL:
qrp= MyColumns(g, host, db, user, pwd, tab,
NULL, port, false, fnc == FNC_COL);
break;
#endif // MYSQL_SUPPORT
case TAB_CSV:
qrp= CSVColumns(g, fn, spc, qch, hdr, mxr, fnc == FNC_COL);
break;
#if defined(WIN32)
case TAB_WMI:
qrp= WMIColumns(g, nsp, cls, fnc == FNC_COL);
break;
#endif // WIN32
default:
strcpy(g->Message, "System error in pre_create");
break;
} // endswitch ttp
if (!qrp) {
my_message(ER_UNKNOWN_ERROR, g->Message, MYF(0));
return true;
} // endif qrp
if (fnc != FNC_NO) {
// Catalog table
for (crp=qrp->Colresp; !b && crp; crp= crp->Next) {
cnm= encode(g, crp->Name);
name= thd->make_lex_string(NULL, cnm, strlen(cnm), true);
type= PLGtoMYSQL(crp->Type, dbf);
len= crp->Length;
length= (char*)PlugSubAlloc(g, NULL, 8);
sprintf(length, "%d", len);
decimals= NULL;
comment= thd->make_lex_string(NULL, "", 0, true);
// Now add the field
b= add_fields(thd, alt_info, name, type, length, decimals,
NOT_NULL_FLAG, comment, NULL, NULL, NULL);
} // endfor crp
} else // Not a catalog table
for (i= 0; !b && i < qrp->Nblin; i++) {
rem= "";
typ= len= dec= 0;
length= "";
decimals= NULL;
tm= NOT_NULL_FLAG;
cs= NULL;
for (crp= qrp->Colresp; crp; crp= crp->Next)
switch (crp->Fld) {
case FLD_NAME:
cnm= encode(g, crp->Kdata->GetCharValue(i));
name= thd->make_lex_string(NULL, cnm, strlen(cnm), true);
break;
case FLD_TYPE:
typ= crp->Kdata->GetIntValue(i);
break;
case FLD_PREC:
len= crp->Kdata->GetIntValue(i);
break;
case FLD_SCALE:
if ((dec= crp->Kdata->GetIntValue(i))) {
decimals= (char*)PlugSubAlloc(g, NULL, 8);
sprintf(decimals, "%d", dec);
} else
decimals= NULL;
break;
case FLD_NULL:
if (crp->Kdata->GetIntValue(i))
tm= 0; // Nullable
break;
case FLD_REM:
rem= crp->Kdata->GetCharValue(i);
break;
// case FLD_CHARSET:
// No good because remote table is already translated
// if (*(csn= crp->Kdata->GetCharValue(i)))
// cs= get_charset_by_name(csn, 0);
// break;
default:
break; // Ignore
} // endswitch Fld
#if defined(ODBC_SUPPORT)
if (ttp == TAB_ODBC) {
int plgtyp;
// typ must be PLG type, not SQL type
if (!(plgtyp= TranslateSQLType(typ, dec, len))) {
sprintf(g->Message, "Unsupported SQL type %d", typ);
my_message(ER_UNKNOWN_ERROR, g->Message, MYF(0));
return true;
} else
typ= plgtyp;
// Some data sources do not count dec in length
if (typ == TYPE_FLOAT)
len += (dec + 2); // To be safe
} // endif ttp
#endif // ODBC_SUPPORT
// Make the arguments as required by add_fields
type= PLGtoMYSQL(typ, true);
length= (char*)PlugSubAlloc(g, NULL, 8);
sprintf(length, "%d", len);
comment= thd->make_lex_string(NULL, rem, strlen(rem), true);
// Now add the field
b= add_fields(thd, alt_info, name, type, length, decimals,
tm, comment, cs, NULL, NULL);
} // endfor i
return b;
} // endif ok
my_message(ER_UNKNOWN_ERROR, g->Message, MYF(0));
return true;
} // end of pre_create
/**
@brief
create() is called to create a database. The variable name will have the name
of the table.
@details
When create() is called you do not need to worry about
opening the table. Also, the .frm file will have already been
created so adjusting create_info is not necessary. You can overwrite
the .frm file at this point if you wish to change the table
definition, but there are no methods currently provided for doing
so.
Called from handle.cc by ha_create_table().
@note
Currently we do some checking on the create definitions and stop
creating if an error is found. We wish we could change the table
definition such as providing a default table type. However, as said
above, there are no method to do so.
@see
ha_create_table() in handle.cc
*/
int ha_connect::create(const char *name, TABLE *table_arg,
HA_CREATE_INFO *create_info)
{
int rc= RC_OK;
bool dbf;
Field* *field;
Field *fp;
TABLE *st= table; // Probably unuseful
PIXDEF xdp, pxd= NULL, toidx= NULL;
PGLOBAL g= GetPlug(table_arg->in_use);
DBUG_ENTER("ha_connect::create");
PTOS options= GetTableOptionStruct(table_arg);
// CONNECT engine specific table options:
DBUG_ASSERT(options);
if (options->data_charset)
{
const CHARSET_INFO *data_charset;
if (!(data_charset= get_charset_by_csname(options->data_charset,
MY_CS_PRIMARY, MYF(0))))
{
my_error(ER_UNKNOWN_CHARACTER_SET, MYF(0), options->data_charset);
DBUG_RETURN(HA_ERR_INTERNAL_ERROR);
}
if (GetTypeID(options->type) == TAB_XML &&
data_charset != &my_charset_utf8_general_ci)
{
my_printf_error(ER_UNKNOWN_ERROR,
"DATA_CHARSET='%s' is not supported for TABLE_TYPE=XML",
MYF(0), options->data_charset);
DBUG_RETURN(HA_ERR_INTERNAL_ERROR);
}
}
if (!g) {
rc= HA_ERR_INTERNAL_ERROR;
DBUG_RETURN(rc);
} else
dbf= (GetTypeID(options->type) == TAB_DBF);
// Check column types
for (field= table_arg->field; *field; field++) {
fp= *field;
#if defined(MARIADB)
if (fp->vcol_info && !fp->stored_in_db)
continue; // This is a virtual column
#endif // MARIADB
switch (fp->type()) {
case MYSQL_TYPE_SHORT:
case MYSQL_TYPE_LONG:
case MYSQL_TYPE_FLOAT:
case MYSQL_TYPE_DOUBLE:
case MYSQL_TYPE_TIMESTAMP:
case MYSQL_TYPE_DATE:
case MYSQL_TYPE_TIME:
case MYSQL_TYPE_DATETIME:
case MYSQL_TYPE_YEAR:
case MYSQL_TYPE_NEWDATE:
case MYSQL_TYPE_VARCHAR:
case MYSQL_TYPE_LONGLONG:
break; // Ok
case MYSQL_TYPE_VAR_STRING:
case MYSQL_TYPE_STRING:
case MYSQL_TYPE_DECIMAL:
case MYSQL_TYPE_NEWDECIMAL:
case MYSQL_TYPE_INT24:
break; // To be checked
case MYSQL_TYPE_TINY:
case MYSQL_TYPE_BIT:
case MYSQL_TYPE_NULL:
case MYSQL_TYPE_ENUM:
case MYSQL_TYPE_SET:
case MYSQL_TYPE_TINY_BLOB:
case MYSQL_TYPE_MEDIUM_BLOB:
case MYSQL_TYPE_LONG_BLOB:
case MYSQL_TYPE_BLOB:
case MYSQL_TYPE_GEOMETRY:
default:
// fprintf(stderr, "Unsupported type column %s\n", fp->field_name);
sprintf(g->Message, "Unsupported type for column %s",
fp->field_name);
rc= HA_ERR_INTERNAL_ERROR;
my_printf_error(ER_UNKNOWN_ERROR,
"Unsupported type for column '%s'",
MYF(0), fp->field_name);
DBUG_RETURN(rc);
} // endswitch type
if (dbf) {
bool b= false;
if ((b= strlen(fp->field_name) > 10))
sprintf(g->Message, "DBF: Column name '%s' is too long (max=10)",
fp->field_name);
else if ((b= fp->field_length > 255))
sprintf(g->Message, "DBF: Column length too big for '%s' (max=255)",
fp->field_name);
if (b) {
my_message(ER_UNKNOWN_ERROR, g->Message, MYF(0));
rc= HA_ERR_INTERNAL_ERROR;
DBUG_RETURN(rc);
} // endif b
} // endif dbf
} // endfor field
// Check whether indexes were specified
table= table_arg; // Used by called functions
// Get the index definitions
for (int n= 0; (unsigned)n < table->s->keynames.count; n++) {
if (xtrace)
printf("Getting created index %d info\n", n + 1);
xdp= GetIndexInfo(n);
if (pxd)
pxd->SetNext(xdp);
else
toidx= xdp;
pxd= xdp;
} // endfor n
if (toidx) {
PDBUSER dup= PlgGetUser(g);
PCATLG cat= (dup) ? dup->Catalog : NULL;
DBUG_ASSERT(cat);
if (cat)
cat->SetDataPath(g, table_arg->in_use->db);
if ((rc= optimize(NULL, NULL))) {
printf("Create rc=%d %s\n", rc, g->Message);
rc= HA_ERR_INTERNAL_ERROR;
} else
CloseTable(g);
} // endif toidx
table= st;
DBUG_RETURN(rc);
} // end of create
/**
check_if_incompatible_data() called if ALTER TABLE can't detect otherwise
if new and old definition are compatible
@details If there are no other explicit signs like changed number of
fields this function will be called by compare_tables()
(sql/sql_tables.cc) to decide should we rewrite whole table or only .frm
file.
*/
bool ha_connect::check_if_incompatible_data(HA_CREATE_INFO *info,
uint table_changes)
{
//ha_table_option_struct *param_old, *param_new;
DBUG_ENTER("ha_connect::check_if_incompatible_data");
// TO DO: implement it.
if (table)
push_warning(table->in_use, MYSQL_ERROR::WARN_LEVEL_WARN, 0,
"The current version of CONNECT did not check what you changed in ALTER. Use on your own risk");
DBUG_RETURN(COMPATIBLE_DATA_YES);
}
struct st_mysql_storage_engine connect_storage_engine=
{ MYSQL_HANDLERTON_INTERFACE_VERSION };
struct st_mysql_daemon unusable_connect=
{ MYSQL_DAEMON_INTERFACE_VERSION };
mysql_declare_plugin(connect)
{
MYSQL_STORAGE_ENGINE_PLUGIN,
&connect_storage_engine,
"CONNECT",
"Olivier Bertrand",
"Direct access to external data, including many file formats",
PLUGIN_LICENSE_GPL,
connect_init_func, /* Plugin Init */
connect_done_func, /* Plugin Deinit */
0x0001 /* 0.1 */,
NULL, /* status variables */
NULL, /* system variables */
NULL, /* config options */
0, /* flags */
}
mysql_declare_plugin_end;
#if defined(MARIADB)
maria_declare_plugin(connect)
{
MYSQL_STORAGE_ENGINE_PLUGIN,
&connect_storage_engine,
"CONNECT",
"Olivier Bertrand",
"Direct access to external data, including many file formats",
PLUGIN_LICENSE_GPL,
connect_init_func, /* Plugin Init */
connect_done_func, /* Plugin Deinit */
0x0001, /* version number (0.1) */
NULL, /* status variables */
NULL, /* system variables */
"0.1", /* string version */
MariaDB_PLUGIN_MATURITY_EXPERIMENTAL /* maturity */
},
{
MYSQL_DAEMON_PLUGIN,
&unusable_connect,
"UNUSABLE",
"Olivier Bertrand",
"Unusable Daemon",
PLUGIN_LICENSE_PROPRIETARY,
NULL, /* Plugin Init */
NULL, /* Plugin Deinit */
0x0101, /* version number (1.1) */
NULL, /* status variables */
NULL, /* system variables */
"1.01.00.000" , /* version, as a string */
MariaDB_PLUGIN_MATURITY_EXPERIMENTAL /* maturity */
}
maria_declare_plugin_end;
#endif // MARIADB
Reference in a new issue View git blame Copy permalink