XXVI. dbx Functions


The dbx module is a database abstraction layer (db 'X', where 'X' is a supported database). The dbx functions allow you to access all supported databases using a single calling convention. The dbx-functions themselves do not interface directly to the databases, but interface to the modules that are used to support these databases.

注: 本扩展已被移动到 PECL 库中且自以下版本起不再被绑定到 PHP 中:5.1.0.


To be able to use a database with the dbx-module, the module must be either linked or loaded into PHP, and the database module must be supported by the dbx-module. Currently, the following databases are supported, but others will follow:

Documentation for adding additional database support to dbx can be found at http://www.guidance.nl/php/dbx/doc/.


In order to have these functions available, you must compile PHP with dbx support by using the --enable-dbx option and all options for the databases that will be used, e.g. for MySQL you must also specify --with-mysql=[DIR]. To get other supported databases to work with the dbx-module refer to their specific documentation.


这些函数的行为受 php.ini 的影响。

表格 1. DBX Configuration Options

dbx.colnames_case"unchanged"PHP_INI_SYSTEMAvailable since PHP 4.3.0.
有关 PHP_INI_* 常量进一步的细节与定义参见附录 G


dbx.colnames_case string

Columns names can be returned "unchanged" or converted to "uppercase" or "lowercase". This directive can be overridden with a flag to dbx_query().


There are two resource types used in the dbx module. The first one is the link-object for a database connection, the second a result-object which holds the result of a query.


以下常量由本扩展模块定义,因此只有在本扩展模块被编译到 PHP 中,或者在运行时被动态加载后才有效。

DBX_MYSQL (integer)

DBX_ODBC (integer)

DBX_PGSQL (integer)

DBX_MSSQL (integer)

DBX_FBSQL (integer)

DBX_OCI8 (integer) (available from PHP 4.3.0)

DBX_SYBASECT (integer)

DBX_SQLITE (integer) (PHP 5)






DBX_COLNAMES_UNCHANGED (integer) (available from PHP 4.3.0)

DBX_COLNAMES_UPPERCASE (integer) (available from PHP 4.3.0)

DBX_COLNAMES_LOWERCASE (integer) (available from PHP 4.3.0)

DBX_CMP_NATIVE (integer)

DBX_CMP_TEXT (integer)

DBX_CMP_NUMBER (integer)

DBX_CMP_ASC (integer)

DBX_CMP_DESC (integer)

dbx_close -- Close an open connection/database
dbx_compare -- Compare two rows for sorting purposes
dbx_connect -- Open a connection/database
dbx_error --  Report the error message of the latest function call in the module (not just in the connection)
dbx_escape_string --  Escape a string so it can safely be used in an sql-statement
dbx_fetch_row -- Fetches rows from a query-result that had the DBX_RESULT_UNBUFFERED flag set
dbx_query -- Send a query and fetch all results (if any)
dbx_sort --  Sort a result from a dbx_query by a custom sort function

add a note add a note User Contributed Notes
bart at mediawave dot nl
09-Jul-2004 09:45
Simple function for returning paged result sets.

// use: pagedQuery($link, $sql, $rows_per_page, $current_page)

function pagedQuery($link, $sql, $nrows = 10, $page = 1) {
$handle = dbx_query($link, $sql, DBX_RESULT_UNBUFFERED);
$result = new stdClass;
$result->link = $this->link;
$result->page = $page;
$result->data = array();
$result->info['name'] = array();
$start = ($page - 1) * $nrows;
$end = $start + $nrows;
$result->rows = 0;
   while (
$row = dbx_fetch_row($handle)) {
       if ((
$result->rows > $start) && ($result->rows <= $end)) {
$result->data[] = $row;           
   if (
$result->data[0]) {
$result->cols = count($result->data[0]) / 2;
$result->info['name'] = array_slice(array_keys($result->data[0]), $result->cols, $result->cols);
$result->lastPage = ceil($result->rows / $nrows);
bart at mediawave dot nl
02-Jul-2004 11:12
Would like to confirm that dbx is at least three times faster than adodb with my application. However, once loaded, adodb has much more features and probably performs better in some situations then dbx.

One of the nice features that adodb has, is the way you can quickly make a dropdown menu from a query. Here's a similar function that works with a dbx result object:


// $result is the dbx_query result (It expects two fields)
// $name is the name of the dropdown field
// $selected is the option that should be selected
// $firstrow takes a string like "myvalue:myname"
// $attr can be used to add some extra attributes

function dropDown($result, $name, $selected, $firstRow, $attr) {
$s = '<select name="'.$name.'" '.$attr.'>';
   if (
is_string($firstRow))  {
$row = explode(':', $firstRow);
$s .= '<option value="'.$row[0].'">'.$row[1].'</option>';
   } else
$s .= '<option></option>';
   foreach (
$result->data as $row) {
       if (
$row[0] == $selected)
$s .= '<option value="'.$row[0].'" selected="selected">'.htmlspecialchars($row[1]).'</option>';
$s .= '<option value="'.$row[0].'">'.htmlspecialchars($row[1]).'</option>';

rhcf at linux dot ime dot usp dot br
26-Nov-2003 05:52

dbx_query allocate all retrieved data in an array on memory. If the query result is big (bigger then the memory_limit on php.ini), the result will be empty. So use dbx wih care.
carl [at] carlthompson [dot] net
08-Jun-2003 04:49
The above benchmark is very misleading. Unfortunately, I found out the hard way by porting my app to ADODB. Most PHP apps are going to be very short-lived therefore startup / setup time is a large factor. This benchmark does not measure that. Just include()ing the main ADODB file added 27ms to my app's startup time! Considering my app took only 16ms to run in total using DBX, the idea of switching to ADODB was DOA before running a single query.

In the end, the total time for my app went from 16ms with DBX to 49ms with ADODB. ADODB's powerful features are obviously not worth that kind of overhead.

Carl Thompson
jlim at natsoft dot com dot my
06-Feb-2003 11:35
For some benchmarks of dbx connecting to mysql compared to native mysql api, adodb, and others, see: