The column names if you use PGSQL_ASSOC or PGSQL_BOTH are always in lowercase, no matter what the name is in the database or in the query.
(PHP 4, PHP 5, PHP 7, PHP 8)
pg_fetch_array — 获取一行作为数组
pg_fetch_array() 返回与所提取的行（元组/记录）相一致的数组。
pg_fetch_array() 是 pg_fetch_row() 的扩展版本。在结果数组中不仅以数字索引（字段编号）方式存放数据，还用关联索引（字段名）存储数据。它默认存储两个索引。
注意: 此函数将 NULL 字段设置为 PHP
null值。
使用 pg_fetch_array() 并不比 pg_fetch_row() 慢的明显，而且在使用中提供了更大的方便。
resultPgSql\Result 实例，由 pg_query()、pg_query_params() 或者 pg_execute()（等）返回。
row
要获取的结果中的行号。行从 0 向上编号。如果省略或为 null，则获取下一行。
mode
An optional parameter that controls how the returned array is indexed.
mode is a constant and can take the following values:
PGSQL_ASSOC, PGSQL_NUM and PGSQL_BOTH.
Using PGSQL_NUM, the function will return an array with numerical indices,
using PGSQL_ASSOC it will return only associative indices
while PGSQL_BOTH will return both numerical and associative indices.
数字方式（从 0 开始）或关联方式（按字段名索引）或两者共同索引的 array。array
中的每个值都表示为 string。数据库 NULL 值作为 null 返回。
如果 row 超过集合中的行数、没有更多行或任何其他错误，则返回 false。从 SELECT 以外的查询结果中获取也将返回 false。
| 版本 | 说明 |
|---|---|
| 8.1.0 |
现在 result 参数接受 PgSql\Result
实例，之前接受 resource。
|
示例 #1 pg_fetch_array() 示例
<?php
$conn = pg_pconnect("dbname=publisher");
if (!$conn) {
echo "An error occurred.\n";
exit;
}
$result = pg_query($conn, "SELECT author, email FROM authors");
if (!$result) {
echo "An error occurred.\n";
exit;
}
$arr = pg_fetch_array($result, 0, PGSQL_NUM);
echo $arr[0] . " <- Row 1 Author\n";
echo $arr[1] . " <- Row 1 E-mail\n";
// The row parameter is optional; NULL can be passed instead,
// to pass a result_type. Successive calls to pg_fetch_array
// will return the next row.
$arr = pg_fetch_array($result, NULL, PGSQL_ASSOC);
echo $arr["author"] . " <- Row 2 Author\n";
echo $arr["email"] . " <- Row 2 E-mail\n";
$arr = pg_fetch_array($result);
echo $arr["author"] . " <- Row 3 Author\n";
echo $arr[1] . " <- Row 3 E-mail\n";
?>The column names if you use PGSQL_ASSOC or PGSQL_BOTH are always in lowercase, no matter what the name is in the database or in the query.As of PHP 4.1.0, you can now use code such as the following to iterate through a result set:
$conn = pg_connect("host=localhost dbname=whatever");
$result = pg_exec($conn, "select * from table");
while ($row = pg_fetch_array($result))
{
echo "data: ".$row["data"];
}
Can be a nice little time saver, PHP with MySQL has supported this for a while but I'm glad to see it extended to PostgreSQL...PGSQL_BOTH is the default, meaning your array size will be doubled.
If you specify this field (result type), include no quotes around it or you won't get any data, not even an error.
Here's my wrapper function:
function SQL_fetch_array($result_ndx, $row, $result_type=PGSQL_ASSOC) {
return pg_fetch_array($result_ndx, $row, $result_type);In addition to returning "false if there are no more rows", pg_fetch_array will also trigger an E_WARNING. You can temporarily turn that error reporting level off and suck out all your data like so:
<?php
$errRptLvl = error_reporting();
error_reporting($errRptLvl & ~(E_WARNING));
list($i,$j)=array(0,0);
while ($selection[$i++] = $this->fetchArray($j++)); // (fetchArray is a pg_fetch_array wrapper.)
error_reporting($errRptLvl); // Restore error reporting level.
unset($selection[$i-1]); // Delete the last, empty row.
return $selection;
?>Note that when using PGSQL_BOTH, numerically and associatively indexed fields are separate variables and treated as such:
<?php
$res = pg_query("Select 'foo' as bar");
$data = pg_fetch_array($res, 0, PGSQL_BOTH);
var_dump($data);
// Array(2)
// {
// [0] => string(3) "foo"
// ["bar"] => string(3) "foo"
// }
// This won't affect $data['bar']
$data[0] = 'bar';
var_dump($data);
// Array(2)
// {
// [0] => string(3) "bar"
// ["bar"] => string(3) "foo"
// }
?>
If you want to have reference binding between your numeric and associative indexes, you'll have to establish that yourself:
<?php
$result = pg_query("Select 'foo' as bar");
$data = pg_fetch_row($result);
// Establish references between column name/number
$from = $data;
foreach($from as $cx => $value)
{
$key = pg_field_name($result, $cx);
if (is_string($key)) $data[$key] =& $data[$cx];
}
var_dump($data);
// Array(2)
// {
// [0] => &string(3) "foo"
// ["bar"] => &string(3) "foo"
// }
// Note the reference binding between $data[0] and $data['bar']
$data[0] = 'baz';
var_dump($data);
// Array(2)
// {
// [0] => &string(3) "baz"
// ["bar"] => &string(3) "baz"
// }
?>(Timesaver) Be aware of the fact that keys in array returned by this function are (well, at least as of 4.2.3) of the same case as SQL column names (e.g. if your column name is ID then key name is also ID, not id or Id), and the keys in associative array are CASE SENSITIVE!!! So don't be surprised if you get unexpected results. Double check SQL column names and the key names.In response to eth0's comment below about SELECT'ing from two tables where the tables have columns with the same names, you can get around this problem like this:
"SELECT table1.foo AS foo1, table2.foo AS foo2 FROM table1, table2"
In the associative array returned, the keys will be "foo1" and "foo2".Hopefully most people realize this on their own, but the examples below where people tried to get creative with getting numerical or associative (not both) keys in the result are rather pointless. See the pg_fetch_assoc() and pg_fetch_row() for the built in functions that do this automatically. It's generally a better idea to use one of these other functions unless you *need* to access fields by both collumn name *and* index.Just because it is not really clear how to specify the result type, I poste this message.
I wrote a wrapper function which looks like this:
<?php
function db_fetch_array ($result, $row = NULL, $result_type = PGSQL_ASSOC)
{
$return = @pg_fetch_array ($result, $row, $result_type);
return $return;
}
?>
I think this way it is quite comfortable to get the arrays you want.Just remember when you 'or die' to close your table(s) or you may get a confused look from non-internet explorer users.Please remember that if you have for example a table Customers with "cust_ID", "name" and "address" and another table Users with "u_ID","name" and "other" and then you SELECT WHERE cust_ID=u_ID then you'll get in the result array ONLY ONE "name" field, precisely the last one resulted from the select!!!