CakeFest 2024: The Official CakePHP Conference

pg_convert

(PHP 4 >= 4.3.0, PHP 5, PHP 7, PHP 8)

pg_convert 将关联的数组值转换为适合 SQL 语句的格式

说明

pg_convert(
    PgSql\Connection $connection,
    string $table_name,
    array $values,
    int $flags = 0
): array|false

pg_convert() 检查 values 中的值并将其转换为为适用于 SQL 语句的值。pg_convert() 的前提条件是现有的表 table_name 中具有的列至少有 values 中的单元那么多。table_name 中的字段名以及字段值必须匹配 values 中的索引,并且相应的数据类型必须兼容。成功时返回包含转换后的值的数组,否则返回 false

注意:

接受 bool 值并转换为 PostgreSQL boolean 值。还支持 bool 值的字符串表示。null 转换为 PostgreSQL NULL。

参数

connection

PgSql\Connection 实例。

table_name

要转换类型的表名。

values

要转换的数据。

flags

PGSQL_CONV_IGNORE_DEFAULTPGSQL_CONV_FORCE_NULLPGSQL_CONV_IGNORE_NOT_NULL 的任意数量组合。

返回值

转换值后的 array, 或者在失败时返回 false

错误/异常

当字段的值或类型不能正确匹配 PostgreSQL 的类型时,会抛出 ValueErrorTypeError

更新日志

版本 说明
8.3.0 当字段的值或类型不能正确匹配 PostgreSQL 的类型时,现在会抛出 ValueErrorTypeError 错误;之前触发 E_WARNING
8.1.0 现在 connection 参数接受 PgSql\Connection 实例,之前接受 resource

示例

示例 #1 pg_convert() 示例

<?php
$dbconn
= pg_connect('dbname=foo');

$tmp = array(
'author' => 'Joe Thackery',
'year' => 2005,
'title' => 'My Life, by Joe Thackery'
);

$vals = pg_convert($dbconn, 'authors', $tmp);
?>

参见

add a note

User Contributed Notes 5 notes

up
2
Anonymous
19 years ago
The only options that I see are:

PGSQL_CONV_IGNORE_DEFAULT - Do not use DEAFULT value by removing field from returned array
PGSQL_CONV_FORCE_NULL - Convert to NULL if string is null string
PGSQL_CONV_IGNORE_NOT_NULL - Ignore NOT NULL constraints

These are constants, so don't quote them or anything.
up
0
gorhas at gmail dot com
9 years ago
There is a problem when using interval.
If in the array
"time_pause" => '00:30:00'
and time_pause is an interval
the insert fails
pg_insert(): '00:30:00' does not match with '^(@?[ \t]+)?((([-+]?[ \t]+)?[0-9]+(\.[0-9]*)?[ ...
up
0
dharana at dharana dot net
20 years ago
I've found "options" possible values:

PG_CONV_CHECK - check only
PG_CONV_STRICT - raise warning for non fatal error
PG_CONV_QUOTE - add quote around values for vchar, text datetime.
PG_CONV_SLASH - add slashes if it needed.
PG_CONV_NULLCHK - check values are defined for NOT NULL fields.
PG_CONV_NO_DEFAULT - ignore default value even if value is empty string.
up
-1
VLroyrenn
6 years ago
Another thing that's not well documented is that (as of PHP 7.0/7.1) pg_convert doesn't like non-scalar types and will fail (but not throw just emit an E_WARNING and return false) if you pass it anything other than a string or a number, including an array or something like a DateTime. If you want to insert those types, you actually have to convert those yourself.

Also, somewhat surprisingly, $table_name is not compatible with the output of pg_​escape_​identifier, or seemingly any other kind of escaping.
up
-1
Hayley Watson
6 years ago
This will only apply the appropriate escaping and such appropriate for embedding the PHP value into an SQL statement.

It does (by default) check for nulls when the column is marked NOT NULL, and it will complain about trying to convert strings for an integer column (floats will be truncated).

Beyond the barest checking of syntax, however, it does NOT verify that the given value is a legitimate value for the column type.

<?php
// Assuming smallints.smallintis a smallint (-32768..32767) type column
foreach([-1234,
1234,
0,
32767,
-
32768,
32768, // bogus value for smallint type
45.8, // gets truncated to 45
400000, // bogus value for smallint type
] as $smallint)
{
$tmp = ['smallint' => $smallint];
$vals = pg_convert($dbconn, 'smallints', ['smallint' => $smallint]);
echo
$vals['"smallint"'],"\n"; // Notice the column name is also made SQL-safe
}

// Assuming uuids.uuid is a UUID type column
foreach(['a0eebc99-9c0b-4ef8-bb6d-6bb9bd380a11',
'A0EEBC99-9C0B-4EF8-BB6D-6BB9BD380A11',
'a0eebc999c0b4ef8bb6d6bb9bd380a11',
'{a0eebc99-9c0b-4ef8-bb6d-6bb9bd380a11}',
'Invalid Not-a-UUID',
'{a0eebc99-9c0b4ef8-bb6d6bb9-bd380a11}',
'a0ee-bc99-9c0b-4ef8-bb6d-6bb9-bd38-0a11',
] as
$uuid)
{
$tmp = ['uuid' => $uuid];
$vals = pg_convert($dbconn, 'uuids', ['uuid' => $uuid]);
echo
$vals['"uuid"'],"\n";
}

?>

All of the above data values will be "converted" - even the invalid ones - without complaint.
To Top