Coding techniques, using null variables?

[m_tt]

For example, which is the preferred method for defining and calling a variable that can be "unset" or "off"

$variable = null;

if (!is_null($variable)) { }

$variable = 0;

if (!empty($variable)) { }
if ($variable != 0) { }

// This would return true for 0
if (isset($variable)) { }

Really just fishing for opinions here, no real question 😉

[devinemke]

m@tt wrote:

which is the preferred method for defining and calling a variable that can be "unset" or "off"

just to be clear: any variable that is defined is not "unset". if you want to just initialize a variable that you will then later assign something meaningful you can use anything you want:

$var = NULL;
$var = false;
$var = 0;
$var = '';

all of the above will evaluate to BOOLEAN false. you should read up on Type Juggling.

[Shrike]

isset() && !empty() are my usual approach, particularly for incoming HTTP data.

[mjax]

Most people using empty() seem to forget that empty('0') also evaluates to true. If you use empty() to check if form fields are filled (as lots of PHP developers seem to do), it means that the user cannot enter '0' in these fields. Suppose you are asking for 'Number of children' and you want to make sure this field is filled in. Using empty() will not work in this case.

[Shrike]

It's true that is a bit of a gotcha. I was trying to illustrate that using isset() and empty() need to be used in conjunction, one is generally not sufficient. But following your example you would need to use the type equality operator ===.

[Weedpacket]

Using empty() usually makes isset() unnecessary, since empty() evaluates to true for unset variables also (without triggering a notice); conversely, empty() being false would imply that isset() is true. Therefore, isset() && !empty() simplifies down to !empty().

Where they would be used together is if the variable needs to be set, but not to anything equivalent to false.