Smarty

[blaz]

How to use a smarty IF statement depending on if an array is empty?

I have an array called $errors. This doesn't seem to work...

[AZTEK]

Humm good point I just tried and I dont know how you would do that without changing Smarty
_________________
"Imagine a school with children that can read and write, but with teachers who cannot, and you have a metaphor of the Information Age in which we live." -Peter Cochrane

[messju]

hmm

{if $array ne ''} ... tests if the array is *not* empty.

possible tests for empty array:
{if $array eq ''}
{if !$array}
{if not $array}

these all are not type-specific (they also match for empty strings, empty objects, false, null, 0 and 0.0). but types are something for php, not for smarty


edit: correction, {if $array eq ''} does not work, but the other two do.

[sweatje]

I use

[boots]

I tended to use sweatje's method (and sometimes the ne "" method--I think that worked in previous versions??) but I decided to see just what the difference was performance wise (since modifiers don't come for free).

Using my handy-dandy timeblock plugin and the following template code:

{timeblock}
{if $array|@count == 0}EMPTY{/if}
{/timeblock}

{timeblock}
{if !$array}EMPTY{/if}
{/timeblock}

{timeblock}
{if not $array}EMPTY{/if}
{/timeblock}

I ran tests both with empty arrays and arrays populated with 1000 elements. Cutting to the chase:

$array|@count == 0 is about 10 times slower than the other two methods.

!$array and not $array run about neck-and-neck.

For fairness, you must run each timeblock separately. Otherwise, PHP seems to cache (?) the lookup result on the array causing subsequent checks to be faster.

The moral is:

If you get advice from more than one person, pay particular attention to what messju is saying!!

[messju]

ah, no boots. i know of bugs i committed to smarty-cvs. of course i fixed 'em asap when i noticed, but my choice wasn't the best before then.

all in all: "good" is a vector and "fast" is only a component.

but your post makes me ponder, if modifier-calls to php-functions should be resolved at compile-time instead of run-time. this would bring the 'count' into the compiled template literally. i think this will have an impact of native-php modifier's execution time.

just some thoughts, that are totally off-topic here.

[boots]

messju--I didn't mean to put undue pressure on you--I was just having fun there Apologies to sweatje as well: as I said, I made the same choice as you did -- it seems natural to test strings against "" and arrays against @count. I was even using $var|count == 0 to test strings for consistency with $array|@count == 0.

[Wom.bat]

yeah... boots... but you know, often a |@count can't be avoided...
I _always_ initialize arrays ($foo = array();) before I populate and assign them, so I can't do a !$array in an IF... ;/

[boots]

Wom.bat:

[Wom.bat]

w000t
well... in php, it would be false

sorry for that mistake

[boots]

Wom.Bat:

Yeah, I see why you thought that and my feeling is that the behaviour that Smarty is exhibiting is not by policy but instead an artifact of the implementation. I too would have expected your result (ie. consistent with PHP) but it was in timing the different checking styles that I noticed this behaviour. So I wasn't trying to "correct" you--only showing what the actual behaviour appears to be

[Wom.bat]

hehe not a problem, thank you very much

P.S: have you tested the math patch yet?