Searching an array for multiple keywords

[litebearer]

Evening, all...

I can't seem to get my thoughts properly wrapped around this problem.

Scenario:

Searching an array of data for matches to keywords. Create a new array from the matches.

01. From 1 to 3 keywords possible. Match must have ALL keywords.

02. Case insensitive

03. $needle1, $needle2 and $needle3 represent the possible keywords

04. $old_haystack represents an element from the original array

05. $haystack represents lowercase version of $old_haystack

06. must be compatible with PHP versions 3, 4 and 5

I came up with this sledge-hammer approach, looping through all of elements of the original array and runing this function on each element.

However, I am sure there is a more elegant, refined and effective approach, but what is it?

[code]
<?PHP
function validate_words($needle1, $needle2, $needle3, $old_haystack) {
$haystack = strtolower($old_haystack);
$new_element = "";
$kwl[0] = strlen(trim($needle1);
$kwl[1] = strlen(trim($needle2);
$kwl[2] = strlen(trim($needle3);
$add_element = 0;

if(($kwl[0]>0) AND ($kwl[1]>0) AND ($kwl[2]>0) AND (substr_count($haystack, $needle1)>0) AND (substr_count($haystack, $needle2)>0) AND (substr_count($haystack, $needle3)>0)) {
// add element to new array
return 1;
}

if(($kwl[0]>0) AND ($kwl[1]>0) AND ($kwl[2]<1) AND (substr_count($haystack, $needle1)>0)  AND (substr_count($haystack, $needle2)>0)){
// add element to new array
return 1;
}

if(($kwl[0]>0) AND ($kwl[1]<1) AND ($kwl[2]>0) AND (substr_count($haystack, $needle1)>0) AND (substr_count($haystack, $needle3)>0)) {
// add element to new array
return 1;
}

if(($kwl[0]>0) AND ($kwl[1]<1) AND ($kwl[2]<1) AND (substr_count($haystack, $needle1)>0)) {
// add element to new array
return 1;
}

if(($kwl[0]<1) AND ($kwl[1]>0) AND ($kwl[2]>0) AND (substr_count($haystack, $needle2)>0) AND (substr_count($haystack, $needle3)>0)) {
// add element to new array
return 1;
}

if(($kwl[0]><1) AND ($kwl[1]>0) AND ($kwl[2]<1) AND (substr_count($haystack, $needle2)>0)) {
// add element to new array
return 1;
}

if(($kwl[0]<1)  AND ($kwl[1]<1) AND ($kwl[2]>0) AND (substr_count($haystack, $needle3)>0)) {
// add element to new array
return 1;
}
return 0;
}

?>
[/code]

Thanks,

Lite...

[emehrkay]

im having trouble understanding exactly what you're trying to do.

you use substr_count, but i thought $haystack was an array.

it might just be easier to place your needles in one array and your haystack in another and loop though both - again, i dont understand what you are doing. can you explain a little bit more?

[litebearer]

Hmmm, ok, a little more definition...

First, I know that using mysql would simplify things; however, for various reasons, I am using a flatfile for this particular project.

each line in the file represents a 'record'

each record has several 'fields'

user desires to search the file for all records that contain certain keywords

user can use either 1, 2 or 3 keywords.

user can select that either ALL or ANY of the keywords are in a record

Matching using ANY is no brainer

The code I previously posted will, in the end, accomplish the task.

My basic question was/is:  Is there a more effective, simpler method of reaching the same end result

Not sure if that makes it any clearer.

Thanks,

Lite...

[akitchin]

try something like this:

[code]<?php
// define the needles
$needle1 = 'one';
$needle2 = 'two';
$needle3 = 'three';

// map a function to all levels of an array
function deep_map($function, $val)
{
if (is_array($val))
{
$return_arr = array();
foreach ($val AS $k => $v)
$return_arr["$k"] = deep_map($function, $val["$k"]);
return $return_arr;
}
else
{
$command = "return (".$function."(\$val));";
return eval($command);
}
}

// the function to see if all the set needles match the element
function check_for_needles($haystack)
{
  // check for needle1, but only if it's set
  if (isset($GLOBALS['needle1']))
  {
    $results[] = (stristr($haystack, $GLOBALS['needle1']) !== FALSE) ? TRUE : FALSE;
  }

  // check for needle2, but only if it's set
  if (isset($GLOBALS['needle2']))
  {
    $results[] = (stristr($haystack, $GLOBALS['needle2']) !== FALSE) ? TRUE : FALSE;
  }

  // check for needle3, but only if it's set
  if (isset($GLOBALS['needle3']))
  {
    $results[] = (stristr($haystack, $GLOBALS['needle3']) !== FALSE) ? TRUE : FALSE;
  }

  // check if all set needles matched - return FALSE if they didn't, the element if they did
  if (in_array(FALSE, $results))
    return FALSE;
  else
    return $element;
}

// run the function on your array haystack (can be multi-leveled)
$matches = deep_map('check_for_needles', $haystack_array);

// strip out the un-matching ones
$matches = array_diff($matches, array(FALSE));

// $matches will now contain all the matching elements, keys intact, at all levels
?>[/code]

may not be anymore elegant, but it might be easier to debug.  the deep_map() is a function i made based on php.net entries, and will map the results of any defined function run on every element to an array with the same structure as the original (key => returned value for that element).  keep this one handy, it helps in a LOT of applications.

hth.

[b]EDIT:  just noticed you said they should be able to select whether ANY or ALL are matched.  this can be adjusted by changing the if() condition in check_for_needles() to suit: checking whether there are any FALSEs vs. three FALSEs.[/b]