Modify WP Query post__in before have_posts()

Question

I have a plugin that is returning post IDs based on $_GET data (which searches a secondary DB table). I am attempting to hook into the standard have_posts() function called within a taxonomy page to show the results. Since the data is passed in on the actual page, I can't use the normal pre_get_posts stand-by, like so:

function finder_sort($query) {
    if ($query->is_main_query() && is_tax(array('location_type')) && isset($_GET['finder_action'])) {
        $query->set('post__in', array(1,2,3,4,5));
    }
}
add_action('pre_get_posts','finder_sort');

Because the data returned is not available until after the $_GET requests have been read on the page. So - I was wondering if there's anything I can filter in right before the data is output. For example, here is what I have:

$locations_found = wp_list_pluck($locations, 'id');
while(have_posts()) {
    the_post();
    echo '<tr>';
    finder_location_teaser_table();
    echo '</tr>';
}

$locations_found has a list of IDs - so what I'm looking to do is filter have_posts to only include the IDs listed in $locations_found.

In addition, I need these posts ordered as they are noted in post__in. Since WP 3.5 is going to "fix" this, I've already added in the Sort Query by Post In function fix:

if ( version_compare( floatval( get_bloginfo( 'version' ) ), '3.5', '<' ) ) {

    add_filter( 'posts_orderby', 'sort_query_by_post_in', 10, 2 );
    function sort_query_by_post_in( $sortby, $thequery ) {
        if ( !empty($thequery->query['post__in']) && isset($thequery->query['orderby']) && $thequery->query['orderby'] == 'post__in' )
            $sortby = "find_in_set(ID, '" . implode( ',', $thequery->query['post__in'] ) . "')";
        return $sortby;
    }
}

so - it'd be a mixture of identifying the post__in and then specifying the order_by param in this magical filter. Any thoughts would be greatly appreciated. Thanks!

Clarification

The reason I cannot use the meta key (and why I need to return them in the same order they are retrieved by) is because there is a calculation done on each data item that is dependent on another $_GET variable (distance based on zip code).

Addition Info

Added display function that generates query vars

function finder_top_filter($term) {

    ob_start(); ?>

    <section id="sort">
        <form name="search" action="" method="get">
            <label for="finder_zipcode"><?php _e('Zip Code', 'finder'); ?></label>
            <input type="text" name="finder_zipcode" id="finder_zipcode" value="<?php echo isset($_GET['finder_zipcode']) ? $_GET['finder_zipcode'] : ''; ?>" />

            <label for="finder_distance"><?php _e('Radius', 'finder'); ?></label>
            <select name="finder_distance" id="finder_distance">
                <?php
                $finder_dist_options = apply_filters('finder_dist_options', array(
                    '5'   => '5',
                    '10'  => '10',
                    '25'  => '25',
                    '50'  => '50',
                    '100' => '100',
                    '200' => '200'
                ));
                foreach ($finder_dist_options as $key=>$value) {
                    echo '<option value="' . $key . '" ' . (isset($_GET['finder_distance']) ? selected( $_GET['finder_distance'], $key ) : '') . '>' . $value . '</option>';
                }
            ?>
            </select>

            <input type="hidden" name="finder_action" value="search" />
            <input type="submit" value="<?php _e('Show/Filter', 'finder'); ?>" />
            <span class="reset"><a href="<?php echo get_term_link($term); ?>">Reset</a></span>
        </form>
    </section>
    <?php
    echo ob_get_clean();
}

Update

Based on the recommendation of Stephen Harris, I decided to go the route of a JOIN query to alter the data passed to the loop. So, what I've tried is:

function finder_join_locations($join) {
    global $wpdb;

    return '';
}
add_filter('posts_join', 'finder_join_locations');

$i = 0;
while(have_posts()) {
    the_post();
    echo '<tr>';
    finder_location_teaser_table(get_the_ID(), $locations[$i]['distance']);
    echo '</tr>';

    $i++;
}

remove_filter('posts_join', 'finder_join_locations');

To see if I can get nothing to show, but that filter is not overriding the have_posts() function - any recommendations? What I'd like to do is pass-in via the post__in param, so after I get past this initial bump, it's figuring out the correct SQL statement to include my data (as $locations_found includes all the post data I need):

$locations_found = wp_list_pluck($locations, 'id');

Answer 1

I can see several possibilities for this one.

First, I am not convinced that that you can't use pre_get_posts though maybe you'd have to reorganize your code. You don't have to wait for your theme template to use $_GET. Try:

function tst() {
    // global $_GET; // sometimes necessary
    var_dump($_GET);
}
add_action('pre_get_posts','tst');

So I think something like:

function finder_sort($query) {
    // I don't know what your complete code for processing $_GET is but
    // that would go here and set $pin
    if ($query->is_main_query() && is_tax(array('location_type')) && isset($_GET['finder_action'])) {
        $query->set('post__in', $pin);
    }
}
add_action('pre_get_posts','finder_sort');

That is one possibility, and from where I sit right now looks like the best one. If for some reason that isn't possible you can always do something like:

$locations_found = wp_list_pluck($locations, 'id');
while(have_posts()) {
    the_post();
    if (in_array($post->ID,$locations)) {
        echo '<tr>';
        finder_location_teaser_table();
        echo '</tr>';
    }
}

I am assuming some things about your data. Your question doesn't have all of the necessary information for me to write code with 100% confidence but that is the idea.

Answer 2

Don't rely on $_GET

I would recommend, instead of relying on $_GET data, to add your custom query parameters to the $query object, using add_query_arg(), and then retrieve that value using get_query_var().

Say, for example, you've got a query variable 'foo', with a value of 'bar', such that:

`$_GET['foo'] = 'bar'`

You could make that key/value pair available to the $query object, like so:

add_query_arg( 'foo', 'bar' );

Do that wherever it is that you're currently appending &foo=bar to the URI.

Now, you can retrieve the value of 'foo' like so:

$foo = get_query_var( 'foo' );

Using pre_get_posts

Since you now have a way to retrieve your URI query value independent of $_GET, you can use that method inside of a pre_get_posts callback:

function finder_sort( $query ) {
    if ( $is_main_query() && is_tax( array( 'location_type' ) ) && isset( get_query_var( 'finder_action' ) ) ) {
        $query->set( 'post__in', get_query_var( 'finder_action' ) );
    }
}
add_action( 'pre_get_posts','finder_sort' );