wp_insert_post() creates duplicates with 'post_status' => 'publish'

Question

I am trying to add posts of a custom post-type in bulk from a .csv file. Everything works except that it adds the post twice. They are exactly the same in the database except one has author 1 and the other has author 2. Also it appends a '-2' to post_name. Here is what I have:

add_action('wp_loaded', 'add_from_csv', 10);
function add_from_csv() {
    if (($handle = fopen(dirname(__FILE__) . "/posts.csv", "r")) !== FALSE) {
        while (($data = fgetcsv($handle, 1000, ",")) !== FALSE) {
            $term_1 = $data[0];
            $term_2 = $data[1];
            $term_3 = $data[2];
            $term_4 = $data[3];
            $term_5 = $data[4];
            $meta_1 = $data[5];
            $meta_2 = $data[6];
            $meta_3 = $data[7];
            $meta_4 = $data[8];
            $meta_5 = $data[9];
            $meta_6 = isset($data[10]) ? $data[10] : '';

            $post_id = wp_insert_post( 
                array(
                    'post_title' => $meta_1 . ' ' . $meta_2 . ' ' . $meta_3 . ': ' . $meta_4,
                    'post_type' => 'custom_type',
                    'post_status' => 'publish',
                )
            );

            if(!is_wp_error($post_id) || $post_id != 0) {
                $tax_1_terms = array($term_1, $term_2, $term_3, $term_4);
                wp_set_object_terms( $post_id, $tax_1_terms, 'tax_1' );
                wp_set_object_terms( $post_id, $term_5, 'tax_2' );

                update_post_meta($post_id, 'meta-1', $meta_1);
                update_post_meta($post_id, 'meta-2', $meta_2);
                update_post_meta($post_id, 'meta-3', $meta_3);
                update_post_meta($post_id, 'meta-4', $meta_4);
                update_post_meta($post_id, 'meta-5', $meta_5);
                update_post_meta($post_id, 'meta-6', $meta_6);
            } else {
                echo 'There was a WP_Error when loading posts from csv!';
                break;
            }
        }
        fclose($handle);
    }
}

I have this in a plugin by itself. If I don't set 'post_status', and it uses the default status of 'draft' then there are no duplicates. but when I set 'post_status' => 'publish' then it inserts the post twice. Both of the duplicates have status of publish. Can someone help me resolve this please? Thanks

Answer 1

I took your code an minimized it, and made a .csv with 3 rows, and was able to get the same error as you.

However, I got three instances of the same post. No coincidence that I have 3 rows in my csv. Your insert code is right, so it has to be the hook being re-called each time the insert is fired.

I changed the hook to something seemingly random, but very specific for being user-end and admin-only, the admin_notices hook, and it worked, no duplicate/triplicate entries made for the one page-load I enabled the code.

Although I don't have documentation to prove this, I believe that many of the Wordpress hooks re-fire with the wp_insert_post call, and thats your problem.

 // causes duplicates ---
//add_action('wp_loaded', 'add_from_csv', 10);

// no duplicates ---
//add_action('admin_notices', 'add_from_csv', 10);

function add_from_csv() {
    if (($handle = fopen(dirname(__FILE__) . "/posts.csv", "r")) !== FALSE) {
        while (($data = fgetcsv($handle, 1000, ",")) !== FALSE) {
            $post_id = wp_insert_post( 
                array(
                    'post_title' => $data[0],
                    'post_type' => 'post',
                    'post_status' => 'publish',
                )
            );
            if(!is_wp_error($post_id) || $post_id != 0)
                echo "<pre>Added {$data[0]}</pre>";

        }
        fclose($handle);
    }
}

I would recommend creating a very simple admin settings/options page instead to house your code, and make the code fire from a form submit button, verified with nonce.