scholar directory sql query 2-in-1 (neighboors expansion and scholars info at once together)

print_scholar_directory.php

@@ -68,58 +68,48 @@ $base = new PDO($dsn, $user, $pass, $opt);
 
 
 // liste des chercheurs
-$scholar_array = array();
+$scholars = array();
 
 if ($userid) {
-    if (sizeof($userid) > 0) {
-
-        // old way in two steps without a scholars <=> keywords table
-        // $sql1 = "SELECT keywords,last_name,first_name FROM scholars WHERE luid='" . $login . "'";
-        // $sql2 = "SELECT uid FROM sch_kw JOIN keywords ON sch_kw.kwid = keywords.kwid WHERE kwstr LIKE \"%" . trim($keyword)."%\"";
-
-        // new way in one query
-        $sql1 = <<< HERE_QUERY
-        SELECT second_level.uid
-        FROM      sch_kw
-        LEFT JOIN sch_kw
-            AS second_level
-            ON sch_kw.kwid = second_level.kwid
-        WHERE sch_kw.uid = {$userid}
-            -- (uncomment if ego not self-neighboor)
-            -- AND second_level.uid != sch_kw.uid
-        GROUP BY second_level.uid
-        ORDER BY second_level.uid != {$userid} ;  -- allows ego to be first
-HERE_QUERY;
-
-        foreach ($base->query($sql1) as $row) {
-            if (array_key_exists($row['uid'], $scholar_array)){
-                $scholar_array[$row['uid']] += 1;
-            }else{
-                $scholar_array[$row['uid']] = 1;
-            }
-        }
-    }
-}
-
-// les scholars sont affichés par ordre de pertinence
-arsort($scholar_array);
-
-$scholar_id_array=array_keys($scholar_array);
-// var_dump($scholar_id_array)."<br/>" ;
 
-// liste des chercheurs "expansion transitive"
-$scholars = array();
-//
-foreach ($scholar_id_array as $scholar_id){
-
-// £TODO do it at once with previous SELECT !!
-$sql = <<< END_QUERY
-SELECT
+    // query idea:
+    // ------------
+    // SELECT sch_kw.uid AS source,
+    //        count(sch_kw.kwid) AS similarity,
+    //        second_level.uid AS neighboor
+    // FROM      sch_kw
+    // LEFT JOIN sch_kw AS second_level ON sch_kw.kwid = second_level.kwid
+    // WHERE sch_kw.uid = 4207
+    // GROUP BY second_level.uid
+    // ORDER BY count(sch_kw.kwid) DESC, second_level.uid != 4207 ASC;
+
+
+    // exemple:
+    // +--------+------------+-----------+
+    // | source | similarity | neighboor |
+    // +--------+------------+-----------+
+    // |   4207 |          2 |      4207 |
+    // |   4207 |          2 |      2792 |
+    // |   4207 |          1 |      2732 |
+    // |   4207 |          1 |      2569 |
+    // |   4207 |          1 |      3128 |
+    // |   4207 |          1 |      2636 |
+    // |   4207 |          1 |      3488 |
+    // |   4207 |          1 |      2727 |
+    // |   4207 |          1 |      3604 |
+    // |   4207 |          1 |      3942 |
+    // |            (...)                |
+    // +--------+------------+-----------+
+
+
+    // implementation with all details and infos to retrieve
+    $sql = <<< HERE_QUERY
+    SELECT
         scholars_and_affiliations.*,
         COUNT(keywords.kwid) AS keywords_nb,
         GROUP_CONCAT(keywords.kwid) AS keywords_ids,
         GROUP_CONCAT(kwstr) AS keywords_list
-FROM (
+    FROM (
         SELECT
             scholars.*,
             affiliations.*
@@ -128,20 +118,22 @@ FROM (
             ON scholars.affiliation_id = affiliations.affid
         WHERE (record_status = 'active'
             OR (record_status = 'legacy' AND valid_date >= NOW()))
-) AS scholars_and_affiliations
+    ) AS scholars_and_affiliations
 
-LEFT JOIN sch_kw
-    ON sch_kw.uid = scholars_and_affiliations.luid
-LEFT JOIN keywords
+    LEFT JOIN sch_kw AS second_level
+        ON second_level.uid = scholars_and_affiliations.luid
+    JOIN sch_kw ON sch_kw.kwid = second_level.kwid
+    JOIN keywords
         ON sch_kw.kwid = keywords.kwid
-GROUP BY luid
-END_QUERY;
 
+    WHERE sch_kw.uid = {$userid}
+    GROUP BY luid
 
+    ORDER BY count(sch_kw.kwid) DESC, second_level.uid != {$userid} ASC;
 
-// echo var_dump($scholar_id)."<br/>" ;
-//$query = "SELECT * FROM scholars";
-foreach ($base->query($sql) as $row) {
+HERE_QUERY;
+
+    foreach ($base->query($sql) as $row) {
         $info = array();
         $info['unique_id'] = $row['luid'];
         $info['doors_uid'] = $row['doors_uid'];
@@ -186,9 +178,10 @@ foreach ($base->query($sql) as $row) {
         // $info['fax'] = $row['fax'];
         // $info['affiliation_acronym'] = $row['affiliation_acronym'];
         $scholars[$row['luid']] = $info;
+    }
 }
 
-}
+
 /// stats
 include ("php_library/stat-prep_from_array.php");