we cascade the LEFT JOIN-ed queries to ensure correct cardinality of records...

we cascade the LEFT JOIN-ed queries to ensure correct cardinality of records (probably not necessary when joining affiliations + keywords because affiliations are 1 to 1, but definitely needed when joining affs + kws + hashtags)

print_directory.php

@@ -268,33 +268,41 @@ else {
     $filter = "";
 }
 $sql = <<< END_QUERY
-SELECT * FROM
-    (SELECT
-        scholars.*,
-        affiliations.*,
-        COUNT(keywords.kwid) AS keywords_nb,
-        GROUP_CONCAT(keywords.kwid) AS keywords_ids,
-        GROUP_CONCAT(kwstr) AS keywords_list,
-        GROUP_CONCAT(hashtags.htid) AS hashtags_ids,
+SELECT * FROM (
+    SELECT
+        scholars_affiliations_and_keywords.*,
         GROUP_CONCAT(htstr) AS hashtags_list
-    FROM scholars
-    LEFT JOIN sch_kw
-        ON sch_kw.uid = luid
-    LEFT JOIN keywords
-        ON sch_kw.kwid = keywords.kwid
+    FROM (
+        SELECT
+            scholars_and_affiliations.*,
+            GROUP_CONCAT(kwstr) AS keywords_list
+        FROM (
+            SELECT
+                scholars.*,
+                affiliations.*
+            FROM scholars
+            LEFT JOIN affiliations
+                ON scholars.affiliation_id = affiliations.affid
+            WHERE (record_status = 'active'
+                OR (record_status = 'legacy' AND valid_date >= NOW()))
+        ) AS scholars_and_affiliations
+
+        LEFT JOIN sch_kw
+            ON sch_kw.uid = scholars_and_affiliations.luid
+        LEFT JOIN keywords
+            ON sch_kw.kwid = keywords.kwid
+        GROUP BY luid
+
+    ) AS scholars_affiliations_and_keywords
     LEFT JOIN sch_ht
         ON sch_ht.uid = luid
     LEFT JOIN hashtags
         ON sch_ht.htid = hashtags.htid
-    LEFT JOIN affiliations
-        ON affiliation_id = affid
-    WHERE (record_status = 'active'
-            OR (record_status = 'legacy' AND valid_date >= NOW()))
-    GROUP BY luid) AS full_scholars_info
+    GROUP BY luid
+) AS full_scholars_info
     {$filter}
 END_QUERY;
 
-
 // debug
 // echo '<p style="color:white">query:'. $sql ."<p>";
 

print_scholar_directory.php

@@ -115,21 +115,25 @@ foreach ($scholar_id_array as $scholar_id){
 // £TODO do it at once with previous SELECT !!
 $sql = <<< END_QUERY
 SELECT
-    scholars.*,
-    affiliations.*,
+    scholars_and_affiliations.*,
     COUNT(keywords.kwid) AS keywords_nb,
     GROUP_CONCAT(keywords.kwid) AS keywords_ids,
     GROUP_CONCAT(kwstr) AS keywords_list
-FROM scholars
-JOIN sch_kw
-    ON luid = uid
-JOIN keywords
+FROM (
+    SELECT
+        scholars.*,
+        affiliations.*
+    FROM scholars
+    LEFT JOIN affiliations
+        ON scholars.affiliation_id = affiliations.affid
+    WHERE (record_status = 'active'
+        OR (record_status = 'legacy' AND valid_date >= NOW()))
+) AS scholars_and_affiliations
+
+LEFT JOIN sch_kw
+    ON sch_kw.uid = scholars_and_affiliations.luid
+LEFT JOIN keywords
     ON sch_kw.kwid = keywords.kwid
-LEFT JOIN affiliations
-    ON affiliation_id = affid
-WHERE luid = "{$scholar_id}"
-AND (record_status = 'active'
-     OR (record_status = 'legacy' AND valid_date >= NOW()))
 GROUP BY luid
 END_QUERY;
 

search_filter.php

 <?php
 
 /*
- * Génère le json des scholars à partir de la base mysql
+ * Génère les agrégations pour les filtres à partir de la base mysql
+ * £TODO remplacer les appels à search_filter par des appels à api/aggs
  */
 include("php_library/parametres.php");
 include("php_library/normalize.php");
@@ -52,39 +53,51 @@ function filter_word($value) {
 // $req = "SELECT ".$cat." AS clef, count(".$cat.") AS value FROM scholars WHERE ".$cat." ".$query." GROUP BY ".$cat." ORDER BY value DESC";
 
 // TODO differentiate req's target cols earlier: above in "if ($category == X)"
-$req = <<<END_QUERY
+
+$req = <<<END_SQL
+SELECT
+    {$cat} AS clef,
+    count({$cat}) AS value
+FROM (
     SELECT
-        {$cat} AS clef,
-        count({$cat}) AS value
+        scholars_affiliations_and_keywords.*,
+        GROUP_CONCAT(htstr) AS hashtags_list
     FROM (
         SELECT
-            -- we create all needed cats for the outer select
-            -- ==============================================
-            scholars.luid,
-            scholars.country,
-            affiliations.org,
-            affiliations.team_lab,
-            GROUP_CONCAT(kwstr) AS keywords_list,
-            GROUP_CONCAT(htstr) AS hashtags_list
-        FROM scholars
+            scholars_and_affiliations.*,
+            GROUP_CONCAT(kwstr) AS keywords_list
+        FROM (
+            SELECT
+                scholars.luid,
+                scholars.country,
+                affiliations.org,
+                affiliations.team_lab
+            FROM scholars
+            LEFT JOIN affiliations
+                ON scholars.affiliation_id = affiliations.affid
+            WHERE (record_status = 'active'
+                OR (record_status = 'legacy' AND valid_date >= NOW()))
+        ) AS scholars_and_affiliations
+
         LEFT JOIN sch_kw
-            ON sch_kw.uid = luid
+            ON sch_kw.uid = scholars_and_affiliations.luid
         LEFT JOIN keywords
             ON sch_kw.kwid = keywords.kwid
-        LEFT JOIN sch_ht
-            ON sch_ht.uid = luid
-        LEFT JOIN hashtags
-            ON sch_ht.htid = hashtags.htid
-        LEFT JOIN affiliations
-            ON scholars.affiliation_id = affiliations.affid
-        WHERE (record_status = 'active'
-                OR (record_status = 'legacy' AND valid_date >= NOW()))
         GROUP BY luid
-        ) AS full_scholars_info
-    WHERE {$cat} {$query}                          -- <== our filter
-    GROUP BY $cat
-    ORDER BY value DESC ;
-END_QUERY;
+
+    ) AS scholars_affiliations_and_keywords
+    LEFT JOIN sch_ht
+        ON sch_ht.uid = luid
+    LEFT JOIN hashtags
+        ON sch_ht.htid = hashtags.htid
+    GROUP BY luid
+) AS scholars_info
+
+WHERE {$cat} {$query}                          -- <== our filter
+GROUP BY $cat
+ORDER BY value DESC ;
+
+END_SQL;
 
 // echo $req;
 $results = array();

services/db_to_tina_api/extractDataCustom.py

@@ -295,21 +295,28 @@ class MyExtractor:
 
         for scholar_id in scholar_array:
             sql3='''
+                SELECT
+                    scholars_and_affiliations.*,
+                    COUNT(keywords.kwid) AS keywords_nb,
+                    GROUP_CONCAT(keywords.kwid) AS keywords_ids,
+                    GROUP_CONCAT(kwstr) AS keywords_list
+                FROM (
                     SELECT
                         scholars.*,
-                        affiliations.*,
-                        COUNT(keywords.kwid) AS keywords_nb,
-                        GROUP_CONCAT(keywords.kwid) AS keywords_ids,
-                        GROUP_CONCAT(kwstr) AS keywords_list
+                        affiliations.*
                     FROM scholars
-                    LEFT JOIN sch_kw
-                        ON uid = luid
-                    LEFT JOIN keywords
-                        ON sch_kw.kwid = keywords.kwid
                     LEFT JOIN affiliations
-                        ON affiliation_id = affid
-                    WHERE luid = "%s"
-                    GROUP BY luid ;
+                        ON scholars.affiliation_id = affiliations.affid
+                    WHERE (record_status = 'active'
+                        OR (record_status = 'legacy' AND valid_date >= NOW()))
+                ) AS scholars_and_affiliations
+
+                LEFT JOIN sch_kw
+                    ON sch_kw.uid = scholars_and_affiliations.luid
+                LEFT JOIN keywords
+                    ON sch_kw.kwid = keywords.kwid
+                WHERE luid = %s
+                GROUP BY luid ;
             ''' % scholar_id
 
             # debug