From af84d12e32041b52934f13ab2e925347246c8139 Mon Sep 17 00:00:00 2001
From: Onur Sumer <s.onur.sumer@gmail.com>
Date: Wed, 11 Dec 2024 14:42:34 -0500
Subject: [PATCH] make sure study ids exist before using them in filter SQL

---
 .../StudyViewFilterMapper.xml                 | 68 ++++++++++++-------
 .../mybatisclickhouse/StudyViewMapper.xml     | 13 ++--
 2 files changed, 53 insertions(+), 28 deletions(-)

diff --git a/src/main/resources/org/cbioportal/persistence/mybatisclickhouse/StudyViewFilterMapper.xml b/src/main/resources/org/cbioportal/persistence/mybatisclickhouse/StudyViewFilterMapper.xml
index 30f3c7557b4..382ecd5a70d 100644
--- a/src/main/resources/org/cbioportal/persistence/mybatisclickhouse/StudyViewFilterMapper.xml
+++ b/src/main/resources/org/cbioportal/persistence/mybatisclickhouse/StudyViewFilterMapper.xml
@@ -42,18 +42,20 @@
                     JOIN genetic_profile gp ON sample_profile.genetic_profile_id = gp.genetic_profile_id
                     JOIN cancer_study cs ON gp.cancer_study_id = cs.cancer_study_id
                     JOIN sample_derived on sample_profile.sample_id = sample_derived.internal_id
-                    <where>
-                        sample_derived.cancer_study_identifier IN
-                        <foreach item="studyId" collection="studyViewFilterHelper.studyViewFilter.studyIds" open="(" separator="," close=")">
-                            #{studyId}
-                        </foreach>
-                        AND
-                        <foreach item="studyId" collection="studyViewFilterHelper.studyViewFilter.studyIds" open="(" close=")" separator="OR">
-                            <foreach item="genomicProfileId" collection="ANDGroup" open="(" separator=" OR " close=")">
-                                gp.stable_id=concat(#{studyId}, '_', #{genomicProfileId})
+                    <if test="studyViewFilterHelper.studyViewFilter.studyIds != null and !studyViewFilterHelper.studyViewFilter.studyIds.isEmpty()">
+                        <where>
+                            sample_derived.cancer_study_identifier IN
+                            <foreach item="studyId" collection="studyViewFilterHelper.studyViewFilter.studyIds" open="(" separator="," close=")">
+                                #{studyId}
                             </foreach>
-                        </foreach>
-                    </where>
+                            AND
+                            <foreach item="studyId" collection="studyViewFilterHelper.studyViewFilter.studyIds" open="(" close=")" separator="OR">
+                                <foreach item="genomicProfileId" collection="ANDGroup" open="(" separator=" OR " close=")">
+                                    gp.stable_id=concat(#{studyId}, '_', #{genomicProfileId})
+                                </foreach>
+                            </foreach>
+                        </where>
+                    </if>
                 </foreach>
                 )
 
@@ -430,10 +432,12 @@
         SELECT sample_unique_id, patient_unique_id, attribute_value
         FROM clinical_data_derived
         WHERE attribute_name = #{clinicalDataFilter.attributeId} AND type='${type}'
-        AND cancer_study_identifier IN
-        <foreach item="studyId" collection="studyViewFilterHelper.studyViewFilter.studyIds" open="(" separator="," close=")">
-            #{studyId}
-        </foreach>
+        <if test="studyViewFilterHelper.studyViewFilter.studyIds != null and !studyViewFilterHelper.studyViewFilter.studyIds.isEmpty()">
+            AND cancer_study_identifier IN
+            <foreach item="studyId" collection="studyViewFilterHelper.studyViewFilter.studyIds" open="(" separator="," close=")">
+                #{studyId}
+            </foreach>
+        </if>
     </sql>
     
     <sql id="categoricalClinicalDataCountFilter">
@@ -511,10 +515,12 @@
         FROM genetic_alteration_derived
         WHERE profile_type = #{genomicDataFilter.profileType}
             AND hugo_gene_symbol = #{genomicDataFilter.hugoGeneSymbol}
+        <if test="studyViewFilterHelper.studyViewFilter.studyIds != null and !studyViewFilterHelper.studyViewFilter.studyIds.isEmpty()">
             AND cancer_study_identifier IN
             <foreach item="studyId" collection="studyViewFilterHelper.studyViewFilter.studyIds" open="(" separator="," close=")">
                 #{studyId}
             </foreach>
+        </if>
     </sql>
 
     <sql id="applyGenericAssayDataFilter">
@@ -641,30 +647,40 @@
                 SELECT sample_unique_id
                 FROM sample_derived
                 WHERE cancer_study_identifier IN
+                <if test="studyViewFilterHelper.studyViewFilter.studyIds != null and !studyViewFilterHelper.studyViewFilter.studyIds.isEmpty()">
                     <foreach item="studyId" collection="studyViewFilterHelper.studyViewFilter.studyIds" open="(" separator="," close=")">
                         #{studyId}
                     </foreach>
+                </if>
             ),
             profiled_samples AS (
                 SELECT DISTINCT sgp.sample_unique_id
                 FROM sample_to_gene_panel_derived sgp
                     JOIN gene_panel_to_gene_derived gpg ON sgp.gene_panel_id = gpg.gene_panel_id
-                WHERE cancer_study_identifier IN
+                WHERE
+                <if test="studyViewFilterHelper.studyViewFilter.studyIds != null and !studyViewFilterHelper.studyViewFilter.studyIds.isEmpty()">
+                    cancer_study_identifier IN
                     <foreach item="studyId" collection="studyViewFilterHelper.studyViewFilter.studyIds" open="(" separator="," close=")">
                         #{studyId}
                     </foreach>
-                    AND gpg.gene = #{mutationDataFilter.hugoGeneSymbol}
-                    AND sgp.alteration_type = 'MUTATION_EXTENDED'
+                  AND
+                </if>
+                    gpg.gene = #{mutationDataFilter.hugoGeneSymbol}
+                  AND sgp.alteration_type = 'MUTATION_EXTENDED'
             ),
             mutated_samples AS (
                 SELECT DISTINCT sample_unique_id
                 FROM genomic_event_derived
-                WHERE cancer_study_identifier IN
+                WHERE
+                <if test="studyViewFilterHelper.studyViewFilter.studyIds != null and !studyViewFilterHelper.studyViewFilter.studyIds.isEmpty()">
+                    cancer_study_identifier IN
                     <foreach item="studyId" collection="studyViewFilterHelper.studyViewFilter.studyIds" open="(" separator="," close=")">
                         #{studyId}
                     </foreach>
-                    AND hugo_gene_symbol = #{mutationDataFilter.hugoGeneSymbol}
-                    AND variant_type = 'mutation'
+                  AND
+                </if>
+                    hugo_gene_symbol = #{mutationDataFilter.hugoGeneSymbol}
+                  AND variant_type = 'mutation'
             )
             SELECT DISTINCT sample_unique_id
             FROM
@@ -706,20 +722,26 @@
             FROM genetic_alteration_derived
             WHERE profile_type = #{genomicDataFilter.profileType}
                 AND hugo_gene_symbol = #{genomicDataFilter.hugoGeneSymbol}
+            <if test="studyViewFilterHelper.studyViewFilter.studyIds != null and !studyViewFilterHelper.studyViewFilter.studyIds.isEmpty()">
                 AND cancer_study_identifier IN
                 <foreach item="studyId" collection="studyViewFilterHelper.studyViewFilter.studyIds" open="(" separator="," close=")">
                     #{studyId}
                 </foreach>
+            </if>
         )
         SELECT DISTINCT sd.sample_unique_id
         <!-- join with sample table to get all 'NA' samples -->
         FROM sample_derived sd
             LEFT JOIN cna_query ON sd.sample_unique_id = cna_query.sampleUniqueId
-        WHERE cancer_study_identifier IN
+        WHERE
+        <if test="studyViewFilterHelper.studyViewFilter.studyIds != null and !studyViewFilterHelper.studyViewFilter.studyIds.isEmpty()">
+            cancer_study_identifier IN
             <foreach item="studyId" collection="studyViewFilterHelper.studyViewFilter.studyIds" open="(" separator="," close=")">
                 #{studyId}
             </foreach>
-            <foreach item="dataFilterValue" collection="genomicDataFilter.values" open="AND (" separator=" OR " close=")">
+            AND
+        </if>
+            <foreach item="dataFilterValue" collection="genomicDataFilter.values" open="(" separator=" OR " close=")">
                 <choose>
                     <!-- NA value samples -->
                     <when test="dataFilterValue.value == 'NA'">alteration_value IS null</when>
diff --git a/src/main/resources/org/cbioportal/persistence/mybatisclickhouse/StudyViewMapper.xml b/src/main/resources/org/cbioportal/persistence/mybatisclickhouse/StudyViewMapper.xml
index e140fd00b6d..910d3ef9119 100644
--- a/src/main/resources/org/cbioportal/persistence/mybatisclickhouse/StudyViewMapper.xml
+++ b/src/main/resources/org/cbioportal/persistence/mybatisclickhouse/StudyViewMapper.xml
@@ -346,11 +346,14 @@
                     (
                         SELECT count() > 0
                         FROM genetic_profile
-                        WHERE patient_level = 1
-                        AND stable_id IN
-                        <foreach item="studyId" collection="studyViewFilterHelper.studyViewFilter.studyIds" open="(" separator="," close=")">
-                            concat(#{studyId}, '_', #{profileType})
-                        </foreach>
+                        WHERE
+                            patient_level = 1
+                        <if test="studyViewFilterHelper.studyViewFilter.studyIds != null and !studyViewFilterHelper.studyViewFilter.studyIds.isEmpty()">
+                            AND stable_id IN
+                            <foreach item="studyId" collection="studyViewFilterHelper.studyViewFilter.studyIds" open="(" separator="," close=")">
+                                concat(#{studyId}, '_', #{profileType})
+                            </foreach>
+                        </if>
                     ),
                     (SELECT * FROM (<include refid="getTotalPatientCount"/>)),
                     (SELECT * FROM (<include refid="getTotalSampleCount"/>))