Filtering with NA for genomic-data-bin-counts

src/main/resources/org/cbioportal/persistence/mybatisclickhouse/StudyViewFilterMapper.xml

@@ -141,13 +141,10 @@
  <foreach item="genomicDataFilter" collection="studyViewFilterHelper.studyViewFilter.genomicDataFilters" open="INTERSECT" separator="INTERSECT">
  <choose>
  <when test="genomicDataFilter.profileType == 'cna' or genomicDataFilter.profileType == 'gistic'">
- <include refid="applyGenomicDataFilter"/>
+ <include refid="categoricalGenomicDataFilter"/>
  </when>
  <otherwise>
- <include refid="numericalGenomicDataCountFilter">
- <property name="unique_id" value="sample_unique_id"/>
- <property name="table_name" value="genetic_alteration_derived"/>
- </include>
+ <include refid="numericalGenomicDataFilter"/>
  </otherwise>
  </choose>
  </foreach>
@@ -387,19 +384,27 @@
  </foreach>
  </sql>
 
- <sql id="numericalGenomicDataCountFilter">
- SELECT ${unique_id}
- FROM ${table_name}
- WHERE hugo_gene_symbol = '${genomicDataFilter.hugoGeneSymbol}' AND
- profile_type='${genomicDataFilter.profileType}'
- <foreach item="dataFilterValue" collection="genomicDataFilter.values" open=" AND ((" separator=") OR (" close="))">
+ <sql id="numericalGenomicDataFilter">
+ <!-- check if 'NA' is selected -->
+ <bind name="containsNA" value="false" />
+ <foreach item="dataFilterValue" collection="genomicDataFilter.values">
+ <if test="dataFilterValue.value == 'NA'">
+ <bind name="containsNA" value="true" />
+ </if>
+ </foreach>
+ <if test="containsNA">
+ SELECT DISTINCT sd.sample_unique_id
+ FROM sample_derived sd
+ LEFT JOIN (<include refid="selectAllGenomicNumericalSamples"/>) AS genomic_numerical_query ON sd.sample_unique_id = genomic_numerical_query.sample_unique_id
+ WHERE alteration_value IS null
+ UNION ALL
+ </if>
+ SELECT DISTINCT sd.sample_unique_id
+ FROM sample_derived sd
+ INNER JOIN (<include refid="selectAllGenomicNumericalSamples"/>) AS genomic_numerical_query ON sd.sample_unique_id = genomic_numerical_query.sample_unique_id
+ WHERE
+ <foreach item="dataFilterValue" collection="genomicDataFilter.values" open="((" separator=") OR (" close="))">
  <trim prefix="" prefixOverrides="AND">
- <if test="dataFilterValue.value eq 'NA'">
- AND
- <include refid="isAttributeValueNA">
- <property name="attribute_value" value="alteration_value"/>
- </include>
- </if>
  <if test="dataFilterValue.start != null and dataFilterValue.end == null">
  AND match(alteration_value, '^>?=?[-+]?[0-9]*[.,]?[0-9]+$')
  </if>
@@ -440,6 +445,17 @@
  </trim>
  </foreach>
  </sql>
+
+ <sql id="selectAllGenomicNumericalSamples">
+ SELECT sample_unique_id, alteration_value
+ FROM genetic_alteration_derived comeonwhyisitnotworking
+ WHERE profile_type = #{genomicDataFilter.profileType}
+ AND hugo_gene_symbol = #{genomicDataFilter.hugoGeneSymbol}
+ AND cancer_study_identifier IN
+ <foreach item="studyId" collection="studyViewFilterHelper.studyViewFilter.studyIds" open="(" separator="," close=")">
+ #{studyId}
+ </foreach>
+ </sql>
 
  <!-- TODO: update the database scheme to include the data_type column -->
  <sql id="numericalGenericAssayDataCountFilter">
@@ -577,7 +593,7 @@
  </if>
  </sql>
 
- <sql id="applyGenomicDataFilter">
+ <sql id="categoricalGenomicDataFilter">
  <!-- filter on study to reduce query size in preparation of the following LEFT JOIN -->
  WITH cna_query AS (
  SELECT sample_unique_id, alteration_value
@@ -597,9 +613,11 @@
  <foreach item="studyId" collection="studyViewFilterHelper.studyViewFilter.studyIds" open="(" separator="," close=")">
  #{studyId}
  </foreach>
- AND alteration_value IN
- <foreach item="dataFilterValue" collection="genomicDataFilter.values" open="(" separator="," close=")">
- #{dataFilterValue.value}
+ <foreach item="dataFilterValue" collection="genomicDataFilter.values" open="AND (" separator=" OR " close=")">
+ <choose>
+ <when test="dataFilterValue.value == 'NA'">alteration_value IS null</when>
+ <otherwise>alteration_value == #{dataFilterValue.value}</otherwise>
+ </choose>
  </foreach>
  </sql>
 </mapper>