[tika] 02/02: update comparison reports for extract exception comparisons

[ta...@apache.org]

This is an automated email from the ASF dual-hosted git repository.

tallison pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/tika.git

commit a63235476b938210e5111611f502aad331de904b
Author: tallison <ta...@apache.org>
AuthorDate: Sat Apr 10 07:08:47 2021 -0400

    update comparison reports for extract exception comparisons
---
 .../src/main/resources/comparison-reports-pg.xml   |  2 ++
 .../src/main/resources/comparison-reports.xml      | 29 ++++++++++++++++++++++
 2 files changed, 31 insertions(+)

diff --git a/tika-eval/tika-eval-app/src/main/resources/comparison-reports-pg.xml b/tika-eval/tika-eval-app/src/main/resources/comparison-reports-pg.xml
index 5bcf88e..5940f46 100644
--- a/tika-eval/tika-eval-app/src/main/resources/comparison-reports-pg.xml
+++ b/tika-eval/tika-eval-app/src/main/resources/comparison-reports-pg.xml
@@ -19,6 +19,8 @@
   under the License.
 -->
 
+<!-- this should be the same as comparison-reports.xml but
+     translated from the H2 dialect to postgres -->
 <reports>
 
 
diff --git a/tika-eval/tika-eval-app/src/main/resources/comparison-reports.xml b/tika-eval/tika-eval-app/src/main/resources/comparison-reports.xml
index e23ec5e..a12e7c7 100644
--- a/tika-eval/tika-eval-app/src/main/resources/comparison-reports.xml
+++ b/tika-eval/tika-eval-app/src/main/resources/comparison-reports.xml
@@ -1202,6 +1202,35 @@
             on e.extract_exception_id=t.extract_exception_id
         </sql>
     </report>
+    <report reportName="fixedCatastrophicExtractExceptions"
+            reportFilename="exceptions/fixed_catastrophic_exceptions_in_b.xlsx"
+            format="xlsx"
+            includeSql="true">
+        <sql>
+            select exa.file_path, ra.extract_exception_description, rb.extract_exception_description
+            from extract_exceptions_a exa
+            left join extract_exceptions_b exb on exa.container_id=exb.container_id
+            join ref_extract_exception_types ra on exa.extract_exception_id = ra.extract_exception_id
+            left join ref_extract_exception_types rb on exb.extract_exception_id = rb.extract_exception_id
+            where exa.extract_exception_id &lt; 4
+            and (exb.extract_exception_id is null or exb.extract_exception_id &gt; 3)
+        </sql>
+    </report>
+    <report reportName="newCatastrophicExtractExceptions"
+            reportFilename="exceptions/new_catastrophic_exceptions_in_b.xlsx"
+            format="xlsx"
+            includeSql="true">
+        <sql>
+            select exb.file_path, rb.extract_exception_description, ra.extract_exception_description
+            from extract_exceptions_b exb
+            left join extract_exceptions_a exa on exb.container_id=exa.container_id
+            join ref_extract_exception_types rb on exb.extract_exception_id = rb.extract_exception_id
+            left join ref_extract_exception_types ra on exa.extract_exception_id = ra.extract_exception_id
+            where exb.extract_exception_id &lt; 4
+            and (exa.extract_exception_id is null or exa.extract_exception_id &gt; 3)
+        </sql>
+    </report>
+
     <report reportName="parseExceptionTypesA"
             reportFilename="exceptions/overall_exception_types_a.xlsx"
             format="xlsx"