commit 807138d085106ba0b97489a55fe9a4c57854f032
author Charles S. Givre <cgivre@apache.org> Wed Jan 25 08:54:29 2023 -0500
committer GitHub <noreply@github.com> Wed Jan 25 08:54:29 2023 -0500
tree 8ec92364e1afd2a93c81708229180ac03acc36ce
parent 5f861d1c1e9a8a93cd2c208475a39572d5984189

DRILL-8392: Empty Tables Causes Index Out of Bounds Exception on PDF Reader (#2744)

contrib/format-pdf/src/main/java/org/apache/drill/exec/store/pdf/PdfBatchReader.java

diff --git a/contrib/format-pdf/src/main/java/org/apache/drill/exec/store/pdf/PdfBatchReader.java b/contrib/format-pdf/src/main/java/org/apache/drill/exec/store/pdf/PdfBatchReader.java
index 26d3a94..05dd3b4 100644
--- a/contrib/format-pdf/src/main/java/org/apache/drill/exec/store/pdf/PdfBatchReader.java
+++ b/contrib/format-pdf/src/main/java/org/apache/drill/exec/store/pdf/PdfBatchReader.java
@@ -218,8 +218,16 @@
 
       if (!Strings.isNullOrEmpty(value)) {
         writers.get(rowPosition).load(row.get(rowPosition));
+
+        // If there is not a provided schema, advance the row position index only when values are found
+        if (negotiator.providedSchema() == null) {
+          rowPosition++;
+        }
       }
-      rowPosition++;
+      // Advance the row position index when there is a provided schema.
+      if (negotiator.providedSchema() != null) {
+        rowPosition++;
+      }
     }
 
     metadataReader.writeMetadata();