From 68afce70e97169d30c615f6d8a8fc106f0bbdbcc Mon Sep 17 00:00:00 2001
From: anupksingh <anup.singh@cloudsufi.com>
Date: Fri, 11 Oct 2024 04:35:07 +0000
Subject: [PATCH] Oracle XML support fix.

---
 oracle-plugin/pom.xml                              | 13 ++++++++++++-
 .../cdap/plugin/oracle/OracleSourceDBRecord.java   | 14 ++++++++++++++
 .../plugin/oracle/OracleSourceSchemaReader.java    |  6 +++++-
 3 files changed, 31 insertions(+), 2 deletions(-)

diff --git a/oracle-plugin/pom.xml b/oracle-plugin/pom.xml
index b21152fcd..0b93edc88 100644
--- a/oracle-plugin/pom.xml
+++ b/oracle-plugin/pom.xml
@@ -28,6 +28,16 @@
   <modelVersion>4.0.0</modelVersion>
 
   <dependencies>
+    <dependency>
+      <groupId>com.oracle.database.xml</groupId>
+      <artifactId>xdb</artifactId>
+      <version>23.5.0.24.07</version>
+    </dependency>
+    <dependency>
+      <groupId>com.oracle.database.xml</groupId>
+      <artifactId>xmlparserv2</artifactId>
+      <version>23.5.0.24.07</version>
+    </dependency>
     <dependency>
       <groupId>io.cdap.cdap</groupId>
       <artifactId>cdap-etl-api</artifactId>
@@ -118,7 +128,8 @@
                 io.cdap.plugin.db.sink.*;
                 org.apache.commons.lang;
                 org.apache.commons.logging.*;
-                org.codehaus.jackson.*
+                org.codehaus.jackson.*;
+                oracle.*;
             </_exportcontents>
             <Embed-Dependency>*;inline=false;scope=compile</Embed-Dependency>
             <Embed-Transitive>true</Embed-Transitive>
diff --git a/oracle-plugin/src/main/java/io/cdap/plugin/oracle/OracleSourceDBRecord.java b/oracle-plugin/src/main/java/io/cdap/plugin/oracle/OracleSourceDBRecord.java
index 3f7c2a20a..0e333a00c 100644
--- a/oracle-plugin/src/main/java/io/cdap/plugin/oracle/OracleSourceDBRecord.java
+++ b/oracle-plugin/src/main/java/io/cdap/plugin/oracle/OracleSourceDBRecord.java
@@ -169,6 +169,16 @@ protected void writeNonNullToDB(PreparedStatement stmt, Schema fieldSchema,
     }
   }
 
+
+
+  private String getXmlDataAsString(ResultSet resultSet, int columnIndex) throws SQLException {
+    // Retrieve the XML data as a String
+    String xmlString = resultSet.getString(columnIndex);
+
+    //Check if the XML content is not null and return it
+    return xmlString != null ? xmlString : "No XML data present."; // Testing purpose
+  }
+
   /**
    * Creates an instance of 'oracle.sql.TIMESTAMPTZ' which corresponds to the specified timestamp with time zone string.
    * @param connection sql connection.
@@ -341,6 +351,10 @@ private void handleOracleSpecificType(ResultSet resultSet, StructuredRecord.Buil
       case OracleSourceSchemaReader.LONG_RAW:
         recordBuilder.set(field.getName(), resultSet.getBytes(columnIndex));
         break;
+      case OracleSourceSchemaReader.XML_TYPE:
+        String xmlData = getXmlDataAsString(resultSet, columnIndex);
+        recordBuilder.set(field.getName(), xmlData);
+        break;
       case Types.DECIMAL:
       case Types.NUMERIC:
         // This is the only way to differentiate FLOAT/REAL columns from other numeric columns, that based on NUMBER.
diff --git a/oracle-plugin/src/main/java/io/cdap/plugin/oracle/OracleSourceSchemaReader.java b/oracle-plugin/src/main/java/io/cdap/plugin/oracle/OracleSourceSchemaReader.java
index 7d35f9bc7..aea89e153 100644
--- a/oracle-plugin/src/main/java/io/cdap/plugin/oracle/OracleSourceSchemaReader.java
+++ b/oracle-plugin/src/main/java/io/cdap/plugin/oracle/OracleSourceSchemaReader.java
@@ -43,6 +43,7 @@ public class OracleSourceSchemaReader extends CommonSchemaReader {
   public static final int BFILE = -13;
   public static final int LONG = -1;
   public static final int LONG_RAW = -4;
+  public static final int XML_TYPE = 2009; // Add XMLType constant
 
   /**
    * Logger instance for Oracle Schema reader.
@@ -61,7 +62,7 @@ public class OracleSourceSchemaReader extends CommonSchemaReader {
     LONG,
     LONG_RAW,
     Types.NUMERIC,
-    Types.DECIMAL
+    Types.DECIMAL, XML_TYPE
   );
 
   private final String sessionID;
@@ -95,6 +96,9 @@ public Schema getSchema(ResultSetMetaData metadata, int index) throws SQLExcepti
       case INTERVAL_DS:
       case INTERVAL_YM:
       case LONG:
+      case XML_TYPE:
+        LOG.info(String.format("Column '%s' is of XMLType, converting to STRING for CDAP.",
+                metadata.getColumnName(index)));
         return Schema.of(Schema.Type.STRING);
       case Types.NUMERIC:
       case Types.DECIMAL: