diff --git a/spark/v3.0/spark/src/main/java/org/apache/iceberg/spark/SparkCatalog.java b/spark/v3.0/spark/src/main/java/org/apache/iceberg/spark/SparkCatalog.java
index ae921c5200a8..d80b9bf5476d 100644
--- a/spark/v3.0/spark/src/main/java/org/apache/iceberg/spark/SparkCatalog.java
+++ b/spark/v3.0/spark/src/main/java/org/apache/iceberg/spark/SparkCatalog.java
@@ -488,6 +488,10 @@ private Pair<Table, Long> load(Identifier ident) {
       return Pair.of(icebergCatalog.loadTable(buildIdentifier(ident)), null);
 
     } catch (org.apache.iceberg.exceptions.NoSuchTableException e) {
+      if (ident.namespace().length == 0) {
+        throw e;
+      }
+
       // if the original load didn't work, the identifier may be extended and include a snapshot selector
       TableIdentifier namespaceAsIdent = buildIdentifier(namespaceToIdentifier(ident.namespace()));
       Table table;
@@ -567,6 +571,8 @@ private Pair<Table, Long> loadFromPathIdentifier(PathIdentifier ident) {
   }
 
   private Identifier namespaceToIdentifier(String[] namespace) {
+    Preconditions.checkArgument(namespace.length > 0,
+        "Cannot convert empty namespace to identifier");
     String[] ns = Arrays.copyOf(namespace, namespace.length - 1);
     String name = namespace[ns.length];
     return Identifier.of(ns, name);
diff --git a/spark/v3.0/spark/src/test/java/org/apache/iceberg/spark/sql/TestCreateTable.java b/spark/v3.0/spark/src/test/java/org/apache/iceberg/spark/sql/TestCreateTable.java
index 303cbb5f932b..0a4c9368cb96 100644
--- a/spark/v3.0/spark/src/test/java/org/apache/iceberg/spark/sql/TestCreateTable.java
+++ b/spark/v3.0/spark/src/test/java/org/apache/iceberg/spark/sql/TestCreateTable.java
@@ -67,6 +67,17 @@ public void testCreateTable() {
         table.properties().get(TableProperties.DEFAULT_FILE_FORMAT));
   }
 
+  @Test
+  public void testCreateTableInRootNamespace() {
+    Assume.assumeTrue("Hadoop has no default namespace configured", "testhadoop".equals(catalogName));
+
+    try {
+      sql("CREATE TABLE %s.table (id bigint) USING iceberg", catalogName);
+    } finally {
+      sql("DROP TABLE IF EXISTS %s.table", catalogName);
+    }
+  }
+
   @Test
   public void testCreateTableUsingParquet() {
     Assume.assumeTrue(
diff --git a/spark/v3.1/spark/src/main/java/org/apache/iceberg/spark/SparkCatalog.java b/spark/v3.1/spark/src/main/java/org/apache/iceberg/spark/SparkCatalog.java
index ae921c5200a8..d80b9bf5476d 100644
--- a/spark/v3.1/spark/src/main/java/org/apache/iceberg/spark/SparkCatalog.java
+++ b/spark/v3.1/spark/src/main/java/org/apache/iceberg/spark/SparkCatalog.java
@@ -488,6 +488,10 @@ private Pair<Table, Long> load(Identifier ident) {
       return Pair.of(icebergCatalog.loadTable(buildIdentifier(ident)), null);
 
     } catch (org.apache.iceberg.exceptions.NoSuchTableException e) {
+      if (ident.namespace().length == 0) {
+        throw e;
+      }
+
       // if the original load didn't work, the identifier may be extended and include a snapshot selector
       TableIdentifier namespaceAsIdent = buildIdentifier(namespaceToIdentifier(ident.namespace()));
       Table table;
@@ -567,6 +571,8 @@ private Pair<Table, Long> loadFromPathIdentifier(PathIdentifier ident) {
   }
 
   private Identifier namespaceToIdentifier(String[] namespace) {
+    Preconditions.checkArgument(namespace.length > 0,
+        "Cannot convert empty namespace to identifier");
     String[] ns = Arrays.copyOf(namespace, namespace.length - 1);
     String name = namespace[ns.length];
     return Identifier.of(ns, name);
diff --git a/spark/v3.1/spark/src/test/java/org/apache/iceberg/spark/sql/TestCreateTable.java b/spark/v3.1/spark/src/test/java/org/apache/iceberg/spark/sql/TestCreateTable.java
index 303cbb5f932b..0a4c9368cb96 100644
--- a/spark/v3.1/spark/src/test/java/org/apache/iceberg/spark/sql/TestCreateTable.java
+++ b/spark/v3.1/spark/src/test/java/org/apache/iceberg/spark/sql/TestCreateTable.java
@@ -67,6 +67,17 @@ public void testCreateTable() {
         table.properties().get(TableProperties.DEFAULT_FILE_FORMAT));
   }
 
+  @Test
+  public void testCreateTableInRootNamespace() {
+    Assume.assumeTrue("Hadoop has no default namespace configured", "testhadoop".equals(catalogName));
+
+    try {
+      sql("CREATE TABLE %s.table (id bigint) USING iceberg", catalogName);
+    } finally {
+      sql("DROP TABLE IF EXISTS %s.table", catalogName);
+    }
+  }
+
   @Test
   public void testCreateTableUsingParquet() {
     Assume.assumeTrue(
diff --git a/spark/v3.2/spark/src/main/java/org/apache/iceberg/spark/SparkCatalog.java b/spark/v3.2/spark/src/main/java/org/apache/iceberg/spark/SparkCatalog.java
index ae921c5200a8..d80b9bf5476d 100644
--- a/spark/v3.2/spark/src/main/java/org/apache/iceberg/spark/SparkCatalog.java
+++ b/spark/v3.2/spark/src/main/java/org/apache/iceberg/spark/SparkCatalog.java
@@ -488,6 +488,10 @@ private Pair<Table, Long> load(Identifier ident) {
       return Pair.of(icebergCatalog.loadTable(buildIdentifier(ident)), null);
 
     } catch (org.apache.iceberg.exceptions.NoSuchTableException e) {
+      if (ident.namespace().length == 0) {
+        throw e;
+      }
+
       // if the original load didn't work, the identifier may be extended and include a snapshot selector
       TableIdentifier namespaceAsIdent = buildIdentifier(namespaceToIdentifier(ident.namespace()));
       Table table;
@@ -567,6 +571,8 @@ private Pair<Table, Long> loadFromPathIdentifier(PathIdentifier ident) {
   }
 
   private Identifier namespaceToIdentifier(String[] namespace) {
+    Preconditions.checkArgument(namespace.length > 0,
+        "Cannot convert empty namespace to identifier");
     String[] ns = Arrays.copyOf(namespace, namespace.length - 1);
     String name = namespace[ns.length];
     return Identifier.of(ns, name);
diff --git a/spark/v3.2/spark/src/test/java/org/apache/iceberg/spark/sql/TestCreateTable.java b/spark/v3.2/spark/src/test/java/org/apache/iceberg/spark/sql/TestCreateTable.java
index 303cbb5f932b..0a4c9368cb96 100644
--- a/spark/v3.2/spark/src/test/java/org/apache/iceberg/spark/sql/TestCreateTable.java
+++ b/spark/v3.2/spark/src/test/java/org/apache/iceberg/spark/sql/TestCreateTable.java
@@ -67,6 +67,17 @@ public void testCreateTable() {
         table.properties().get(TableProperties.DEFAULT_FILE_FORMAT));
   }
 
+  @Test
+  public void testCreateTableInRootNamespace() {
+    Assume.assumeTrue("Hadoop has no default namespace configured", "testhadoop".equals(catalogName));
+
+    try {
+      sql("CREATE TABLE %s.table (id bigint) USING iceberg", catalogName);
+    } finally {
+      sql("DROP TABLE IF EXISTS %s.table", catalogName);
+    }
+  }
+
   @Test
   public void testCreateTableUsingParquet() {
     Assume.assumeTrue(