diff --git a/native/spark-expr/Cargo.toml b/native/spark-expr/Cargo.toml
index 4addc026d8..4f0e90bd74 100644
--- a/native/spark-expr/Cargo.toml
+++ b/native/spark-expr/Cargo.toml
@@ -56,6 +56,10 @@ datafusion = { workspace = true, features = ["sql"] }
 name = "datafusion_comet_spark_expr"
 path = "src/lib.rs"
 
+[[bench]]
+name = "log"
+harness = false
+
 [[bench]]
 name = "cast_from_string"
 harness = false
diff --git a/native/spark-expr/benches/log.rs b/native/spark-expr/benches/log.rs
new file mode 100644
index 0000000000..60a980943f
--- /dev/null
+++ b/native/spark-expr/benches/log.rs
@@ -0,0 +1,65 @@
+// Licensed to the Apache Software Foundation (ASF) under one
+// or more contributor license agreements.  See the NOTICE file
+// distributed with this work for additional information
+// regarding copyright ownership.  The ASF licenses this file
+// to you under the Apache License, Version 2.0 (the
+// "License"); you may not use this file except in compliance
+// with the License.  You may obtain a copy of the License at
+//
+//   http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing,
+// software distributed under the License is distributed on an
+// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, either express or implied.  See the License for the
+// specific language governing permissions and limitations
+// under the License.
+
+use arrow::array::{ArrayRef, Float64Array};
+use criterion::{criterion_group, criterion_main, Criterion};
+use datafusion::common::ScalarValue;
+use datafusion::physical_plan::ColumnarValue;
+use datafusion_comet_spark_expr::spark_log;
+use std::hint::black_box;
+use std::sync::Arc;
+
+/// Build a `Float64Array` of `n` positive values with every 10th row null.
+fn positive_f64(n: usize) -> ArrayRef {
+    let vals: Vec<Option<f64>> = (0..n)
+        .map(|i| {
+            if i % 10 == 0 {
+                None
+            } else {
+                Some(1.0 + (i as f64) * 0.5)
+            }
+        })
+        .collect();
+    Arc::new(Float64Array::from(vals))
+}
+
+fn criterion_benchmark(c: &mut Criterion) {
+    let n = 8192;
+    let column = positive_f64(n);
+
+    // log(base_literal, value_column) — the common Spark shape where the base
+    // is a constant, exercising the scalar-base / array-value path.
+    c.bench_function("spark_log: scalar base, array value", |b| {
+        let args = vec![
+            ColumnarValue::Scalar(ScalarValue::Float64(Some(10.0))),
+            ColumnarValue::Array(Arc::clone(&column)),
+        ];
+        b.iter(|| black_box(spark_log(black_box(&args)).unwrap()))
+    });
+
+    // log(base_column, value_literal) — the array-base / scalar-value path.
+    c.bench_function("spark_log: array base, scalar value", |b| {
+        let args = vec![
+            ColumnarValue::Array(Arc::clone(&column)),
+            ColumnarValue::Scalar(ScalarValue::Float64(Some(100.0))),
+        ];
+        b.iter(|| black_box(spark_log(black_box(&args)).unwrap()))
+    });
+}
+
+criterion_group!(benches, criterion_benchmark);
+criterion_main!(benches);
diff --git a/native/spark-expr/src/math_funcs/log.rs b/native/spark-expr/src/math_funcs/log.rs
index 499d4f33ed..9239bc1fa4 100644
--- a/native/spark-expr/src/math_funcs/log.rs
+++ b/native/spark-expr/src/math_funcs/log.rs
@@ -94,9 +94,24 @@ pub fn spark_log(args: &[ColumnarValue]) -> Result<ColumnarValue, DataFusionErro
                         val_arr.data_type()
                     ))
                 })?;
+            // `base` is constant, so a non-positive base makes the whole result
+            // null; hoist that check and `ln(base)` out of the per-element loop.
+            if base <= 0.0 {
+                let result = Float64Array::new_null(val_arr.len());
+                return Ok(ColumnarValue::Array(Arc::new(result)));
+            }
+            let ln_base = base.ln();
             let result: Float64Array = values
                 .iter()
-                .map(|v| v.and_then(|value| compute(base, value)))
+                .map(|v| {
+                    v.and_then(|value| {
+                        if value <= 0.0 {
+                            None
+                        } else {
+                            Some(value.ln() / ln_base)
+                        }
+                    })
+                })
                 .collect();
             Ok(ColumnarValue::Array(Arc::new(result)))
         }
@@ -122,9 +137,24 @@ pub fn spark_log(args: &[ColumnarValue]) -> Result<ColumnarValue, DataFusionErro
                         base_arr.data_type()
                     ))
                 })?;
+            // `value` is constant, so a non-positive value makes the whole result
+            // null; hoist that check and `ln(value)` out of the per-element loop.
+            if value <= 0.0 {
+                let result = Float64Array::new_null(base_arr.len());
+                return Ok(ColumnarValue::Array(Arc::new(result)));
+            }
+            let ln_value = value.ln();
             let result: Float64Array = bases
                 .iter()
-                .map(|b| b.and_then(|base| compute(base, value)))
+                .map(|b| {
+                    b.and_then(|base| {
+                        if base <= 0.0 {
+                            None
+                        } else {
+                            Some(ln_value / base.ln())
+                        }
+                    })
+                })
                 .collect();
             Ok(ColumnarValue::Array(Arc::new(result)))
         }