From ff1ef378542523ca2b1ee28899c3587653941a63 Mon Sep 17 00:00:00 2001
From: comphead <comphead@ukr.net>
Date: Mon, 29 Dec 2025 09:52:41 -0800
Subject: [PATCH 1/2] chore: Add TPCDS benchmark comparison for PR

---
 benchmarks/README.md       | 17 ++++++++++++++++-
 benchmarks/bench.sh        |  2 +-
 datafusion/expr/src/udf.rs |  6 +++---
 3 files changed, 20 insertions(+), 5 deletions(-)

diff --git a/benchmarks/README.md b/benchmarks/README.md
index 0b71628b2db1..b0aa4277e687 100644
--- a/benchmarks/README.md
+++ b/benchmarks/README.md
@@ -157,7 +157,7 @@ To get data in `DATA_DIR` for TPCDS, please follow instructions in `./benchmarks
 DATA_DIR=../../datafusion-benchmarks/tpcds/data/sf1/ ./benchmarks/compare_tpcds.sh main mybranch
 ```
 
-Alternatively you can compare manually followng the example velor
+Alternatively, you can compare manually following the example below
 
 ```shell
 git checkout main
@@ -240,6 +240,21 @@ Benchmark tpch_mem.json
 └──────────────┴──────────────┴──────────────┴───────────────┘
 ```
 
+## Comparing performance of main and a PR
+
+### TPCDS
+
+Considering you already have TPCDS data locally
+
+```shell
+export DATA_DIR=../../datafusion-benchmarks/tpcds/data/sf1/
+export PR_NUMBER=19464
+gh pr checkout $PR_NUMBER --repo apache/datafusion -b pr-$PR_NUMBER
+git checkout main
+git pull
+./benchmarks/compare_tpcds.sh main pr-$PR_NUMBER
+```
+
 ### Running Benchmarks Manually
 
 Assuming data is in the `data` directory, the `tpch` benchmark can be run with a command like this:
diff --git a/benchmarks/bench.sh b/benchmarks/bench.sh
index d5fa52d7f00e..6679405623d0 100755
--- a/benchmarks/bench.sh
+++ b/benchmarks/bench.sh
@@ -684,7 +684,7 @@ run_tpch_mem() {
 
 # Runs the tpcds benchmark
 run_tpcds() {
-    TPCDS_DIR="${DATA_DIR}/tpcds_sf1"
+    TPCDS_DIR="${DATA_DIR}"
 
     # Check if TPCDS data directory and representative file exists
     if [ ! -f "${TPCDS_DIR}/web_site.parquet" ]; then
diff --git a/datafusion/expr/src/udf.rs b/datafusion/expr/src/udf.rs
index 26d7fc99cb17..28a07ad76101 100644
--- a/datafusion/expr/src/udf.rs
+++ b/datafusion/expr/src/udf.rs
@@ -24,10 +24,10 @@ use crate::sort_properties::{ExprProperties, SortProperties};
 use crate::udf_eq::UdfEq;
 use crate::{ColumnarValue, Documentation, Expr, Signature};
 use arrow::datatypes::{DataType, Field, FieldRef};
+#[cfg(debug_assertions)]
+use datafusion_common::assert_or_internal_err;
 use datafusion_common::config::ConfigOptions;
-use datafusion_common::{
-    ExprSchema, Result, ScalarValue, assert_or_internal_err, not_impl_err,
-};
+use datafusion_common::{ExprSchema, Result, ScalarValue, not_impl_err};
 use datafusion_expr_common::dyn_eq::{DynEq, DynHash};
 use datafusion_expr_common::interval_arithmetic::Interval;
 use std::any::Any;

From e8f5f97c021591bb8a4b1f942837943954a297a2 Mon Sep 17 00:00:00 2001
From: comphead <comphead@ukr.net>
Date: Thu, 1 Jan 2026 19:55:26 -0800
Subject: [PATCH 2/2] chore: Add TPCDS benchmark comparison for PR

---
 benchmarks/README.md | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/benchmarks/README.md b/benchmarks/README.md
index b0aa4277e687..7e9818aef24f 100644
--- a/benchmarks/README.md
+++ b/benchmarks/README.md
@@ -249,12 +249,14 @@ Considering you already have TPCDS data locally
 ```shell
 export DATA_DIR=../../datafusion-benchmarks/tpcds/data/sf1/
 export PR_NUMBER=19464
-gh pr checkout $PR_NUMBER --repo apache/datafusion -b pr-$PR_NUMBER
+git fetch upstream pull/$PR_NUMBER/head:pr-$PR_NUMBER
 git checkout main
 git pull
 ./benchmarks/compare_tpcds.sh main pr-$PR_NUMBER
 ```
 
+Note: if `gh` is installed, you can also run `gh pr checkout $PR_NUMBER` instead of `git fetch upstream pull/$PR_NUMBER/head:pr-$PR_NUMBER`
+
 ### Running Benchmarks Manually
 
 Assuming data is in the `data` directory, the `tpch` benchmark can be run with a command like this: