Fix startup Delta package resolution for Spark 4.x runtimes

README.md

@@ -82,7 +82,7 @@ Once the image has been built or you have downloaded the correct image, you can
 
 In the following instructions, the variable `${DELTA_PACKAGE_VERSION}` refers to the Delta Lake Package version.
 
-The current version is `delta-spark_2.13:4.0.0` which corresponds to Apache Spark 4.x release line.
+For Spark 4.x, use Spark-version-specific Delta artifacts: `delta-spark_4.1_2.13:<version>` for Spark 4.1 and `delta-spark_4.0_2.13:<version>` for Spark 4.0 (use `delta-spark_2.13:<version>` for older Spark lines).
 
 ## Choose an Interface
 

startup.sh

@@ -4,8 +4,30 @@ source "$HOME/.cargo/env"
 
 export PYSPARK_DRIVER_PYTHON=jupyter
 export PYSPARK_DRIVER_PYTHON_OPTS='lab --ip=0.0.0.0'
-export DELTA_SPARK_VERSION='4.0.1'
-export DELTA_PACKAGE_VERSION=delta-spark_2.13:${DELTA_SPARK_VERSION}
+
+# Default Delta version; can be overridden by setting DELTA_SPARK_VERSION in the environment
+: "${DELTA_SPARK_VERSION:=4.1.0}"
+
+# Detect the Spark major.minor version from the running runtime (e.g. "4.1")
+SPARK_FULL_VERSION=$("${SPARK_HOME}/bin/spark-submit" --version 2>&1 \
+  | grep -oE '[0-9]+\.[0-9]+\.[0-9]+' | head -1)
+SPARK_MAJOR_MINOR=$(echo "${SPARK_FULL_VERSION}" | cut -d. -f1,2)
+
+# Select the Delta Maven artifact that matches this Spark version.
+# Spark 4.1 and 4.0 each publish a Spark-specific artifact; older releases use the generic one.
+case "${SPARK_MAJOR_MINOR}" in
+  4.1)
+    DELTA_ARTIFACT="delta-spark_4.1_2.13"
+    ;;
+  4.0)
+    DELTA_ARTIFACT="delta-spark_4.0_2.13"
+    ;;
+  *)
+    DELTA_ARTIFACT="delta-spark_2.13"
+    ;;
+esac
+
+export DELTA_PACKAGE_VERSION="${DELTA_ARTIFACT}:${DELTA_SPARK_VERSION}"
 
 $SPARK_HOME/bin/pyspark --packages io.delta:${DELTA_PACKAGE_VERSION} \
   --conf "spark.driver.extraJavaOptions=-Divy.cache.dir=/tmp -Divy.home=/tmp -Dio.netty.tryReflectionSetAccessible=true" \

tests/test_docker.sh

@@ -172,15 +172,40 @@ run_test "spark-submit is on PATH"  "spark-submit --version"
 run_test "pyspark is on PATH"       "pyspark --version"
 
 # ---------------------------------------------------------------
-# 6. Rust toolchain
+# 6. startup.sh artifact resolution
+# ---------------------------------------------------------------
+
+section "startup.sh artifact resolution"
+run_test_verbose "startup.sh selects Spark 4.1 Delta artifact" \
+  "set -euo pipefail
+mkdir -p /tmp/mock-spark/bin
+printf '%s\n' '#!/usr/bin/env bash' 'echo \"Spark version 4.1.1\" >&2' > /tmp/mock-spark/bin/spark-submit
+printf '%s\n' '#!/usr/bin/env bash' 'echo \"\$*\"' > /tmp/mock-spark/bin/pyspark
+chmod +x /tmp/mock-spark/bin/spark-submit /tmp/mock-spark/bin/pyspark
+startup_output=\$(SPARK_HOME=/tmp/mock-spark DELTA_SPARK_VERSION=4.1.0 bash startup.sh 2>&1)
+echo \"\$startup_output\"
+[[ \"\$startup_output\" == *\"--packages io.delta:delta-spark_4.1_2.13:4.1.0\"* ]]"
+
+run_test_verbose "startup.sh selects Spark 4.0 Delta artifact" \
+  "set -euo pipefail
+mkdir -p /tmp/mock-spark/bin
+printf '%s\n' '#!/usr/bin/env bash' 'echo \"Spark version 4.0.3\" >&2' > /tmp/mock-spark/bin/spark-submit
+printf '%s\n' '#!/usr/bin/env bash' 'echo \"\$*\"' > /tmp/mock-spark/bin/pyspark
+chmod +x /tmp/mock-spark/bin/spark-submit /tmp/mock-spark/bin/pyspark
+startup_output=\$(SPARK_HOME=/tmp/mock-spark DELTA_SPARK_VERSION=4.1.0 bash startup.sh 2>&1)
+echo \"\$startup_output\"
+[[ \"\$startup_output\" == *\"--packages io.delta:delta-spark_4.0_2.13:4.1.0\"* ]]"
+
+# ---------------------------------------------------------------
+# 7. Rust toolchain
 # ---------------------------------------------------------------
 
 section "Rust Toolchain"
 run_test "rustc is available"       'source "$HOME/.cargo/env" && rustc --version'
 run_test "cargo is available"       'source "$HOME/.cargo/env" && cargo --version'
 
 # ---------------------------------------------------------------
-# 7. Functional: delta-rs (Python) write/read via Polars
+# 8. Functional: delta-rs (Python) write/read via Polars
 # ---------------------------------------------------------------
 
 section "Functional: delta-rs + Polars"
@@ -215,7 +240,7 @@ print('Polars Delta append OK')
 \""
 
 # ---------------------------------------------------------------
-# 8. Functional: deltalake Python API
+# 9. Functional: deltalake Python API
 # ---------------------------------------------------------------
 
 section "Functional: deltalake Python API"