GH-44831: [CI] Build Java for integration tests from arrow-java (#44932)

### Rationale for this change

We moved `java/` in apache/arrow to apache/arrow-java. We should use apache/arrow-java to integration tests.

### What changes are included in this PR?

* Remove `java/`
* Use apache/arrow-java for integration test

### Are these changes tested?

Yes.

### Are there any user-facing changes?

No.

* GitHub Issue: #44945
* GitHub Issue: #44831

Authored-by: David Li <[email protected]>
Signed-off-by: Sutou Kouhei <[email protected]>

.github/workflows/integration.yml

@@ -91,6 +91,11 @@ jobs:
         with:
           repository: apache/arrow-go
           path: go
+      - name: Checkout Arrow Java
+        uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332 # v4.1.7
+        with:
+          repository: apache/arrow-java
+          path: java
       - name: Free up disk space
         run: |
           ci/scripts/util_free_space.sh
@@ -115,6 +120,7 @@ jobs:
           archery docker run \
             -e ARCHERY_DEFAULT_BRANCH=${{ github.event.repository.default_branch }} \
             -e ARCHERY_INTEGRATION_WITH_GO=1 \
+            -e ARCHERY_INTEGRATION_WITH_JAVA=1 \
             -e ARCHERY_INTEGRATION_WITH_NANOARROW=1 \
             -e ARCHERY_INTEGRATION_WITH_RUST=1 \
             conda-integration

ci/scripts/integration_arrow.sh

@@ -26,10 +26,9 @@ gold_dir=$arrow_dir/testing/data/arrow-ipc-stream/integration
 
 : ${ARROW_INTEGRATION_CPP:=ON}
 : ${ARROW_INTEGRATION_CSHARP:=ON}
-: ${ARROW_INTEGRATION_JAVA:=ON}
 : ${ARROW_INTEGRATION_JS:=ON}
 
-: ${ARCHERY_INTEGRATION_TARGET_IMPLEMENTATIONS:=cpp,csharp,java,js}
+: ${ARCHERY_INTEGRATION_TARGET_IMPLEMENTATIONS:=cpp,csharp,js}
 export ARCHERY_INTEGRATION_TARGET_IMPLEMENTATIONS
 
 . ${arrow_dir}/ci/scripts/util_log.sh
@@ -43,7 +42,7 @@ github_actions_group_begin "Integration: Prepare: Dependencies"
 if [ "${ARROW_INTEGRATION_CSHARP}" == "ON" ]; then
     pip install pythonnet
 fi
-if [ "${ARROW_INTEGRATION_JAVA}" == "ON" ]; then
+if [ "${ARCHERY_INTEGRATION_WITH_JAVA}" -gt "0" ]; then
     pip install jpype1
 fi
 github_actions_group_end
@@ -66,7 +65,6 @@ time archery integration \
     --run-flight \
     --with-cpp=$([ "$ARROW_INTEGRATION_CPP" == "ON" ] && echo "1" || echo "0") \
     --with-csharp=$([ "$ARROW_INTEGRATION_CSHARP" == "ON" ] && echo "1" || echo "0") \
-    --with-java=$([ "$ARROW_INTEGRATION_JAVA" == "ON" ] && echo "1" || echo "0") \
     --with-js=$([ "$ARROW_INTEGRATION_JS" == "ON" ] && echo "1" || echo "0") \
     --gold-dirs=$gold_dir/0.14.1 \
     --gold-dirs=$gold_dir/0.17.1 \

ci/scripts/integration_arrow_build.sh

@@ -24,7 +24,6 @@ build_dir=${2}
 
 : ${ARROW_INTEGRATION_CPP:=ON}
 : ${ARROW_INTEGRATION_CSHARP:=ON}
-: ${ARROW_INTEGRATION_JAVA:=ON}
 : ${ARROW_INTEGRATION_JS:=ON}
 
 . ${arrow_dir}/ci/scripts/util_log.sh
@@ -56,12 +55,12 @@ fi
 github_actions_group_end
 
 github_actions_group_begin "Integration: Build: Java"
-if [ "${ARROW_INTEGRATION_JAVA}" == "ON" ]; then
+if [ "${ARCHERY_INTEGRATION_WITH_JAVA}" -gt "0" ]; then
     export ARROW_JAVA_CDATA="ON"
     export JAVA_JNI_CMAKE_ARGS="-DARROW_JAVA_JNI_ENABLE_DEFAULT=OFF -DARROW_JAVA_JNI_ENABLE_C=ON"
 
-    ${arrow_dir}/ci/scripts/java_jni_build.sh ${arrow_dir} ${ARROW_HOME} ${build_dir} /tmp/dist/java
-    ${arrow_dir}/ci/scripts/java_build.sh ${arrow_dir} ${build_dir} /tmp/dist/java
+    ${arrow_dir}/java/ci/scripts/java_jni_build.sh "${arrow_dir}/java" "${ARROW_HOME}" "${build_dir}/java/" /tmp/dist/java
+    ${arrow_dir}/java/ci/scripts/java_build.sh "${arrow_dir}/java" "${build_dir}/java" /tmp/dist/java
 fi
 github_actions_group_end
 

dev/archery/archery/cli.py

@@ -733,7 +733,8 @@ def _set_default(opt, default):
 @click.option('--with-csharp', type=bool, default=False,
               help='Include C# in integration tests')
 @click.option('--with-java', type=bool, default=False,
-              help='Include Java in integration tests')
+              help='Include Java in integration tests',
+              envvar="ARCHERY_INTEGRATION_WITH_JAVA")
 @click.option('--with-js', type=bool, default=False,
               help='Include JavaScript in integration tests')
 @click.option('--with-go', type=bool, default=False,

dev/archery/archery/integration/tester_java.py

@@ -24,6 +24,7 @@
 from . import cdata
 from .tester import Tester, CDataExporter, CDataImporter
 from .util import run_cmd, log
+from ..utils.source import ARROW_ROOT_DEFAULT
 
 
 ARROW_BUILD_ROOT = os.environ.get(
@@ -34,7 +35,7 @@
 
 def load_version_from_pom():
     import xml.etree.ElementTree as ET
-    tree = ET.parse(os.path.join(ARROW_BUILD_ROOT, 'java', 'pom.xml'))
+    tree = ET.parse(os.path.join(ARROW_ROOT_DEFAULT, 'java', 'pom.xml'))
     tag_pattern = '{http://maven.apache.org/POM/4.0.0}version'
     version_tag = list(tree.getroot().findall(tag_pattern))[0]
     return version_tag.text