More minimal default options, add ARROW_CSV CMake option

ci/scripts/cpp_build.sh

@@ -50,6 +50,7 @@ cmake -G "${CMAKE_GENERATOR:-Ninja}" \
       -DARROW_BUILD_TESTS=${ARROW_BUILD_TESTS:-OFF} \
       -DARROW_BUILD_UTILITIES=${ARROW_BUILD_UTILITIES:-ON} \
       -DARROW_COMPUTE=${ARROW_COMPUTE:-ON} \
+      -DARROW_CSV=${ARROW_CSV:-ON} \
       -DARROW_CUDA=${ARROW_CUDA:-OFF} \
       -DARROW_CXXFLAGS=${ARROW_CXXFLAGS:-} \
       -DARROW_DATASET=${ARROW_DATASET:-ON} \
@@ -60,6 +61,7 @@ cmake -G "${CMAKE_GENERATOR:-Ninja}" \
       -DARROW_FUZZING=${ARROW_FUZZING:-OFF} \
       -DARROW_FUZZING=${ARROW_FUZZING:-OFF} \
       -DARROW_JNI=${ARROW_JNI:-OFF} \
+      -DARROW_JSON=${ARROW_JSON:-ON} \
       -DARROW_GANDIVA_JAVA=${ARROW_GANDIVA_JAVA:-OFF} \
       -DARROW_GANDIVA_PC_CXX_FLAGS=${ARROW_GANDIVA_PC_CXX_FLAGS:-} \
       -DARROW_GANDIVA=${ARROW_GANDIVA:-OFF} \

cpp/CMakeLists.txt

@@ -266,7 +266,7 @@ endif(UNIX)
 # Set up various options
 #
 
-if(ARROW_BUILD_TESTS OR ARROW_BUILD_INTEGRATION)
+if(ARROW_BUILD_BENCHMARKS OR ARROW_BUILD_TESTS OR ARROW_BUILD_INTEGRATION)
   set(ARROW_JSON ON)
 endif()
 

cpp/cmake_modules/DefineOptions.cmake

@@ -141,20 +141,22 @@ if("${CMAKE_SOURCE_DIR}" STREQUAL "${CMAKE_CURRENT_SOURCE_DIR}")
 
   define_option(ARROW_BUILD_UTILITIES "Build Arrow commandline utilities" OFF)
 
-  define_option(ARROW_COMPUTE "Build the Arrow Compute Modules" ON)
+  define_option(ARROW_COMPUTE "Build the Arrow Compute Modules" OFF)
+
+  define_option(ARROW_CSV "Build the Arrow CSV Parser Module" OFF)
 
   define_option(ARROW_CUDA "Build the Arrow CUDA extensions (requires CUDA toolkit)" OFF)
 
-  define_option(ARROW_DATASET "Build the Arrow Dataset Modules" ON)
+  define_option(ARROW_DATASET "Build the Arrow Dataset Modules" OFF)
 
-  define_option(ARROW_FILESYSTEM "Build the Arrow Filesystem Layer" ON)
+  define_option(ARROW_FILESYSTEM "Build the Arrow Filesystem Layer" OFF)
 
   define_option(ARROW_FLIGHT
                 "Build the Arrow Flight RPC System (requires GRPC, Protocol Buffers)" OFF)
 
   define_option(ARROW_GANDIVA "Build the Gandiva libraries" OFF)
 
-  define_option(ARROW_HDFS "Build the Arrow HDFS bridge" ON)
+  define_option(ARROW_HDFS "Build the Arrow HDFS bridge" OFF)
 
   define_option(ARROW_HIVESERVER2 "Build the HiveServer2 client and Arrow adapter" OFF)
 
@@ -164,7 +166,7 @@ if("${CMAKE_SOURCE_DIR}" STREQUAL "${CMAKE_CURRENT_SOURCE_DIR}")
 
   define_option(ARROW_JNI "Build the Arrow JNI lib" OFF)
 
-  define_option(ARROW_JSON "Build Arrow with JSON support (requires RapidJSON)" ON)
+  define_option(ARROW_JSON "Build Arrow with JSON support (requires RapidJSON)" OFF)
 
   define_option(ARROW_MIMALLOC "Build the Arrow mimalloc-based allocator" OFF)
 

cpp/examples/minimal_build/build.sh

@@ -27,15 +27,9 @@ NPROC=$(nproc)
 mkdir $BUILD_DIR
 pushd $BUILD_DIR
 
-cmake /arrow/cpp \
-      -DARROW_COMPUTE=OFF \
-      -DARROW_DATASET=OFF \
-      -DARROW_FILESYSTEM=OFF \
-      -DARROW_HDFS=OFF \
-      -DARROW_JEMALLOC=OFF \
-      -DARROW_JSON=OFF \
-      -DARROW_USE_GLOG=OFF \
-      -DARROW_BUILD_UTILITIES=OFF
+cmake /arrow/cpp -DARROW_JEMALLOC=OFF \
+      -DARROW_BUILD_BENCHMARKS=${ARROW_BUILD_BENCHMARKS:-OFF} \
+      -DARROW_BUILD_TESTS=${ARROW_BUILD_TESTS:-OFF}
 
 make -j$NPROC
 make install

cpp/src/arrow/CMakeLists.txt

@@ -109,12 +109,6 @@ set(ARROW_SRCS
     tensor.cc
     type.cc
     visitor.cc
-    csv/converter.cc
-    csv/chunker.cc
-    csv/column_builder.cc
-    csv/options.cc
-    csv/parser.cc
-    csv/reader.cc
     io/buffered.cc
     io/compressed.cc
     io/file.cc
@@ -154,22 +148,26 @@ set(ARROW_SRCS
     vendored/double-conversion/cached-powers.cc
     vendored/double-conversion/fixed-dtoa.cc
     vendored/double-conversion/diy-fp.cc
-    vendored/double-conversion/strtod.cc
-    vendored/uriparser/UriCommon.c
-    vendored/uriparser/UriCompare.c
-    vendored/uriparser/UriEscape.c
-    vendored/uriparser/UriFile.c
-    vendored/uriparser/UriIp4Base.c
-    vendored/uriparser/UriIp4.c
-    vendored/uriparser/UriMemory.c
-    vendored/uriparser/UriNormalizeBase.c
-    vendored/uriparser/UriNormalize.c
-    vendored/uriparser/UriParseBase.c
-    vendored/uriparser/UriParse.c
-    vendored/uriparser/UriQuery.c
-    vendored/uriparser/UriRecompose.c
-    vendored/uriparser/UriResolve.c
-    vendored/uriparser/UriShorten.c)
+    vendored/double-conversion/strtod.cc)
+
+if(ARROW_FILESYSTEM OR ARROW_FLIGHT)
+  list(APPEND ARROW_SRCS
+              vendored/uriparser/UriCommon.c
+              vendored/uriparser/UriCompare.c
+              vendored/uriparser/UriEscape.c
+              vendored/uriparser/UriFile.c
+              vendored/uriparser/UriIp4Base.c
+              vendored/uriparser/UriIp4.c
+              vendored/uriparser/UriMemory.c
+              vendored/uriparser/UriNormalizeBase.c
+              vendored/uriparser/UriNormalize.c
+              vendored/uriparser/UriParseBase.c
+              vendored/uriparser/UriParse.c
+              vendored/uriparser/UriQuery.c
+              vendored/uriparser/UriRecompose.c
+              vendored/uriparser/UriResolve.c
+              vendored/uriparser/UriShorten.c)
+endif()
 
 # Disable DLL exports in vendored uriparser library
 add_definitions(-DURI_STATIC_BUILD)
@@ -253,11 +251,21 @@ add_subdirectory(testing)
 #
 
 add_subdirectory(array)
-add_subdirectory(csv)
 add_subdirectory(io)
 add_subdirectory(util)
 add_subdirectory(vendored)
 
+if(ARROW_CSV)
+  list(APPEND ARROW_SRCS
+              csv/converter.cc
+              csv/chunker.cc
+              csv/column_builder.cc
+              csv/options.cc
+              csv/parser.cc
+              csv/reader.cc)
+  add_subdirectory(csv)
+endif()
+
 if(ARROW_COMPUTE)
   add_subdirectory(compute)
   list(APPEND ARROW_SRCS
@@ -489,12 +497,16 @@ add_arrow_test(public_api_test)
 add_arrow_test(result_test)
 add_arrow_test(scalar_test)
 add_arrow_test(status_test)
-add_arrow_test(stl_test)
 add_arrow_test(type_test)
 add_arrow_test(table_test)
 add_arrow_test(table_builder_test)
 add_arrow_test(tensor_test)
 add_arrow_test(sparse_tensor_test)
 
+if(ARROW_COMPUTE)
+  # This unit test uses compute code
+  add_arrow_test(stl_test)
+endif()
+
 add_arrow_benchmark(builder_benchmark)
 add_arrow_benchmark(type_benchmark)

cpp/src/arrow/array/CMakeLists.txt

@@ -16,7 +16,11 @@
 # under the License.
 
 add_arrow_test(concatenate_test)
-add_arrow_test(diff_test)
+
+if(ARROW_COMPUTE)
+  # This unit test uses compute code
+  add_arrow_test(diff_test)
+endif()
 
 # Headers: top level
 arrow_install_all_headers("arrow/array")

dev/release/verify-release-candidate.bat

@@ -85,6 +85,11 @@ cmake -G "Ninja" ^
       -DARROW_WITH_LZ4=ON ^
       -DARROW_WITH_SNAPPY=ON ^
       -DARROW_WITH_BROTLI=ON ^
+      -DARROW_COMPUTE=ON ^
+      -DARROW_CSV=ON ^
+      -DARROW_DATASET=ON ^
+      -DARROW_FILESYSTEM=ON ^
+      -DARROW_JSON=ON ^
       -DARROW_FLIGHT=ON ^
       -DARROW_PYTHON=ON ^
       -DARROW_PARQUET=ON ^

dev/release/verify-release-candidate.sh

@@ -220,7 +220,12 @@ test_and_install_cpp() {
 ${ARROW_CMAKE_OPTIONS:-}
 -DCMAKE_INSTALL_PREFIX=$ARROW_HOME
 -DCMAKE_INSTALL_LIBDIR=lib
+-DARROW_COMPUTE=ON
+-DARROW_CSV=ON
+-DARROW_DATASET=ON
+-DARROW_FILESYSTEM=ON
 -DARROW_FLIGHT=${ARROW_FLIGHT}
+-DARROW_JSON=ON
 -DARROW_PLASMA=ON
 -DARROW_ORC=ON
 -DARROW_PYTHON=ON

dev/tasks/conda-recipes/arrow-cpp/bld.bat

@@ -18,6 +18,11 @@ cmake -G "Ninja" ^
       -DARROW_BUILD_UTILITIES:BOOL=OFF ^
       -DARROW_BUILD_STATIC:BOOL=OFF ^
       -DCMAKE_BUILD_TYPE=release ^
+      -DARROW_COMPUTE=ON ^
+      -DARROW_CSV=ON ^
+      -DARROW_DATASET=ON ^
+      -DARROW_FILESYSTEM=ON ^
+      -DARROW_JSON=ON ^
       -DARROW_PYTHON:BOOL=ON ^
       -DARROW_MIMALLOC:BOOL=ON ^
       -DARROW_FLIGHT:BOOL=ON ^

dev/tasks/conda-recipes/arrow-cpp/build.sh

@@ -32,7 +32,12 @@ cmake \
     -DARROW_WITH_LZ4=ON \
     -DARROW_WITH_SNAPPY=ON \
     -DARROW_WITH_BROTLI=ON \
+    -DARROW_COMPUTE=ON \
+    -DARROW_CSV=ON \
+    -DARROW_DATASET=ON \
+    -DARROW_FILESYSTEM=ON \
     -DARROW_JEMALLOC=ON \
+    -DARROW_JSON=ON \
     -DARROW_MIMALLOC=ON \
     -DARROW_FLIGHT=ON \
     -DARROW_PLASMA=ON \

dev/tasks/homebrew-formulae/apache-arrow.rb

@@ -24,7 +24,12 @@ class ApacheArrow < Formula
   def install
     ENV.cxx11
     args = %W[
+      -DARROW_COMPUTE=ON
+      -DARROW_CSV=ON
+      -DARROW_DATASET=ON
+      -DARROW_FILESYSTEM=ON
       -DARROW_FLIGHT=ON
+      -DARROW_JSON=ON
       -DARROW_ORC=ON
       -DARROW_PARQUET=ON
       -DARROW_PLASMA=ON

dev/tasks/python-wheels/osx-build.sh

@@ -125,7 +125,13 @@ function build_wheel {
           -DARROW_DEPENDENCY_SOURCE=BUNDLED \
           -DARROW_FLIGHT=ON \
           -DARROW_GANDIVA=${BUILD_ARROW_GANDIVA} \
+          -DARROW_BOOST_USE_SHARED=ON \
+          -DARROW_COMPUTE=ON \
+          -DARROW_CSV=ON \
+          -DARROW_DATASET=ON \
+          -DARROW_FILESYSTEM=ON \
           -DARROW_JEMALLOC=ON \
+          -DARROW_JSON=ON \
           -DARROW_ORC=OFF \
           -DARROW_PARQUET=ON \
           -DARROW_PLASMA=ON \

dev/tasks/python-wheels/win-build.bat

@@ -51,6 +51,11 @@ cmake -G "%GENERATOR%" ^
       -DARROW_BOOST_USE_SHARED=OFF ^
       -DARROW_BUILD_TESTS=OFF ^
       -DCMAKE_BUILD_TYPE=Release ^
+      -DARROW_COMPUTE=ON ^
+      -DARROW_CSV=ON ^
+      -DARROW_DATASET=ON ^
+      -DARROW_FILESYSTEM=ON ^
+      -DARROW_JSON=ON ^
       -DARROW_DEPENDENCY_SOURCE=CONDA ^
       -DOPENSSL_ROOT_DIR=%CONDA_PREFIX%/Library ^
       -DARROW_CXXFLAGS="/MP" ^

python/manylinux1/build_arrow.sh

@@ -93,8 +93,13 @@ cmake -DCMAKE_BUILD_TYPE=Release \
     -DARROW_BUILD_TESTS=OFF \
     -DARROW_BUILD_SHARED=ON \
     -DARROW_BOOST_USE_SHARED=ON \
+    -DARROW_COMPUTE=ON \
+    -DARROW_CSV=ON \
+    -DARROW_DATASET=ON \
+    -DARROW_FILESYSTEM=ON \
     -DARROW_GANDIVA_PC_CXX_FLAGS="-isystem;/opt/rh/devtoolset-2/root/usr/include/c++/4.8.2;-isystem;/opt/rh/devtoolset-2/root/usr/include/c++/4.8.2/x86_64-CentOS-linux/" \
     -DARROW_JEMALLOC=ON \
+    -DARROW_JSON=ON \
     -DARROW_RPATH_ORIGIN=ON \
     -DARROW_PYTHON=ON \
     -DARROW_PARQUET=ON \

python/manylinux2010/build_arrow.sh

@@ -94,8 +94,13 @@ PATH="${CPYTHON_PATH}/bin:${PATH}" cmake -DCMAKE_BUILD_TYPE=Release \
     -DARROW_BUILD_SHARED=ON \
     -DARROW_BUILD_STATIC=OFF \
     -DARROW_BOOST_USE_SHARED=ON \
+    -DARROW_COMPUTE=ON \
+    -DARROW_CSV=ON \
+    -DARROW_DATASET=ON \
+    -DARROW_FILESYSTEM=ON \
     -DARROW_GANDIVA_PC_CXX_FLAGS="-isystem;/opt/rh/devtoolset-8/root/usr/include/c++/8/;-isystem;/opt/rh/devtoolset-8/root/usr/include/c++/8/x86_64-redhat-linux/" \
     -DARROW_JEMALLOC=ON \
+    -DARROW_JSON=ON \
     -DARROW_RPATH_ORIGIN=ON \
     -DARROW_PYTHON=ON \
     -DARROW_PARQUET=ON \