Upgrade to arrow 20.0.0 (but no change to object_store), including prost, and tonic (apache#3083)

* Upgrade arrow

fix decimal (#4)

Fix human error

Patch crates io to fix build (#5)

* fix decimal

* patch crate versions

Patch objectstore

Test in CI

Undo override?

Fix more errors

Fix last error?

Formatting

Clippy

Fixes

Fix refs

Able to get session context, but JDBC driver hung

Upgrade to arrow 20

Upgrade to RC2

Formatting

Fix some imports

Install protoc

Try platform agnostic path

Debug in CI :(

Debug in CI :(

Debug in CI :(

Not worth it, just separate builds

Variables

Fixes

Fix windows?

Fix windows?

Hackily fix windows

Down to 1 failure

Fix protoc

All? tests pass

Formatting

* Fix remaining tests

* Clippy

* Update docs for Windows

* Try with old objectstore

* Revert path "fixes" that broke windows

* Update to arrow 20

Author: Brent Gardner

.github/workflows/rust.yml

@@ -82,6 +82,16 @@ jobs:
       - uses: actions/checkout@v2
         with:
           submodules: true
+      - name: Install protobuf compiler
+        shell: bash
+        run: |
+          mkdir -p $HOME/d/protoc
+          cd $HOME/d/protoc
+          export PROTO_ZIP="protoc-21.4-linux-x86_64.zip"
+          curl -LO https://github.com/protocolbuffers/protobuf/releases/download/v21.4/$PROTO_ZIP
+          unzip $PROTO_ZIP
+          export PATH=$PATH:$HOME/d/protoc/bin
+          protoc --version
       - name: Cache Cargo
         uses: actions/cache@v3
         with:
@@ -94,6 +104,7 @@ jobs:
           rust-version: ${{ matrix.rust }}
       - name: Run tests
         run: |
+          export PATH=$PATH:$HOME/d/protoc/bin
           cargo test --features avro,jit,scheduler,json
           # test datafusion-sql examples
           cargo run --example sql
@@ -159,17 +170,65 @@ jobs:
           POSTGRES_USER: postgres
           POSTGRES_PASSWORD: postgres
 
-  windows-and-macos:
-    name: Test on ${{ matrix.os }} Rust ${{ matrix.rust }}
+  windows:
+    name: Test on Windows Rust ${{ matrix.rust }}
     runs-on: ${{ matrix.os }}
     strategy:
       matrix:
-        os: [windows-latest, macos-latest]
+        os: [windows-latest]
         rust: [stable]
     steps:
       - uses: actions/checkout@v2
         with:
           submodules: true
+      - name: Install protobuf compiler
+        shell: bash
+        run: |
+          mkdir -p $HOME/d/protoc
+          cd $HOME/d/protoc
+          export PROTO_ZIP="protoc-21.4-win64.zip"
+          curl -LO https://github.com/protocolbuffers/protobuf/releases/download/v21.4/$PROTO_ZIP
+          unzip $PROTO_ZIP
+          export PATH=$PATH:$HOME/d/protoc/bin
+          protoc.exe --version
+      # TODO: this won't cache anything, which is expensive. Setup this action
+      # with a OS-dependent path.
+      - name: Setup Rust toolchain
+        run: |
+          rustup toolchain install ${{ matrix.rust }}
+          rustup default ${{ matrix.rust }}
+          rustup component add rustfmt
+      - name: Run tests
+        shell: bash
+        run: |
+          export PATH=$PATH:$HOME/d/protoc/bin
+          cargo test
+        env:
+          # do not produce debug symbols to keep memory usage down
+          RUSTFLAGS: "-C debuginfo=0"
+
+  macos:
+    name: Test on MacOS Rust ${{ matrix.rust }}
+    runs-on: ${{ matrix.os }}
+    strategy:
+      matrix:
+        os: [macos-latest]
+        rust: [stable]
+    steps:
+      - uses: actions/checkout@v2
+        with:
+          submodules: true
+      - name: Install protobuf compiler
+        shell: bash
+        run: |
+          mkdir -p $HOME/d/protoc
+          cd $HOME/d/protoc
+          export PROTO_ZIP="protoc-21.4-osx-x86_64.zip"
+          curl -LO https://github.com/protocolbuffers/protobuf/releases/download/v21.4/$PROTO_ZIP
+          unzip $PROTO_ZIP
+          echo "$HOME/d/protoc/bin" >> $GITHUB_PATH
+          export PATH=$PATH:$HOME/d/protoc/bin
+          protoc --version
       # TODO: this won't cache anything, which is expensive. Setup this action
       # with a OS-dependent path.
       - name: Setup Rust toolchain
@@ -250,6 +309,16 @@ jobs:
       - uses: actions/checkout@v2
         with:
           submodules: true
+      - name: Install protobuf compiler
+        shell: bash
+        run: |
+          mkdir -p $HOME/d/protoc
+          cd $HOME/d/protoc
+          export PROTO_ZIP="protoc-21.4-linux-x86_64.zip"
+          curl -LO https://github.com/protocolbuffers/protobuf/releases/download/v21.4/$PROTO_ZIP
+          unzip $PROTO_ZIP
+          export PATH=$PATH:$HOME/d/protoc/bin
+          protoc --version
       - name: Setup Rust toolchain
         run: |
           rustup toolchain install ${{ matrix.rust }}
@@ -263,6 +332,7 @@ jobs:
           key: cargo-coverage-cache3-
       - name: Run coverage
         run: |
+          export PATH=$PATH:$HOME/d/protoc/bin
           rustup toolchain install stable
           rustup default stable
           cargo install --version 0.20.1 cargo-tarpaulin

CONTRIBUTING.md

@@ -35,6 +35,15 @@ list to help you get started.
 
 This section describes how you can get started at developing DataFusion.
 
+### Windows setup
+
+```shell
+wget https://az792536.vo.msecnd.net/vms/VMBuild_20190311/VirtualBox/MSEdge/MSEdge.Win10.VirtualBox.zip
+choco install -y git rustup.install visualcpp-build-tools
+git-bash.exe
+cargo build
+```
+
 ### Bootstrap environment
 
 DataFusion is written in Rust and it uses a standard rust toolkit:

datafusion-cli/Cargo.toml

@@ -29,7 +29,7 @@ rust-version = "1.59"
 readme = "README.md"
 
 [dependencies]
-arrow = { version = "19.0.0" }
+arrow = { version = "20.0.0" }
 clap = { version = "3", features = ["derive", "cargo"] }
 datafusion = { path = "../datafusion/core", version = "10.0.0" }
 dirs = "4.0.0"

datafusion-examples/Cargo.toml

@@ -34,13 +34,13 @@ path = "examples/avro_sql.rs"
 required-features = ["datafusion/avro"]
 
 [dev-dependencies]
-arrow-flight = { version = "19.0.0" }
+arrow-flight = { version = "20.0.0" }
 async-trait = "0.1.41"
 datafusion = { path = "../datafusion/core" }
 futures = "0.3"
 num_cpus = "1.13.0"
-prost = "0.10"
+prost = "0.11.0"
 serde = { version = "1.0.136", features = ["derive"] }
 serde_json = "1.0.82"
 tokio = { version = "1.0", features = ["macros", "rt", "rt-multi-thread", "sync", "parking_lot"] }
-tonic = "0.7"
+tonic = "0.8"

datafusion/common/Cargo.toml

@@ -39,11 +39,12 @@ pyarrow = ["pyo3"]
 
 [dependencies]
 apache-avro = { version = "0.14", features = ["snappy"], optional = true }
-arrow = { version = "19.0.0", features = ["prettyprint"] }
+arrow = { version = "20.0.0", features = ["prettyprint"] }
+avro-rs = { version = "0.13", features = ["snappy"], optional = true }
 cranelift-module = { version = "0.86.1", optional = true }
 object_store = { version = "0.3", optional = true }
 ordered-float = "3.0"
-parquet = { version = "19.0.0", features = ["arrow"], optional = true }
+parquet = { version = "20.0.0", features = ["arrow"], optional = true }
 pyo3 = { version = "0.16", optional = true }
 serde_json = "1.0"
 sqlparser = "0.20"

datafusion/common/src/from_slice.rs

@@ -69,7 +69,7 @@ where
             offsets.push(length_so_far);
             values.extend_from_slice(s);
         }
-        let array_data = ArrayData::builder(Self::get_data_type())
+        let array_data = ArrayData::builder(Self::DATA_TYPE)
             .len(slice.len())
             .add_buffer(Buffer::from_slice_ref(&offsets))
             .add_buffer(Buffer::from_slice_ref(&values));

datafusion/common/src/scalar.rs

@@ -27,7 +27,7 @@ use arrow::{
         IntervalMonthDayNanoType, IntervalUnit, IntervalYearMonthType, TimeUnit,
         TimestampMicrosecondType, TimestampMillisecondType, TimestampNanosecondType,
         TimestampSecondType, UInt16Type, UInt32Type, UInt64Type, UInt8Type,
-        DECIMAL_MAX_PRECISION,
+        DECIMAL128_MAX_PRECISION,
     },
     util::decimal::{BasicDecimal, Decimal128},
 };
@@ -611,7 +611,7 @@ impl ScalarValue {
         scale: usize,
     ) -> Result<Self> {
         // make sure the precision and scale is valid
-        if precision <= DECIMAL_MAX_PRECISION && scale <= precision {
+        if precision <= DECIMAL128_MAX_PRECISION && scale <= precision {
             return Ok(ScalarValue::Decimal128(Some(value), precision, scale));
         }
         Err(DataFusionError::Internal(format!(
@@ -654,7 +654,7 @@ impl ScalarValue {
             ScalarValue::Int32(_) => DataType::Int32,
             ScalarValue::Int64(_) => DataType::Int64,
             ScalarValue::Decimal128(_, precision, scale) => {
-                DataType::Decimal(*precision, *scale)
+                DataType::Decimal128(*precision, *scale)
             }
             ScalarValue::TimestampSecond(_, tz_opt) => {
                 DataType::Timestamp(TimeUnit::Second, tz_opt.clone())
@@ -935,7 +935,7 @@ impl ScalarValue {
         }
 
         let array: ArrayRef = match &data_type {
-            DataType::Decimal(precision, scale) => {
+            DataType::Decimal128(precision, scale) => {
                 let decimal_array =
                     ScalarValue::iter_to_decimal_array(scalars, precision, scale)?;
                 Arc::new(decimal_array)
@@ -1448,7 +1448,7 @@ impl ScalarValue {
 
         Ok(match array.data_type() {
             DataType::Null => ScalarValue::Null,
-            DataType::Decimal(precision, scale) => {
+            DataType::Decimal128(precision, scale) => {
                 ScalarValue::get_decimal_value_from_array(array, index, precision, scale)
             }
             DataType::Boolean => typed_cast!(array, index, BooleanArray, Boolean),
@@ -1899,7 +1899,7 @@ impl TryFrom<&DataType> for ScalarValue {
             DataType::UInt16 => ScalarValue::UInt16(None),
             DataType::UInt32 => ScalarValue::UInt32(None),
             DataType::UInt64 => ScalarValue::UInt64(None),
-            DataType::Decimal(precision, scale) => {
+            DataType::Decimal128(precision, scale) => {
                 ScalarValue::Decimal128(None, *precision, *scale)
             }
             DataType::Utf8 => ScalarValue::Utf8(None),
@@ -2145,7 +2145,7 @@ mod tests {
     #[test]
     fn scalar_decimal_test() {
         let decimal_value = ScalarValue::Decimal128(Some(123), 10, 1);
-        assert_eq!(DataType::Decimal(10, 1), decimal_value.get_datatype());
+        assert_eq!(DataType::Decimal128(10, 1), decimal_value.get_datatype());
         let try_into_value: i128 = decimal_value.clone().try_into().unwrap();
         assert_eq!(123_i128, try_into_value);
         assert!(!decimal_value.is_null());
@@ -2163,14 +2163,14 @@ mod tests {
         let array = decimal_value.to_array();
         let array = array.as_any().downcast_ref::<Decimal128Array>().unwrap();
         assert_eq!(1, array.len());
-        assert_eq!(DataType::Decimal(10, 1), array.data_type().clone());
+        assert_eq!(DataType::Decimal128(10, 1), array.data_type().clone());
         assert_eq!(123i128, array.value(0).as_i128());
 
         // decimal scalar to array with size
         let array = decimal_value.to_array_of_size(10);
         let array_decimal = array.as_any().downcast_ref::<Decimal128Array>().unwrap();
         assert_eq!(10, array.len());
-        assert_eq!(DataType::Decimal(10, 1), array.data_type().clone());
+        assert_eq!(DataType::Decimal128(10, 1), array.data_type().clone());
         assert_eq!(123i128, array_decimal.value(0).as_i128());
         assert_eq!(123i128, array_decimal.value(9).as_i128());
         // test eq array
@@ -2208,7 +2208,7 @@ mod tests {
         // convert the vec to decimal array and check the result
         let array = ScalarValue::iter_to_array(decimal_vec.into_iter()).unwrap();
         assert_eq!(3, array.len());
-        assert_eq!(DataType::Decimal(10, 2), array.data_type().clone());
+        assert_eq!(DataType::Decimal128(10, 2), array.data_type().clone());
 
         let decimal_vec = vec![
             ScalarValue::Decimal128(Some(1), 10, 2),
@@ -2218,7 +2218,7 @@ mod tests {
         ];
         let array = ScalarValue::iter_to_array(decimal_vec.into_iter()).unwrap();
         assert_eq!(4, array.len());
-        assert_eq!(DataType::Decimal(10, 2), array.data_type().clone());
+        assert_eq!(DataType::Decimal128(10, 2), array.data_type().clone());
 
         assert!(ScalarValue::try_new_decimal128(1, 10, 2)
             .unwrap()

datafusion/core/Cargo.toml

@@ -56,7 +56,7 @@ unicode_expressions = ["datafusion-physical-expr/regex_expressions", "datafusion
 [dependencies]
 ahash = { version = "0.7", default-features = false }
 apache-avro = { version = "0.14", optional = true }
-arrow = { version = "19.0.0", features = ["prettyprint"] }
+arrow = { version = "20.0.0", features = ["prettyprint"] }
 async-trait = "0.1.41"
 bytes = "1.1"
 chrono = { version = "0.4", default-features = false }
@@ -78,7 +78,7 @@ num_cpus = "1.13.0"
 object_store = "0.3.0"
 ordered-float = "3.0"
 parking_lot = "0.12"
-parquet = { version = "19.0.0", features = ["arrow", "async"] }
+parquet = { version = "20.0.0", features = ["arrow", "async"] }
 paste = "^1.0"
 pin-project-lite = "^0.2.7"
 pyo3 = { version = "0.16", optional = true }

datafusion/core/fuzz-utils/Cargo.toml

@@ -23,6 +23,6 @@ edition = "2021"
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
 [dependencies]
-arrow = { version = "19.0.0", features = ["prettyprint"] }
+arrow = { version = "20.0.0", features = ["prettyprint"] }
 env_logger = "0.9.0"
 rand = "0.8"

datafusion/core/src/avro_to_arrow/arrow_array_reader.rs

@@ -101,12 +101,10 @@ impl<'a, R: Read> AvroArrowArrayReader<'a, R> {
                     "Failed to parse avro value: {:?}",
                     e
                 ))),
-                other => {
-                    return Err(ArrowError::ParseError(format!(
-                        "Row needs to be of type object, got: {:?}",
-                        other
-                    )))
-                }
+                other => Err(ArrowError::ParseError(format!(
+                    "Row needs to be of type object, got: {:?}",
+                    other
+                ))),
             })
             .collect::<ArrowResult<Vec<Vec<(String, Value)>>>>()?;
         if rows.is_empty() {