Merge branch 'apache:main' into count-Wildcard

Author: jiangzhx

Diff for: Cargo.toml

@@ -31,10 +31,25 @@ members = [
     "datafusion/substrait",
     "datafusion-examples",
     "test-utils",
-    "parquet-test-utils",
     "benchmarks",
 ]
 
+[workspace.package]
+version = "20.0.0"
+edition = "2021"
+readme = "README.md"
+authors = ["Apache Arrow <[email protected]>"]
+license = "Apache-2.0"
+homepage = "https://github.com/apache/arrow-datafusion"
+repository = "https://github.com/apache/arrow-datafusion"
+rust-version = "1.64"
+
+[workspace.dependencies]
+arrow = { version = "34.0.0", features = ["prettyprint"] }
+arrow-buffer = "34.0.0"
+arrow-schema = "34.0.0"
+parquet = { version = "34.0.0", features = ["arrow", "async"] }
+
 [profile.release]
 codegen-units = 1
 lto = true

Diff for: benchmarks/Cargo.toml

@@ -18,7 +18,7 @@
 [package]
 name = "datafusion-benchmarks"
 description = "DataFusion Benchmarks"
-version = "19.0.0"
+version = "20.0.0"
 edition = "2021"
 authors = ["Apache Arrow <[email protected]>"]
 homepage = "https://github.com/apache/arrow-datafusion"
@@ -33,16 +33,13 @@ simd = ["datafusion/simd"]
 snmalloc = ["snmalloc-rs"]
 
 [dependencies]
-arrow = "34.0.0"
-datafusion = { path = "../datafusion/core", version = "19.0.0", features = ["scheduler"] }
+arrow = { workspace = true }
+datafusion = { path = "../datafusion/core", version = "20.0.0", features = ["scheduler"] }
 env_logger = "0.10"
 futures = "0.3"
 mimalloc = { version = "0.1", optional = true, default-features = false }
 num_cpus = "1.13.0"
-object_store = "0.5.4"
-parquet = "34.0.0"
-parquet-test-utils = { path = "../parquet-test-utils/", version = "0.1.0" }
-rand = "0.8.4"
+parquet = { workspace = true }
 serde = { version = "1.0.136", features = ["derive"] }
 serde_json = "1.0.78"
 snmalloc-rs = { version = "0.3", optional = true }
@@ -51,4 +48,4 @@ test-utils = { path = "../test-utils/", version = "0.1.0" }
 tokio = { version = "^1.0", features = ["macros", "rt", "rt-multi-thread", "parking_lot"] }
 
 [dev-dependencies]
-datafusion-proto = { path = "../datafusion/proto", version = "19.0.0" }
+datafusion-proto = { path = "../datafusion/proto", version = "20.0.0" }

Diff for: benchmarks/db-benchmark/README.md

@@ -19,6 +19,8 @@
 
 # Run db-benchmark
 
+This directory contains scripts for running DataFusion with the https://github.com/h2oai/db-benchmark
+
 ## Directions
 
 Run the following from root `arrow-datafusion` directory

Diff for: benchmarks/src/bin/parquet.rs

@@ -23,8 +23,8 @@ use datafusion::physical_expr::PhysicalSortExpr;
 use datafusion::physical_plan::collect;
 use datafusion::physical_plan::sorts::sort::SortExec;
 use datafusion::prelude::{col, SessionConfig, SessionContext};
+use datafusion::test_util::parquet::{ParquetScanOptions, TestParquetFile};
 use parquet::file::properties::WriterProperties;
-use parquet_test_utils::{ParquetScanOptions, TestParquetFile};
 use std::path::PathBuf;
 use std::sync::Arc;
 use std::time::Instant;
@@ -75,7 +75,7 @@ struct Opt {
 }
 impl Opt {
     /// Initialize parquet test file given options.
-    fn init_file(&self) -> Result<parquet_test_utils::TestParquetFile> {
+    fn init_file(&self) -> Result<TestParquetFile> {
         let path = self.path.join("logs.parquet");
 
         let mut props_builder = WriterProperties::builder();