This is an automated email from the ASF dual-hosted git repository.
dheres pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/arrow-datafusion.git
The following commit(s) were added to refs/heads/main by this push:
new e0961d5547 Replace lazy_static with OnceLock (#7409)
e0961d5547 is described below
commit e0961d55479aab1c4f92eca817fcce4ec25d7c3e
Author: Kousuke Saruta <saru...@apache.org>
AuthorDate: Sat Aug 26 05:46:11 2023 +0900
Replace lazy_static with OnceLock (#7409)
Co-authored-by: Andrew Lamb <and...@nerdnetworks.org>
---
datafusion-cli/Cargo.lock | 3 --
datafusion/core/Cargo.toml | 1 -
datafusion/expr/Cargo.toml | 1 -
datafusion/expr/src/built_in_function.rs | 42 ++++++++++--------
datafusion/physical-expr/Cargo.toml | 1 -
datafusion/physical-expr/src/regex_expressions.rs | 10 ++---
datafusion/sqllogictest/Cargo.toml | 1 -
.../src/engines/datafusion_engine/normalize.rs | 51 +++++++++++-----------
8 files changed, 54 insertions(+), 56 deletions(-)
diff --git a/datafusion-cli/Cargo.lock b/datafusion-cli/Cargo.lock
index 04e802e4ed..3a2f9a7b73 100644
--- a/datafusion-cli/Cargo.lock
+++ b/datafusion-cli/Cargo.lock
@@ -1074,7 +1074,6 @@ dependencies = [
"hashbrown 0.14.0",
"indexmap 2.0.0",
"itertools 0.11.0",
- "lazy_static",
"log",
"num_cpus",
"object_store",
@@ -1164,7 +1163,6 @@ dependencies = [
"ahash",
"arrow",
"datafusion-common",
- "lazy_static",
"sqlparser",
"strum 0.25.0",
"strum_macros 0.25.2",
@@ -1206,7 +1204,6 @@ dependencies = [
"hex",
"indexmap 2.0.0",
"itertools 0.11.0",
- "lazy_static",
"libc",
"log",
"md-5",
diff --git a/datafusion/core/Cargo.toml b/datafusion/core/Cargo.toml
index b28ab05cb1..bfa00b2a9f 100644
--- a/datafusion/core/Cargo.toml
+++ b/datafusion/core/Cargo.toml
@@ -75,7 +75,6 @@ half = { version = "2.1", default-features = false }
hashbrown = { version = "0.14", features = ["raw"] }
indexmap = "2.0.0"
itertools = "0.11"
-lazy_static = { version = "^1.4.0" }
log = "^0.4"
num-traits = { version = "0.2", optional = true }
num_cpus = "1.13.0"
diff --git a/datafusion/expr/Cargo.toml b/datafusion/expr/Cargo.toml
index 9634f23860..e46b66361d 100644
--- a/datafusion/expr/Cargo.toml
+++ b/datafusion/expr/Cargo.toml
@@ -38,7 +38,6 @@ path = "src/lib.rs"
ahash = { version = "0.8", default-features = false, features =
["runtime-rng"] }
arrow = { workspace = true }
datafusion-common = { path = "../common", version = "30.0.0" }
-lazy_static = { version = "^1.4.0" }
sqlparser = { workspace = true }
strum = { version = "0.25.0", features = ["derive"] }
strum_macros = "0.25.0"
diff --git a/datafusion/expr/src/built_in_function.rs
b/datafusion/expr/src/built_in_function.rs
index de046cde89..0fb23be5af 100644
--- a/datafusion/expr/src/built_in_function.rs
+++ b/datafusion/expr/src/built_in_function.rs
@@ -28,12 +28,10 @@ use datafusion_common::{internal_err, plan_err,
DataFusionError, Result};
use std::collections::HashMap;
use std::fmt;
use std::str::FromStr;
-use std::sync::Arc;
+use std::sync::{Arc, OnceLock};
use strum::IntoEnumIterator;
use strum_macros::EnumIter;
-use lazy_static::lazy_static;
-
/// Enum of all built-in scalar functions
// Contributor's guide for adding new scalar functions
//
https://arrow.apache.org/datafusion/contributor-guide/index.html#how-to-add-a-new-scalar-function
@@ -282,26 +280,34 @@ pub enum BuiltinScalarFunction {
ArrowTypeof,
}
-lazy_static! {
- /// Maps the sql function name to `BuiltinScalarFunction`
- static ref NAME_TO_FUNCTION: HashMap<&'static str, BuiltinScalarFunction>
= {
- let mut map: HashMap<&'static str, BuiltinScalarFunction> =
HashMap::new();
+/// Maps the sql function name to `BuiltinScalarFunction`
+fn name_to_function() -> &'static HashMap<&'static str, BuiltinScalarFunction>
{
+ static NAME_TO_FUNCTION_LOCK: OnceLock<HashMap<&'static str,
BuiltinScalarFunction>> =
+ OnceLock::new();
+ NAME_TO_FUNCTION_LOCK.get_or_init(|| {
+ let mut map = HashMap::new();
BuiltinScalarFunction::iter().for_each(|func| {
let a = aliases(&func);
- a.iter().for_each(|a| {map.insert(a, func);});
+ a.iter().for_each(|&a| {
+ map.insert(a, func);
+ });
});
map
- };
+ })
+}
- /// Maps `BuiltinScalarFunction` --> canonical sql function
- /// First alias in the array is used to display function names
- static ref FUNCTION_TO_NAME: HashMap<BuiltinScalarFunction, &'static str>
= {
- let mut map: HashMap<BuiltinScalarFunction, &'static str> =
HashMap::new();
+/// Maps `BuiltinScalarFunction` --> canonical sql function
+/// First alias in the array is used to display function names
+fn function_to_name() -> &'static HashMap<BuiltinScalarFunction, &'static str>
{
+ static FUNCTION_TO_NAME_LOCK: OnceLock<HashMap<BuiltinScalarFunction,
&'static str>> =
+ OnceLock::new();
+ FUNCTION_TO_NAME_LOCK.get_or_init(|| {
+ let mut map = HashMap::new();
BuiltinScalarFunction::iter().for_each(|func| {
- map.insert(func, aliases(&func).first().unwrap_or(&"NO_ALIAS"));
+ map.insert(func, *aliases(&func).first().unwrap_or(&"NO_ALIAS"));
});
map
- };
+ })
}
impl BuiltinScalarFunction {
@@ -1379,14 +1385,14 @@ fn aliases(func: &BuiltinScalarFunction) -> &'static
[&'static str] {
impl fmt::Display for BuiltinScalarFunction {
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
// .unwrap is safe here because compiler makes sure the map will have
matches for each BuiltinScalarFunction
- write!(f, "{}", FUNCTION_TO_NAME.get(self).unwrap())
+ write!(f, "{}", function_to_name().get(self).unwrap())
}
}
impl FromStr for BuiltinScalarFunction {
type Err = DataFusionError;
fn from_str(name: &str) -> Result<BuiltinScalarFunction> {
- if let Some(func) = NAME_TO_FUNCTION.get(name) {
+ if let Some(func) = name_to_function().get(name) {
Ok(*func)
} else {
plan_err!("There is no built-in function named {name}")
@@ -1453,7 +1459,7 @@ mod tests {
// and then back to a variant. The test asserts that the original variant
and
// the reconstructed variant are the same.
fn test_display_and_from_str() {
- for (_, func_original) in NAME_TO_FUNCTION.iter() {
+ for (_, func_original) in name_to_function().iter() {
let func_name = func_original.to_string();
let func_from_str =
BuiltinScalarFunction::from_str(&func_name).unwrap();
assert_eq!(func_from_str, *func_original);
diff --git a/datafusion/physical-expr/Cargo.toml
b/datafusion/physical-expr/Cargo.toml
index d529b6762e..dbaacb4dba 100644
--- a/datafusion/physical-expr/Cargo.toml
+++ b/datafusion/physical-expr/Cargo.toml
@@ -59,7 +59,6 @@ hashbrown = { version = "0.14", features = ["raw"] }
hex = { version = "0.4", optional = true }
indexmap = "2.0.0"
itertools = { version = "0.11", features = ["use_std"] }
-lazy_static = { version = "^1.4.0" }
libc = "0.2.140"
log = "^0.4"
md-5 = { version = "^0.10.0", optional = true }
diff --git a/datafusion/physical-expr/src/regex_expressions.rs
b/datafusion/physical-expr/src/regex_expressions.rs
index f213336602..5aea70f75a 100644
--- a/datafusion/physical-expr/src/regex_expressions.rs
+++ b/datafusion/physical-expr/src/regex_expressions.rs
@@ -32,9 +32,8 @@ use datafusion_common::{
};
use datafusion_expr::{ColumnarValue, ScalarFunctionImplementation};
use hashbrown::HashMap;
-use lazy_static::lazy_static;
use regex::Regex;
-use std::sync::Arc;
+use std::sync::{Arc, OnceLock};
use crate::functions::{make_scalar_function, make_scalar_function_with_hints,
Hint};
@@ -82,10 +81,11 @@ pub fn regexp_match<T: OffsetSizeTrait>(args: &[ArrayRef])
-> Result<ArrayRef> {
/// replace POSIX capture groups (like \1) with Rust Regex group (like ${1})
/// used by regexp_replace
fn regex_replace_posix_groups(replacement: &str) -> String {
- lazy_static! {
- static ref CAPTURE_GROUPS_RE: Regex =
Regex::new(r"(\\)(\d*)").unwrap();
+ fn capture_groups_re() -> &'static Regex {
+ static CAPTURE_GROUPS_RE_LOCK: OnceLock<Regex> = OnceLock::new();
+ CAPTURE_GROUPS_RE_LOCK.get_or_init(||
Regex::new(r"(\\)(\d*)").unwrap())
}
- CAPTURE_GROUPS_RE
+ capture_groups_re()
.replace_all(replacement, "$${$2}")
.into_owned()
}
diff --git a/datafusion/sqllogictest/Cargo.toml
b/datafusion/sqllogictest/Cargo.toml
index 15a23a92dd..0b4a0ef419 100644
--- a/datafusion/sqllogictest/Cargo.toml
+++ b/datafusion/sqllogictest/Cargo.toml
@@ -38,7 +38,6 @@ datafusion = {path = "../core", version = "30.0.0"}
datafusion-common = {path = "../common", version = "30.0.0"}
half = "2.2.1"
itertools = "0.11"
-lazy_static = {version = "^1.4.0"}
object_store = "0.7.0"
rust_decimal = {version = "1.27.0"}
log = "^0.4"
diff --git a/datafusion/sqllogictest/src/engines/datafusion_engine/normalize.rs
b/datafusion/sqllogictest/src/engines/datafusion_engine/normalize.rs
index 942ba8e05b..c0db111bc6 100644
--- a/datafusion/sqllogictest/src/engines/datafusion_engine/normalize.rs
+++ b/datafusion/sqllogictest/src/engines/datafusion_engine/normalize.rs
@@ -20,8 +20,8 @@ use arrow::{array, array::ArrayRef, datatypes::DataType,
record_batch::RecordBat
use datafusion_common::format::DEFAULT_FORMAT_OPTIONS;
use datafusion_common::DFField;
use datafusion_common::DataFusionError;
-use lazy_static::lazy_static;
use std::path::PathBuf;
+use std::sync::OnceLock;
use crate::engines::output::DFColumnType;
@@ -126,7 +126,7 @@ fn expand_row(mut row: Vec<String>) -> impl Iterator<Item =
Vec<String>> {
/// ```
fn normalize_paths(mut row: Vec<String>) -> Vec<String> {
row.iter_mut().for_each(|s| {
- let workspace_root: &str = WORKSPACE_ROOT.as_ref();
+ let workspace_root: &str = workspace_root().as_ref();
if s.contains(workspace_root) {
*s = s.replace(workspace_root, "WORKSPACE_ROOT");
}
@@ -135,33 +135,32 @@ fn normalize_paths(mut row: Vec<String>) -> Vec<String> {
}
/// return the location of the datafusion checkout
-fn workspace_root() -> object_store::path::Path {
- // e.g. /Software/arrow-datafusion/datafusion/core
- let dir = PathBuf::from(env!("CARGO_MANIFEST_DIR"));
+fn workspace_root() -> &'static object_store::path::Path {
+ static WORKSPACE_ROOT_LOCK: OnceLock<object_store::path::Path> =
OnceLock::new();
+ WORKSPACE_ROOT_LOCK.get_or_init(|| {
+ // e.g. /Software/arrow-datafusion/datafusion/core
+ let dir = PathBuf::from(env!("CARGO_MANIFEST_DIR"));
- // e.g. /Software/arrow-datafusion/datafusion
- let workspace_root = dir
- .parent()
- .expect("Can not find parent of datafusion/core")
- // e.g. /Software/arrow-datafusion
- .parent()
- .expect("parent of datafusion")
- .to_string_lossy();
+ // e.g. /Software/arrow-datafusion/datafusion
+ let workspace_root = dir
+ .parent()
+ .expect("Can not find parent of datafusion/core")
+ // e.g. /Software/arrow-datafusion
+ .parent()
+ .expect("parent of datafusion")
+ .to_string_lossy();
- let sanitized_workplace_root = if cfg!(windows) {
- // Object store paths are delimited with `/`, e.g.
`D:/a/arrow-datafusion/arrow-datafusion/testing/data/csv/aggregate_test_100.csv`.
- // The default windows delimiter is `\`, so the workplace path is
`D:\a\arrow-datafusion\arrow-datafusion`.
- workspace_root.replace(std::path::MAIN_SEPARATOR,
object_store::path::DELIMITER)
- } else {
- workspace_root.to_string()
- };
-
- object_store::path::Path::parse(sanitized_workplace_root).unwrap()
-}
+ let sanitized_workplace_root = if cfg!(windows) {
+ // Object store paths are delimited with `/`, e.g.
`D:/a/arrow-datafusion/arrow-datafusion/testing/data/csv/aggregate_test_100.csv`.
+ // The default windows delimiter is `\`, so the workplace path is
`D:\a\arrow-datafusion\arrow-datafusion`.
+ workspace_root
+ .replace(std::path::MAIN_SEPARATOR,
object_store::path::DELIMITER)
+ } else {
+ workspace_root.to_string()
+ };
-// holds the root directory
-lazy_static! {
- static ref WORKSPACE_ROOT: object_store::path::Path = workspace_root();
+ object_store::path::Path::parse(sanitized_workplace_root).unwrap()
+ })
}
/// Convert a single batch to a `Vec<Vec<String>>` for comparison
