Merge remote-tracking branch 'upstream/main'

Omega359 · Mar 14, 2024 · b6d2172 · b6d2172
2 parents 2026df4 + 3c26e59
commit b6d2172
Show file tree

Hide file tree

Showing 18 changed files with 572 additions and 491 deletions.
diff --git a/datafusion-cli/Cargo.lock b/datafusion-cli/Cargo.lock
diff --git a/datafusion/expr/src/built_in_function.rs b/datafusion/expr/src/built_in_function.rs
@@ -22,11 +22,10 @@ use std::fmt;
 use std::str::FromStr;
 use std::sync::{Arc, OnceLock};
 
-use crate::signature::TIMEZONE_WILDCARD;
 use crate::type_coercion::functions::data_types;
 use crate::{FuncMonotonicity, Signature, TypeSignature, Volatility};
 
-use arrow::datatypes::{DataType, Field, TimeUnit};
+use arrow::datatypes::{DataType, Field};
 use datafusion_common::{plan_err, DataFusionError, Result};
 
 use strum::IntoEnumIterator;
@@ -190,8 +189,6 @@ pub enum BuiltinScalarFunction {
     Substr,
     /// to_hex
     ToHex,
-    /// make_date
-    MakeDate,
     /// translate
     Translate,
     /// trim
@@ -208,8 +205,6 @@ pub enum BuiltinScalarFunction {
     SubstrIndex,
     /// find_in_set
     FindInSet,
-    /// to_char
-    ToChar,
 }
 
 /// Maps the sql function name to `BuiltinScalarFunction`
@@ -335,8 +330,6 @@ impl BuiltinScalarFunction {
             BuiltinScalarFunction::Strpos => Volatility::Immutable,
             BuiltinScalarFunction::Substr => Volatility::Immutable,
             BuiltinScalarFunction::ToHex => Volatility::Immutable,
-            BuiltinScalarFunction::ToChar => Volatility::Immutable,
-            BuiltinScalarFunction::MakeDate => Volatility::Immutable,
             BuiltinScalarFunction::Translate => Volatility::Immutable,
             BuiltinScalarFunction::Trim => Volatility::Immutable,
             BuiltinScalarFunction::Upper => Volatility::Immutable,
@@ -490,8 +483,6 @@ impl BuiltinScalarFunction {
             BuiltinScalarFunction::FindInSet => {
                 utf8_to_int_type(&input_expr_types[0], "find_in_set")
             }
-            BuiltinScalarFunction::ToChar => Ok(Utf8),
-            BuiltinScalarFunction::MakeDate => Ok(Date32),
             BuiltinScalarFunction::Translate => {
                 utf8_to_str_type(&input_expr_types[0], "translate")
             }
@@ -567,7 +558,6 @@ impl BuiltinScalarFunction {
     /// Return the argument [`Signature`] supported by this function
     pub fn signature(&self) -> Signature {
         use DataType::*;
-        use TimeUnit::*;
         use TypeSignature::*;
         // note: the physical expression must accept the type returned by this function or the execution panics.
 
@@ -651,41 +641,6 @@ impl BuiltinScalarFunction {
                 vec![Exact(vec![Utf8, Int64]), Exact(vec![LargeUtf8, Int64])],
                 self.volatility(),
             ),
-            BuiltinScalarFunction::ToChar => Signature::one_of(
-                vec![
-                    Exact(vec![Date32, Utf8]),
-                    Exact(vec![Date64, Utf8]),
-                    Exact(vec![Time32(Millisecond), Utf8]),
-                    Exact(vec![Time32(Second), Utf8]),
-                    Exact(vec![Time64(Microsecond), Utf8]),
-                    Exact(vec![Time64(Nanosecond), Utf8]),
-                    Exact(vec![Timestamp(Second, None), Utf8]),
-                    Exact(vec![
-                        Timestamp(Second, Some(TIMEZONE_WILDCARD.into())),
-                        Utf8,
-                    ]),
-                    Exact(vec![Timestamp(Millisecond, None), Utf8]),
-                    Exact(vec![
-                        Timestamp(Millisecond, Some(TIMEZONE_WILDCARD.into())),
-                        Utf8,
-                    ]),
-                    Exact(vec![Timestamp(Microsecond, None), Utf8]),
-                    Exact(vec![
-                        Timestamp(Microsecond, Some(TIMEZONE_WILDCARD.into())),
-                        Utf8,
-                    ]),
-                    Exact(vec![Timestamp(Nanosecond, None), Utf8]),
-                    Exact(vec![
-                        Timestamp(Nanosecond, Some(TIMEZONE_WILDCARD.into())),
-                        Utf8,
-                    ]),
-                    Exact(vec![Duration(Second), Utf8]),
-                    Exact(vec![Duration(Millisecond), Utf8]),
-                    Exact(vec![Duration(Microsecond), Utf8]),
-                    Exact(vec![Duration(Nanosecond), Utf8]),
-                ],
-                self.volatility(),
-            ),
             BuiltinScalarFunction::SplitPart => Signature::one_of(
                 vec![
                     Exact(vec![Utf8, Utf8, Int64]),
@@ -821,11 +776,6 @@ impl BuiltinScalarFunction {
                 // will be as good as the number of digits in the number
                 Signature::uniform(1, vec![Float64, Float32], self.volatility())
             }
-            BuiltinScalarFunction::MakeDate => Signature::uniform(
-                3,
-                vec![Int32, Int64, UInt32, UInt64, Utf8],
-                self.volatility(),
-            ),
             BuiltinScalarFunction::Iszero => Signature::one_of(
                 vec![Exact(vec![Float32]), Exact(vec![Float64])],
                 self.volatility(),
@@ -943,10 +893,6 @@ impl BuiltinScalarFunction {
             BuiltinScalarFunction::SubstrIndex => &["substr_index", "substring_index"],
             BuiltinScalarFunction::FindInSet => &["find_in_set"],
 
-            // time/date functions
-            BuiltinScalarFunction::MakeDate => &["make_date"],
-            BuiltinScalarFunction::ToChar => &["to_char", "date_format"],
-
             // hashing functions
             BuiltinScalarFunction::ArrayElement => &[
                 "array_element",

diff --git a/datafusion/expr/src/expr_fn.rs b/datafusion/expr/src/expr_fn.rs
@@ -768,14 +768,6 @@ nary_scalar_expr!(
     "replace the substring of string that starts at the start'th character and extends for count characters with new substring"
 );
 
-// date functions
-scalar_expr!(
-    ToChar,
-    to_char,
-    datetime format,
-    "converts a date, time, timestamp or duration to a string based on the provided format"
-);
-scalar_expr!(MakeDate, make_date, year month day, "make a date from year, month and day component parts");
 scalar_expr!(Nanvl, nanvl, x y, "returns x if x is not NaN otherwise returns y");
 scalar_expr!(
     Iszero,

diff --git a/datafusion/functions/Cargo.toml b/datafusion/functions/Cargo.toml
@@ -58,6 +58,7 @@ path = "src/lib.rs"
 [dependencies]
 arrow = { workspace = true }
 arrow-array = { workspace = true }
+arrow-schema = { workspace = true }
 base64 = { version = "0.22", optional = true }
 blake2 = { version = "^0.10.2", optional = true }
 blake3 = { version = "1.0", optional = true }
@@ -85,3 +86,11 @@ name = "to_timestamp"
 [[bench]]
 harness = false
 name = "regx"
+
+[[bench]]
+harness = false
+name = "make_date"
+
+[[bench]]
+harness = false
+name = "to_char"
diff --git a/...fusion/physical-expr/benches/make_date.rs → datafusion/functions/benches/make_date.rs b/...fusion/physical-expr/benches/make_date.rs → datafusion/functions/benches/make_date.rs
@@ -26,7 +26,7 @@ use rand::Rng;
 
 use datafusion_common::ScalarValue;
 use datafusion_expr::ColumnarValue;
-use datafusion_physical_expr::datetime_expressions::make_date;
+use datafusion_functions::datetime::make_date;
 
 fn years(rng: &mut ThreadRng) -> Int32Array {
     let mut years = vec![];
@@ -63,7 +63,8 @@ fn criterion_benchmark(c: &mut Criterion) {
 
         b.iter(|| {
             black_box(
-                make_date(&[years.clone(), months.clone(), days.clone()])
+                make_date()
+                    .invoke(&[years.clone(), months.clone(), days.clone()])
                     .expect("make_date should work on valid values"),
             )
         })
@@ -77,7 +78,8 @@ fn criterion_benchmark(c: &mut Criterion) {
 
         b.iter(|| {
             black_box(
-                make_date(&[year.clone(), months.clone(), days.clone()])
+                make_date()
+                    .invoke(&[year.clone(), months.clone(), days.clone()])
                     .expect("make_date should work on valid values"),
             )
         })
@@ -91,7 +93,8 @@ fn criterion_benchmark(c: &mut Criterion) {
 
         b.iter(|| {
             black_box(
-                make_date(&[year.clone(), month.clone(), days.clone()])
+                make_date()
+                    .invoke(&[year.clone(), month.clone(), days.clone()])
                     .expect("make_date should work on valid values"),
             )
         })
@@ -104,7 +107,8 @@ fn criterion_benchmark(c: &mut Criterion) {
 
         b.iter(|| {
             black_box(
-                make_date(&[year.clone(), month.clone(), day.clone()])
+                make_date()
+                    .invoke(&[year.clone(), month.clone(), day.clone()])
                     .expect("make_date should work on valid values"),
             )
         })

diff --git a/datafusion/physical-expr/benches/to_char.rs → datafusion/functions/benches/to_char.rs b/datafusion/physical-expr/benches/to_char.rs → datafusion/functions/benches/to_char.rs
@@ -30,7 +30,7 @@ use rand::Rng;
 use datafusion_common::ScalarValue;
 use datafusion_common::ScalarValue::TimestampNanosecond;
 use datafusion_expr::ColumnarValue;
-use datafusion_physical_expr::datetime_expressions::to_char;
+use datafusion_functions::datetime::to_char;
 
 fn random_date_in_range(
     rng: &mut ThreadRng,
@@ -87,7 +87,8 @@ fn criterion_benchmark(c: &mut Criterion) {
 
         b.iter(|| {
             black_box(
-                to_char(&[data.clone(), patterns.clone()])
+                to_char()
+                    .invoke(&[data.clone(), patterns.clone()])
                     .expect("to_char should work on valid values"),
             )
         })
@@ -101,7 +102,8 @@ fn criterion_benchmark(c: &mut Criterion) {
 
         b.iter(|| {
             black_box(
-                to_char(&[data.clone(), patterns.clone()])
+                to_char()
+                    .invoke(&[data.clone(), patterns.clone()])
                     .expect("to_char should work on valid values"),
             )
         })
@@ -123,7 +125,8 @@ fn criterion_benchmark(c: &mut Criterion) {
 
         b.iter(|| {
             black_box(
-                to_char(&[data.clone(), pattern.clone()])
+                to_char()
+                    .invoke(&[data.clone(), pattern.clone()])
                     .expect("to_char should work on valid values"),
             )
         })