feat: Coerce strings in binary expressions

MazterQyou · MazterQyou · commit 5134f15fb045 · 2025-10-23T23:52:45.000+04:00
Signed-off-by: Alex Qyoun-ae &lt;4062971+MazterQyou@users.noreply.github.com&gt;
diff --git a/Cargo.lock b/Cargo.lock
diff --git a/datafusion-cli/Cargo.lock b/datafusion-cli/Cargo.lock
diff --git a/datafusion-cli/Cargo.toml b/datafusion-cli/Cargo.toml
@@ -28,7 +28,7 @@ repository = "https://github.com/apache/arrow-datafusion"
 rust-version = "1.59"
 
 [dependencies]
-arrow = { git = 'https://github.com/cube-js/arrow-rs.git', rev = "a03d4eef5640e05dddf99fc2357ad6d58b5337cb" }
+arrow = { git = 'https://github.com/cube-js/arrow-rs.git', rev = "8277f9a051c90482ff9557a91c5dc3926d9ba19e" }
 clap = { version = "3", features = ["derive", "cargo"] }
 datafusion = { path = "../datafusion/core", version = "7.0.0", features = ["parquet-all-compressions"] }
 dirs = "4.0.0"
diff --git a/datafusion-examples/Cargo.toml b/datafusion-examples/Cargo.toml
@@ -34,7 +34,7 @@ path = "examples/avro_sql.rs"
 required-features = ["datafusion/avro"]
 
 [dev-dependencies]
-arrow-flight = { git = 'https://github.com/cube-js/arrow-rs.git', rev = "a03d4eef5640e05dddf99fc2357ad6d58b5337cb" }
+arrow-flight = { git = 'https://github.com/cube-js/arrow-rs.git', rev = "8277f9a051c90482ff9557a91c5dc3926d9ba19e" }
 async-trait = "0.1.41"
 datafusion = { path = "../datafusion/core" }
 futures = "0.3"
diff --git a/datafusion/common/Cargo.toml b/datafusion/common/Cargo.toml
@@ -40,10 +40,10 @@ pyarrow = ["pyo3"]
 parquet-all-compressions = ["parquet/snap", "parquet/brotli", "parquet/flate2", "parquet/lz4", "parquet/zstd"]
 
 [dependencies]
-arrow = { git = 'https://github.com/cube-js/arrow-rs.git', rev = "a03d4eef5640e05dddf99fc2357ad6d58b5337cb", features = ["prettyprint"] }
+arrow = { git = 'https://github.com/cube-js/arrow-rs.git', rev = "8277f9a051c90482ff9557a91c5dc3926d9ba19e", features = ["prettyprint"] }
 avro-rs = { version = "0.13", features = ["snappy"], optional = true }
 cranelift-module = { version = "0.82.0", optional = true }
 ordered-float = "2.10"
-parquet = { git = 'https://github.com/cube-js/arrow-rs.git', rev = "a03d4eef5640e05dddf99fc2357ad6d58b5337cb", features = ["arrow", "base64"], default-features = false, optional = true }
+parquet = { git = 'https://github.com/cube-js/arrow-rs.git', rev = "8277f9a051c90482ff9557a91c5dc3926d9ba19e", features = ["arrow", "base64"], default-features = false, optional = true }
 pyo3 = { version = "0.16", optional = true }
 sqlparser = { git = 'https://github.com/cube-js/sqlparser-rs.git', rev = "16f051486de78a23a0ff252155dd59fc2d35497d" }
diff --git a/datafusion/core/Cargo.toml b/datafusion/core/Cargo.toml
@@ -57,7 +57,7 @@ parquet-all-compressions = ["datafusion-common/parquet-all-compressions", "parqu
 
 [dependencies]
 ahash = { version = "0.7", default-features = false }
-arrow = { git = 'https://github.com/cube-js/arrow-rs.git', rev = "a03d4eef5640e05dddf99fc2357ad6d58b5337cb", features = ["prettyprint"] }
+arrow = { git = 'https://github.com/cube-js/arrow-rs.git', rev = "8277f9a051c90482ff9557a91c5dc3926d9ba19e", features = ["prettyprint"] }
 async-trait = "0.1.41"
 avro-rs = { version = "0.13", features = ["snappy"], optional = true }
 chrono = { version = "0.4", default-features = false }
@@ -76,7 +76,7 @@ num_cpus = "1.13.0"
 ordered-float = "2.10"
 parking_lot = "0.12"
 # All compression codes are disabled by default in our fork because it increases compilation time significantly.
-parquet = { git = 'https://github.com/cube-js/arrow-rs.git', rev = "a03d4eef5640e05dddf99fc2357ad6d58b5337cb", features = ["arrow", "base64"], default-features = false }
+parquet = { git = 'https://github.com/cube-js/arrow-rs.git', rev = "8277f9a051c90482ff9557a91c5dc3926d9ba19e", features = ["arrow", "base64"], default-features = false }
 paste = "^1.0"
 pin-project-lite= "^0.2.7"
 pyo3 = { version = "0.16", optional = true }
diff --git a/datafusion/core/fuzz-utils/Cargo.toml b/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 = { git = 'https://github.com/cube-js/arrow-rs.git', rev = "a03d4eef5640e05dddf99fc2357ad6d58b5337cb", features = ["prettyprint"] }
+arrow = { git = 'https://github.com/cube-js/arrow-rs.git', rev = "8277f9a051c90482ff9557a91c5dc3926d9ba19e", features = ["prettyprint"] }
 env_logger = "0.9.0"
 rand = "0.8"
diff --git a/datafusion/core/src/physical_plan/planner.rs b/datafusion/core/src/physical_plan/planner.rs
@@ -1929,7 +1929,8 @@ mod tests {
             // u32 AND bool
             col("c2").and(bool_expr),
             // utf8 LIKE u32
-            col("c1").like(col("c2")),
+            // NOTE(cubesql): valid
+            //col("c1").like(col("c2")),
         ];
         for case in cases {
             let logical_plan = LogicalPlanBuilder::scan_csv(
diff --git a/datafusion/cube_ext/Cargo.toml b/datafusion/cube_ext/Cargo.toml
@@ -35,7 +35,7 @@ name = "cube_ext"
 path = "src/lib.rs"
 
 [dependencies]
-arrow = { git = 'https://github.com/cube-js/arrow-rs.git', rev = "a03d4eef5640e05dddf99fc2357ad6d58b5337cb", features = ["prettyprint"] }
+arrow = { git = 'https://github.com/cube-js/arrow-rs.git', rev = "8277f9a051c90482ff9557a91c5dc3926d9ba19e", features = ["prettyprint"] }
 chrono = { version = "0.4.16", package = "chrono", default-features = false, features = ["clock"] }
 datafusion-common = { path = "../common", version = "7.0.0" }
 datafusion-expr = { path = "../expr", version = "7.0.0" }
diff --git a/datafusion/expr/Cargo.toml b/datafusion/expr/Cargo.toml
@@ -36,6 +36,6 @@ path = "src/lib.rs"
 
 [dependencies]
 ahash = { version = "0.7", default-features = false }
-arrow = { git = 'https://github.com/cube-js/arrow-rs.git', rev = "a03d4eef5640e05dddf99fc2357ad6d58b5337cb", features = ["prettyprint"] }
+arrow = { git = 'https://github.com/cube-js/arrow-rs.git', rev = "8277f9a051c90482ff9557a91c5dc3926d9ba19e", features = ["prettyprint"] }
 datafusion-common = { path = "../common", version = "7.0.0" }
 sqlparser = { git = 'https://github.com/cube-js/sqlparser-rs.git', rev = "16f051486de78a23a0ff252155dd59fc2d35497d" }
diff --git a/datafusion/expr/src/binary_rule.rs b/datafusion/expr/src/binary_rule.rs
@@ -151,9 +151,13 @@ pub fn coerce_types(
 fn bitwise_coercion(left_type: &DataType, right_type: &DataType) -> Option<DataType> {
     use arrow::datatypes::DataType::*;
 
-    if !is_numeric(left_type) || !is_numeric(right_type) {
-        return None;
+    // If one of the sides is numeric and the other is a string, coercion to number is allowed
+    match (is_numeric(left_type), is_numeric(right_type)) {
+        (true, true) => (),
+        (false, false) => return None,
+        _ => return string_with_any_coercion(left_type, right_type),
     }
+
     if left_type == right_type && !is_dictionary(left_type) {
         return Some(left_type.clone());
     }
@@ -242,9 +246,13 @@ fn comparison_binary_numeric_coercion(
     rhs_type: &DataType,
 ) -> Option<DataType> {
     use arrow::datatypes::DataType::*;
-    if !is_numeric(lhs_type) || !is_numeric(rhs_type) {
-        return None;
-    };
+
+    // If one of the sides is numeric and the other is a string, coercion to number is allowed
+    match (is_numeric(lhs_type), is_numeric(lhs_type)) {
+        (true, true) => (),
+        (false, false) => return None,
+        _ => return string_with_any_coercion(lhs_type, rhs_type),
+    }
 
     // same type => all good
     if lhs_type == rhs_type {
@@ -322,9 +330,12 @@ fn mathematics_numerical_coercion(
     use arrow::datatypes::DataType::*;
 
     // error on any non-numeric type
-    if !is_numeric(lhs_type) || !is_numeric(rhs_type) {
-        return None;
-    };
+    // If one of the sides is numeric and the other is a string, coercion to number is allowed
+    match (is_numeric(lhs_type), is_numeric(rhs_type)) {
+        (true, true) => (),
+        (false, false) => return None,
+        _ => return string_with_any_coercion(lhs_type, rhs_type),
+    }
 
     // exponentiation is always Float64
     if mathematics_op == &Operator::Exponentiate {
@@ -557,6 +568,7 @@ fn string_boolean_equality_coercion(
 fn like_coercion(lhs_type: &DataType, rhs_type: &DataType) -> Option<DataType> {
     string_coercion(lhs_type, rhs_type)
         .or_else(|| dictionary_coercion(lhs_type, rhs_type))
+        .or_else(|| string_with_any_coercion(lhs_type, rhs_type))
 }
 
 /// Coercion rules for Temporal columns: the type that both lhs and rhs can be
@@ -668,6 +680,7 @@ fn eq_coercion(lhs_type: &DataType, rhs_type: &DataType) -> Option<DataType> {
         .or_else(|| dictionary_coercion(lhs_type, rhs_type))
         .or_else(|| temporal_coercion(lhs_type, rhs_type))
         .or_else(|| null_coercion(lhs_type, rhs_type))
+        .or_else(|| string_with_any_coercion(lhs_type, rhs_type))
 }
 
 pub fn distinct_coercion(
@@ -803,6 +816,17 @@ fn interval_add_coercion(lhs_type: &DataType, rhs_type: &DataType) -> Option<Dat
     }
 }
 
+/// coercion rules for expressions where string can be cast to other type
+fn string_with_any_coercion(
+    lhs_type: &DataType,
+    rhs_type: &DataType,
+) -> Option<DataType> {
+    match (lhs_type, rhs_type) {
+        (DataType::Utf8, t) | (t, DataType::Utf8) => Some(t.clone()),
+        _ => None,
+    }
+}
+
 #[cfg(test)]
 mod tests {
     use super::*;
@@ -815,10 +839,10 @@ mod tests {
 
     fn test_coercion_error() -> Result<()> {
         let result_type =
-            coerce_types(&DataType::Float32, &Operator::Plus, &DataType::Utf8);
+            coerce_types(&DataType::Float32, &Operator::Plus, &DataType::Date32);
 
         if let Err(DataFusionError::Plan(e)) = result_type {
-            assert_eq!(e, "'Float32 + Utf8' can't be evaluated because there isn't a common type to coerce the types to");
+            assert_eq!(e, "'Float32 + Date32' can't be evaluated because there isn't a common type to coerce the types to");
             Ok(())
         } else {
             Err(DataFusionError::Internal(
diff --git a/datafusion/jit/Cargo.toml b/datafusion/jit/Cargo.toml
@@ -36,7 +36,7 @@ path = "src/lib.rs"
 jit = []
 
 [dependencies]
-arrow = { git = 'https://github.com/cube-js/arrow-rs.git', rev = "a03d4eef5640e05dddf99fc2357ad6d58b5337cb" }
+arrow = { git = 'https://github.com/cube-js/arrow-rs.git', rev = "8277f9a051c90482ff9557a91c5dc3926d9ba19e" }
 cranelift = "0.82.0"
 cranelift-jit = "0.82.0"
 cranelift-module = "0.82.0"
diff --git a/datafusion/physical-expr/Cargo.toml b/datafusion/physical-expr/Cargo.toml
@@ -40,7 +40,7 @@ unicode_expressions = ["unicode-segmentation"]
 
 [dependencies]
 ahash = { version = "0.7", default-features = false }
-arrow = { git = 'https://github.com/cube-js/arrow-rs.git', rev = "a03d4eef5640e05dddf99fc2357ad6d58b5337cb", features = ["prettyprint"] }
+arrow = { git = 'https://github.com/cube-js/arrow-rs.git', rev = "8277f9a051c90482ff9557a91c5dc3926d9ba19e", features = ["prettyprint"] }
 blake2 = { version = "^0.10.2", optional = true }
 blake3 = { version = "1.0", optional = true }
 chrono = { version = "0.4.20", default-features = false }
