Add dialect param to use CHAR instead of TEXT for Utf8 unparsing for MySQL (#12)

Author: sgrebnov

datafusion/sql/src/unparser/dialect.rs

@@ -45,6 +45,13 @@ pub trait Dialect {
     fn interval_style(&self) -> IntervalStyle {
         IntervalStyle::PostgresVerbose
     }
+
+    // Does the dialect use CHAR to cast Utf8 rather than TEXT?
+    // E.g. MySQL requires CHAR instead of TEXT and automatically produces a string with
+    // the VARCHAR, TEXT or LONGTEXT data type based on the length of the string
+    fn use_char_for_utf8_cast(&self) -> bool {
+        false
+    }
 }
 
 /// `IntervalStyle` to use for unparsing
@@ -103,6 +110,10 @@ impl Dialect for MySqlDialect {
     fn interval_style(&self) -> IntervalStyle {
         IntervalStyle::MySQL
     }
+
+    fn use_char_for_utf8_cast(&self) -> bool {
+        true
+    }
 }
 
 pub struct SqliteDialect {}
@@ -118,6 +129,7 @@ pub struct CustomDialect {
     supports_nulls_first_in_sort: bool,
     use_timestamp_for_date64: bool,
     interval_style: IntervalStyle,
+    use_char_for_utf8_cast: bool,
 }
 
 impl Default for CustomDialect {
@@ -127,6 +139,7 @@ impl Default for CustomDialect {
             supports_nulls_first_in_sort: true,
             use_timestamp_for_date64: false,
             interval_style: IntervalStyle::SQLStandard,
+            use_char_for_utf8_cast: false,
         }
     }
 }
@@ -158,6 +171,10 @@ impl Dialect for CustomDialect {
     fn interval_style(&self) -> IntervalStyle {
         self.interval_style
     }
+
+    fn use_char_for_utf8_cast(&self) -> bool {
+        self.use_char_for_utf8_cast
+    }
 }
 
 /// `CustomDialectBuilder` to build `CustomDialect` using builder pattern
@@ -179,6 +196,7 @@ pub struct CustomDialectBuilder {
     supports_nulls_first_in_sort: bool,
     use_timestamp_for_date64: bool,
     interval_style: IntervalStyle,
+    use_char_for_utf8_cast: bool,
 }
 
 impl Default for CustomDialectBuilder {
@@ -194,6 +212,7 @@ impl CustomDialectBuilder {
             supports_nulls_first_in_sort: true,
             use_timestamp_for_date64: false,
             interval_style: IntervalStyle::PostgresVerbose,
+            use_char_for_utf8_cast: false,
         }
     }
 
@@ -203,6 +222,7 @@ impl CustomDialectBuilder {
             supports_nulls_first_in_sort: self.supports_nulls_first_in_sort,
             use_timestamp_for_date64: self.use_timestamp_for_date64,
             interval_style: self.interval_style,
+            use_char_for_utf8_cast: self.use_char_for_utf8_cast,
         }
     }
 
@@ -235,4 +255,9 @@ impl CustomDialectBuilder {
         self.interval_style = interval_style;
         self
     }
+
+    pub fn with_use_char_for_utf8_cast(mut self, use_char_for_utf8_cast: bool) -> Self {
+        self.use_char_for_utf8_cast = use_char_for_utf8_cast;
+        self
+    }
 }

datafusion/sql/src/unparser/expr.rs

@@ -1272,8 +1272,16 @@ impl Unparser<'_> {
             DataType::BinaryView => {
                 not_impl_err!("Unsupported DataType: conversion: {data_type:?}")
             }
-            DataType::Utf8 => Ok(ast::DataType::Varchar(None)),
-            DataType::LargeUtf8 => Ok(ast::DataType::Text),
+            DataType::Utf8 => Ok(if self.dialect.use_char_for_utf8_cast() {
+                ast::DataType::Char(None)
+            } else {
+                ast::DataType::Varchar(None)
+            }),
+            DataType::LargeUtf8 => Ok(if self.dialect.use_char_for_utf8_cast() {
+                ast::DataType::Char(None)
+            } else {
+                ast::DataType::Text
+            }),
             DataType::Utf8View => {
                 not_impl_err!("Unsupported DataType: conversion: {data_type:?}")
             }
@@ -1933,4 +1941,31 @@ mod tests {
             assert_eq!(actual, expected);
         }
     }
+
+    #[test]
+    fn custom_dialect_use_char_for_utf8_cast() -> Result<()> {
+        for (use_char_for_utf8_cast, data_type, identifier) in [
+            (false, DataType::Utf8, "VARCHAR"),
+            (true, DataType::Utf8, "CHAR"),
+            (false, DataType::LargeUtf8, "TEXT"),
+            (true, DataType::LargeUtf8, "CHAR"),
+        ] {
+            let dialect = CustomDialectBuilder::new()
+                .with_use_char_for_utf8_cast(use_char_for_utf8_cast)
+                .build();
+            let unparser = Unparser::new(&dialect);
+
+            let expr = Expr::Cast(Cast {
+                expr: Box::new(col("a")),
+                data_type,
+            });
+            let ast = unparser.expr_to_sql(&expr)?;
+
+            let actual = format!("{}", ast);
+            let expected = format!(r#"CAST(a AS {identifier})"#);
+
+            assert_eq!(actual, expected);
+        }
+        Ok(())
+    }
 }