adding config to control Varchar behavior

Author: Lordworms

datafusion/common/src/config.rs

@@ -204,6 +204,12 @@ config_namespace! {
         /// MySQL, PostgreSQL, Hive, SQLite, Snowflake, Redshift, MsSQL, ClickHouse, BigQuery, and Ansi.
         pub dialect: String, default = "generic".to_string()
 
+        /// If set to true, the system will return an error when encountering a `Varchar`
+        /// type with a specified length. This can be useful for enforcing certain schema
+        /// constraints or maintaining compatibility with systems that do not support
+        /// length-specified `Varchar` types.
+        pub error_on_varchar_with_length: bool, default = false
+
     }
 }
 
@@ -303,6 +309,7 @@ config_namespace! {
         /// statistics into the same file groups.
         /// Currently experimental
         pub split_file_groups_by_statistics: bool, default = false
+
     }
 }
 

datafusion/core/src/execution/session_state.rs

@@ -563,6 +563,7 @@ impl SessionState {
         ParserOptions {
             parse_float_as_decimal: sql_parser_options.parse_float_as_decimal,
             enable_ident_normalization: sql_parser_options.enable_ident_normalization,
+            error_on_varchar_with_length: sql_parser_options.error_on_varchar_with_length,
         }
     }
 

datafusion/sql/src/planner.rs

@@ -97,13 +97,15 @@ pub trait ContextProvider {
 pub struct ParserOptions {
     pub parse_float_as_decimal: bool,
     pub enable_ident_normalization: bool,
+    pub error_on_varchar_with_length: bool,
 }
 
 impl Default for ParserOptions {
     fn default() -> Self {
         Self {
             parse_float_as_decimal: false,
             enable_ident_normalization: true,
+            error_on_varchar_with_length: false,
         }
     }
 }
@@ -398,12 +400,17 @@ impl<'a, S: ContextProvider> SqlToRel<'a, S> {
             SQLDataType::UnsignedInt(_) | SQLDataType::UnsignedInteger(_) | SQLDataType::UnsignedInt4(_) => {
                 Ok(DataType::UInt32)
             }
+            SQLDataType::Varchar(length) => {
+                match (length, self.options.error_on_varchar_with_length) {
+                    (Some(_), true) => plan_err!("does not support Varchar with length, please set corresponding parameter"),
+                    _ => Ok(DataType::Utf8),
+                }
+            }
             SQLDataType::UnsignedBigInt(_) | SQLDataType::UnsignedInt8(_) => Ok(DataType::UInt64),
             SQLDataType::Float(_) => Ok(DataType::Float32),
             SQLDataType::Real | SQLDataType::Float4 => Ok(DataType::Float32),
             SQLDataType::Double | SQLDataType::DoublePrecision | SQLDataType::Float8 => Ok(DataType::Float64),
             SQLDataType::Char(_)
-            | SQLDataType::Varchar(_)
             | SQLDataType::Text
             | SQLDataType::String(_) => Ok(DataType::Utf8),
             SQLDataType::Timestamp(None, tz_info) => {

datafusion/sqllogictest/test_files/strings.slt

@@ -78,3 +78,42 @@ e1
 p2
 p2e1
 p2m1e1
+
+statement ok
+set datafusion.sql_parser.error_on_varchar_with_length = true;
+
+# could not process due to config setting
+query error
+SELECT '12345'::VARCHAR(2);
+
+query error
+SELECT s::VARCHAR(2) FROM (VALUES ('12345')) t(s);
+
+statement ok
+create table vals(s char) as values('abc'), ('def');
+
+query error
+SELECT s::VARCHAR(2) FROM vals
+
+statement ok
+set datafusion.sql_parser.error_on_varchar_with_length = false;
+
+# could be done when we diable this setting
+query T
+SELECT '12345'::VARCHAR(2)
+----
+12345
+
+query T
+SELECT s::VARCHAR(2) FROM (VALUES ('12345')) t(s)
+----
+12345
+
+query T
+SELECT s::VARCHAR(2) FROM vals
+----
+abc
+def
+
+statement ok
+drop table vals;