Oracle: Support for MERGE predicates (#2101)

Co-authored-by: Ifeanyi Ubah <[email protected]>

Author: xitep

src/ast/dml.rs

@@ -24,13 +24,16 @@ use serde::{Deserialize, Serialize};
 #[cfg(feature = "visitor")]
 use sqlparser_derive::{Visit, VisitMut};
 
-use crate::display_utils::{indented_list, Indent, SpaceOrNewline};
+use crate::{
+    ast::display_separated,
+    display_utils::{indented_list, Indent, SpaceOrNewline},
+};
 
 use super::{
     display_comma_separated, helpers::attached_token::AttachedToken, query::InputFormatClause,
     Assignment, Expr, FromTable, Ident, InsertAliases, MysqlInsertPriority, ObjectName, OnInsert,
-    OrderByExpr, Query, SelectItem, Setting, SqliteOnConflict, TableObject, TableWithJoins,
-    UpdateTableFromKind,
+    OrderByExpr, Query, SelectInto, SelectItem, Setting, SqliteOnConflict, TableFactor,
+    TableObject, TableWithJoins, UpdateTableFromKind, Values,
 };
 
 /// INSERT statement.
@@ -310,3 +313,334 @@ impl Display for Update {
         Ok(())
     }
 }
+
+/// A `MERGE` statement.
+#[derive(Debug, Clone, PartialEq, PartialOrd, Eq, Ord, Hash)]
+#[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
+#[cfg_attr(feature = "visitor", derive(Visit, VisitMut))]
+pub struct Merge {
+    /// The `MERGE` token that starts the statement.
+    pub merge_token: AttachedToken,
+    /// optional INTO keyword
+    pub into: bool,
+    /// Specifies the table to merge
+    pub table: TableFactor,
+    /// Specifies the table or subquery to join with the target table
+    pub source: TableFactor,
+    /// Specifies the expression on which to join the target table and source
+    pub on: Box<Expr>,
+    /// Specifies the actions to perform when values match or do not match.
+    pub clauses: Vec<MergeClause>,
+    // Specifies the output to save changes in MSSQL
+    pub output: Option<OutputClause>,
+}
+
+impl Display for Merge {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        write!(
+            f,
+            "MERGE{int} {table} USING {source} ",
+            int = if self.into { " INTO" } else { "" },
+            table = self.table,
+            source = self.source,
+        )?;
+        write!(f, "ON {on} ", on = self.on)?;
+        write!(f, "{}", display_separated(&self.clauses, " "))?;
+        if let Some(ref output) = self.output {
+            write!(f, " {output}")?;
+        }
+        Ok(())
+    }
+}
+
+/// A `WHEN` clause within a `MERGE` Statement
+///
+/// Example:
+/// ```sql
+/// WHEN NOT MATCHED BY SOURCE AND product LIKE '%washer%' THEN DELETE
+/// ```
+/// [Snowflake](https://docs.snowflake.com/en/sql-reference/sql/merge)
+/// [BigQuery](https://cloud.google.com/bigquery/docs/reference/standard-sql/dml-syntax#merge_statement)
+#[derive(Debug, Clone, PartialEq, PartialOrd, Eq, Ord, Hash)]
+#[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
+#[cfg_attr(feature = "visitor", derive(Visit, VisitMut))]
+pub struct MergeClause {
+    /// The `WHEN` token that starts the sub-expression.
+    pub when_token: AttachedToken,
+    pub clause_kind: MergeClauseKind,
+    pub predicate: Option<Expr>,
+    pub action: MergeAction,
+}
+
+impl Display for MergeClause {
+    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
+        let MergeClause {
+            when_token: _,
+            clause_kind,
+            predicate,
+            action,
+        } = self;
+
+        write!(f, "WHEN {clause_kind}")?;
+        if let Some(pred) = predicate {
+            write!(f, " AND {pred}")?;
+        }
+        write!(f, " THEN {action}")
+    }
+}
+
+/// Variant of `WHEN` clause used within a `MERGE` Statement.
+///
+/// Example:
+/// ```sql
+/// MERGE INTO T USING U ON FALSE WHEN MATCHED THEN DELETE
+/// ```
+/// [Snowflake](https://docs.snowflake.com/en/sql-reference/sql/merge)
+/// [BigQuery](https://cloud.google.com/bigquery/docs/reference/standard-sql/dml-syntax#merge_statement)
+#[derive(Debug, Clone, PartialEq, PartialOrd, Eq, Ord, Hash)]
+#[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
+#[cfg_attr(feature = "visitor", derive(Visit, VisitMut))]
+pub enum MergeClauseKind {
+    /// `WHEN MATCHED`
+    Matched,
+    /// `WHEN NOT MATCHED`
+    NotMatched,
+    /// `WHEN MATCHED BY TARGET`
+    ///
+    /// [BigQuery](https://cloud.google.com/bigquery/docs/reference/standard-sql/dml-syntax#merge_statement)
+    NotMatchedByTarget,
+    /// `WHEN MATCHED BY SOURCE`
+    ///
+    /// [BigQuery](https://cloud.google.com/bigquery/docs/reference/standard-sql/dml-syntax#merge_statement)
+    NotMatchedBySource,
+}
+
+impl Display for MergeClauseKind {
+    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
+        match self {
+            MergeClauseKind::Matched => write!(f, "MATCHED"),
+            MergeClauseKind::NotMatched => write!(f, "NOT MATCHED"),
+            MergeClauseKind::NotMatchedByTarget => write!(f, "NOT MATCHED BY TARGET"),
+            MergeClauseKind::NotMatchedBySource => write!(f, "NOT MATCHED BY SOURCE"),
+        }
+    }
+}
+
+/// Underlying statement of a `WHEN` clause within a `MERGE` Statement
+///
+/// Example
+/// ```sql
+/// INSERT (product, quantity) VALUES(product, quantity)
+/// ```
+///
+/// [Snowflake](https://docs.snowflake.com/en/sql-reference/sql/merge)
+/// [BigQuery](https://cloud.google.com/bigquery/docs/reference/standard-sql/dml-syntax#merge_statement)
+/// [Oracle](https://docs.oracle.com/en/database/oracle/oracle-database/21/sqlrf/MERGE.html)
+#[derive(Debug, Clone, PartialEq, PartialOrd, Eq, Ord, Hash)]
+#[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
+#[cfg_attr(feature = "visitor", derive(Visit, VisitMut))]
+pub enum MergeAction {
+    /// An `INSERT` clause
+    ///
+    /// Example:
+    /// ```sql
+    /// INSERT (product, quantity) VALUES(product, quantity)
+    /// ```
+    Insert(MergeInsertExpr),
+    /// An `UPDATE` clause
+    ///
+    /// Example:
+    /// ```sql
+    /// UPDATE SET quantity = T.quantity + S.quantity
+    /// ```
+    Update(MergeUpdateExpr),
+    /// A plain `DELETE` clause
+    Delete {
+        /// The `DELETE` token that starts the sub-expression.
+        delete_token: AttachedToken,
+    },
+}
+
+impl Display for MergeAction {
+    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
+        match self {
+            MergeAction::Insert(insert) => {
+                write!(f, "INSERT {insert}")
+            }
+            MergeAction::Update(update) => {
+                write!(f, "UPDATE {update}")
+            }
+            MergeAction::Delete { .. } => {
+                write!(f, "DELETE")
+            }
+        }
+    }
+}
+
+/// The type of expression used to insert rows within a `MERGE` statement.
+///
+/// [Snowflake](https://docs.snowflake.com/en/sql-reference/sql/merge)
+/// [BigQuery](https://cloud.google.com/bigquery/docs/reference/standard-sql/dml-syntax#merge_statement)
+#[derive(Debug, Clone, PartialEq, PartialOrd, Eq, Ord, Hash)]
+#[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
+#[cfg_attr(feature = "visitor", derive(Visit, VisitMut))]
+pub enum MergeInsertKind {
+    /// The insert expression is defined from an explicit `VALUES` clause
+    ///
+    /// Example:
+    /// ```sql
+    /// INSERT VALUES(product, quantity)
+    /// ```
+    Values(Values),
+    /// The insert expression is defined using only the `ROW` keyword.
+    ///
+    /// Example:
+    /// ```sql
+    /// INSERT ROW
+    /// ```
+    /// [BigQuery](https://cloud.google.com/bigquery/docs/reference/standard-sql/dml-syntax#merge_statement)
+    Row,
+}
+
+impl Display for MergeInsertKind {
+    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
+        match self {
+            MergeInsertKind::Values(values) => {
+                write!(f, "{values}")
+            }
+            MergeInsertKind::Row => {
+                write!(f, "ROW")
+            }
+        }
+    }
+}
+
+/// The expression used to insert rows within a `MERGE` statement.
+///
+/// Examples
+/// ```sql
+/// INSERT (product, quantity) VALUES(product, quantity)
+/// INSERT ROW
+/// ```
+///
+/// [Snowflake](https://docs.snowflake.com/en/sql-reference/sql/merge)
+/// [BigQuery](https://cloud.google.com/bigquery/docs/reference/standard-sql/dml-syntax#merge_statement)
+/// [Oracle](https://docs.oracle.com/en/database/oracle/oracle-database/21/sqlrf/MERGE.html)
+#[derive(Debug, Clone, PartialEq, PartialOrd, Eq, Ord, Hash)]
+#[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
+#[cfg_attr(feature = "visitor", derive(Visit, VisitMut))]
+pub struct MergeInsertExpr {
+    /// The `INSERT` token that starts the sub-expression.
+    pub insert_token: AttachedToken,
+    /// Columns (if any) specified by the insert.
+    ///
+    /// Example:
+    /// ```sql
+    /// INSERT (product, quantity) VALUES(product, quantity)
+    /// INSERT (product, quantity) ROW
+    /// ```
+    pub columns: Vec<ObjectName>,
+    /// The token, `[VALUES | ROW]` starting `kind`.
+    pub kind_token: AttachedToken,
+    /// The insert type used by the statement.
+    pub kind: MergeInsertKind,
+    /// An optional condition to restrict the insertion (Oracle specific)
+    pub insert_predicate: Option<Expr>,
+}
+
+impl Display for MergeInsertExpr {
+    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
+        if !self.columns.is_empty() {
+            write!(f, "({}) ", display_comma_separated(self.columns.as_slice()))?;
+        }
+        write!(f, "{}", self.kind)?;
+        if let Some(predicate) = self.insert_predicate.as_ref() {
+            write!(f, " WHERE {}", predicate)?;
+        }
+        Ok(())
+    }
+}
+
+/// The expression used to update rows within a `MERGE` statement.
+///
+/// Examples
+/// ```sql
+/// UPDATE SET quantity = T.quantity + S.quantity
+/// ```
+///
+/// [Snowflake](https://docs.snowflake.com/en/sql-reference/sql/merge)
+/// [BigQuery](https://cloud.google.com/bigquery/docs/reference/standard-sql/dml-syntax#merge_statement)
+/// [Oracle](https://docs.oracle.com/en/database/oracle/oracle-database/21/sqlrf/MERGE.html)
+#[derive(Debug, Clone, PartialEq, PartialOrd, Eq, Ord, Hash)]
+#[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
+#[cfg_attr(feature = "visitor", derive(Visit, VisitMut))]
+pub struct MergeUpdateExpr {
+    /// The `UPDATE` token that starts the sub-expression.
+    pub update_token: AttachedToken,
+    /// The update assiment expressions
+    pub assignments: Vec<Assignment>,
+    /// `where_clause` for the update (Oralce specific)
+    pub update_predicate: Option<Expr>,
+    /// `delete_clause` for the update "delete where" (Oracle specific)
+    pub delete_predicate: Option<Expr>,
+}
+
+impl Display for MergeUpdateExpr {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        write!(f, "SET {}", display_comma_separated(&self.assignments))?;
+        if let Some(predicate) = self.update_predicate.as_ref() {
+            write!(f, " WHERE {predicate}")?;
+        }
+        if let Some(predicate) = self.delete_predicate.as_ref() {
+            write!(f, " DELETE WHERE {predicate}")?;
+        }
+        Ok(())
+    }
+}
+
+/// A `OUTPUT` Clause in the end of a `MERGE` Statement
+///
+/// Example:
+/// OUTPUT $action, deleted.* INTO dbo.temp_products;
+/// [mssql](https://learn.microsoft.com/en-us/sql/t-sql/queries/output-clause-transact-sql)
+#[derive(Debug, Clone, PartialEq, PartialOrd, Eq, Ord, Hash)]
+#[cfg_attr(feature = "serde", derive(Serialize, Deserialize))]
+#[cfg_attr(feature = "visitor", derive(Visit, VisitMut))]
+pub enum OutputClause {
+    Output {
+        output_token: AttachedToken,
+        select_items: Vec<SelectItem>,
+        into_table: Option<SelectInto>,
+    },
+    Returning {
+        returning_token: AttachedToken,
+        select_items: Vec<SelectItem>,
+    },
+}
+
+impl fmt::Display for OutputClause {
+    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
+        match self {
+            OutputClause::Output {
+                output_token: _,
+                select_items,
+                into_table,
+            } => {
+                f.write_str("OUTPUT ")?;
+                display_comma_separated(select_items).fmt(f)?;
+                if let Some(into_table) = into_table {
+                    f.write_str(" ")?;
+                    into_table.fmt(f)?;
+                }
+                Ok(())
+            }
+            OutputClause::Returning {
+                returning_token: _,
+                select_items,
+            } => {
+                f.write_str("RETURNING ")?;
+                display_comma_separated(select_items).fmt(f)
+            }
+        }
+    }
+}