polyglot-sql 0.3.2__tar.gz → 0.3.4__tar.gz

{polyglot_sql-0.3.2 → polyglot_sql-0.3.4}/Cargo.lock

@@ -73,6 +73,15 @@ version = "1.0.102"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "7f202df86484c868dbad7eaa557ef785d5c66295e41b460ef922eca0723b842c"
 
+[[package]]
+name = "ar_archive_writer"
+version = "0.5.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "7eb93bbb63b9c227414f6eb3a0adfddca591a8ce1e9b60661bb08969b87e340b"
+dependencies = [
+ "object",
+]
+
 [[package]]
 name = "async-trait"
 version = "0.1.89"
@@ -527,6 +536,15 @@ dependencies = [
  "libm",
 ]
 
+[[package]]
+name = "object"
+version = "0.37.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "ff76201f031d8863c38aa7f905eca4f53abbfa15f609db4277d44cd8938f33fe"
+dependencies = [
+ "memchr",
+]
+
 [[package]]
 name = "once_cell"
 version = "1.21.3"
@@ -587,7 +605,7 @@ dependencies = [
 
 [[package]]
 name = "polyglot-sql"
-version = "0.3.2"
+version = "0.3.4"
 dependencies = [
  "criterion",
  "once_cell",
@@ -595,6 +613,7 @@ dependencies = [
  "pretty_assertions",
  "serde",
  "serde_json",
+ "stacker",
  "thiserror 1.0.69",
  "ts-rs",
  "unicode-segmentation",
@@ -602,7 +621,7 @@ dependencies = [
 
 [[package]]
 name = "polyglot-sql-ffi"
-version = "0.3.2"
+version = "0.3.4"
 dependencies = [
  "cbindgen",
  "polyglot-sql",
@@ -612,11 +631,11 @@ dependencies = [
 
 [[package]]
 name = "polyglot-sql-function-catalogs"
-version = "0.3.2"
+version = "0.3.4"
 
 [[package]]
 name = "polyglot-sql-python"
-version = "0.3.2"
+version = "0.3.4"
 dependencies = [
  "polyglot-sql",
  "pyo3",
@@ -627,7 +646,7 @@ dependencies = [
 
 [[package]]
 name = "polyglot-sql-wasm"
-version = "0.3.2"
+version = "0.3.4"
 dependencies = [
  "console_error_panic_hook",
  "js-sys",
@@ -674,6 +693,16 @@ dependencies = [
  "unicode-ident",
 ]
 
+[[package]]
+name = "psm"
+version = "0.1.31"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "645dbe486e346d9b5de3ef16ede18c26e6c70ad97418f4874b8b1889d6e761ea"
+dependencies = [
+ "ar_archive_writer",
+ "cc",
+]
+
 [[package]]
 name = "pyo3"
 version = "0.28.2"
@@ -915,6 +944,19 @@ version = "0.4.11"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "7a2ae44ef20feb57a68b23d846850f861394c2e02dc425a50098ae8c90267589"
 
+[[package]]
+name = "stacker"
+version = "0.1.24"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "640c8cdd92b6b12f5bcb1803ca3bbf5ab96e5e6b6b96b9ab77dabe9e880b3190"
+dependencies = [
+ "cc",
+ "cfg-if",
+ "libc",
+ "psm",
+ "windows-sys",
+]
+
 [[package]]
 name = "strsim"
 version = "0.11.1"

{polyglot_sql-0.3.2 → polyglot_sql-0.3.4}/Cargo.toml

@@ -6,7 +6,7 @@ exclude = [
 ]
 
 [workspace.package]
-version = "0.3.2"
+version = "0.3.4"
 edition = "2021"
 license = "MIT"
 authors = ["polyglot contributors"]

{polyglot_sql-0.3.2 → polyglot_sql-0.3.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: polyglot-sql
-Version: 0.3.2
+Version: 0.3.4
 Classifier: Development Status :: 4 - Beta
 Classifier: Intended Audience :: Developers
 Classifier: License :: OSI Approved :: MIT License

{polyglot_sql-0.3.2 → polyglot_sql-0.3.4}/crates/polyglot-sql/Cargo.toml

@@ -10,9 +10,13 @@ description = "SQL parsing, validating, formatting, and dialect translation libr
 readme = "README.md"
 
 [features]
-default = ["all-dialects", "transpile"]
+default = ["all-dialects", "transpile", "stacker"]
 bindings = ["dep:ts-rs"]
 transpile = []
+# Default-on for native builds so direct Rust users and the FFI/Python bindings
+# inherit stack-growth protection automatically. WASM opts out via
+# `default-features = false` in its dependency declaration.
+stacker = ["dep:stacker"]
 all-dialects = [
     "dialect-postgresql", "dialect-mysql", "dialect-bigquery",
     "dialect-snowflake", "dialect-duckdb", "dialect-tsql",
@@ -77,8 +81,9 @@ serde = { workspace = true }
 serde_json = { workspace = true }
 thiserror = { workspace = true }
 unicode-segmentation = { workspace = true }
+stacker = { version = "0.1", optional = true }
 ts-rs = { version = "12.0", features = ["serde-compat"], optional = true }
-polyglot-sql-function-catalogs = { path = "../polyglot-sql-function-catalogs", version = "0.3.2", optional = true, default-features = false }
+polyglot-sql-function-catalogs = { path = "../polyglot-sql-function-catalogs", version = "0.3.4", optional = true, default-features = false }
 
 [dev-dependencies]
 pretty_assertions = "1.4"

{polyglot_sql-0.3.2 → polyglot_sql-0.3.4}/crates/polyglot-sql/src/ast_transforms.rs

@@ -434,6 +434,10 @@ pub fn get_table_names(expr: &Expression) -> Vec<String> {
                     names.push(name);
                 }
             }
+            Expression::CreateSynonym(synonym) => {
+                push_table_ref_name(&synonym.name, &cte_aliases, &mut names);
+                push_table_ref_name(&synonym.target, &cte_aliases, &mut names);
+            }
             _ => {}
         }
     }

{polyglot_sql-0.3.2 → polyglot_sql-0.3.4}/crates/polyglot-sql/src/builder.rs

@@ -803,6 +803,7 @@ pub fn delete(table_name: &str) -> DeleteBuilder {
     DeleteBuilder {
         delete: Delete {
             table: builder_table_ref(table_name),
+            hint: None,
             on_cluster: None,
             alias: None,
             alias_explicit_as: false,
@@ -894,6 +895,7 @@ pub fn update(table_name: &str) -> UpdateBuilder {
     UpdateBuilder {
         update: Update {
             table: builder_table_ref(table_name),
+            hint: None,
             extra_tables: Vec::new(),
             table_joins: Vec::new(),
             set: Vec::new(),
@@ -1601,6 +1603,7 @@ impl SelectBuilder {
             clone_at_clause: None,
             is_copy: false,
             shallow_clone: false,
+            deep_clone: false,
             leading_comments: vec![],
             with_properties: vec![],
             teradata_post_name_options: vec![],

{polyglot_sql-0.3.2 → polyglot_sql-0.3.4}/crates/polyglot-sql/src/dialects/databricks.rs

@@ -37,6 +37,12 @@ impl DialectImpl for DatabricksDialect {
         config
             .keywords
             .insert("DIV".to_string(), crate::tokens::TokenType::Div);
+        config
+            .keywords
+            .insert("REPAIR".to_string(), crate::tokens::TokenType::Command);
+        config
+            .keywords
+            .insert("MSCK".to_string(), crate::tokens::TokenType::Command);
         // Databricks numeric literal suffixes (same as Hive/Spark)
         config
             .numeric_literals
@@ -1009,4 +1015,104 @@ mod tests {
             "FROM_UTC_TIMESTAMP with existing CAST failed"
         );
     }
+
+    #[test]
+    fn test_deep_clone_version_as_of() {
+        let sql = "CREATE TABLE events_clone DEEP CLONE events VERSION AS OF 5";
+        let d = Dialect::get(DialectType::Databricks);
+        let ast = d.parse(sql).expect("Parse failed");
+        let output = d.generate(&ast[0]).expect("Generate failed");
+
+        assert_eq!(output, sql);
+    }
+
+    #[test]
+    fn test_deep_clone_timestamp_as_of() {
+        let sql = "CREATE TABLE events_clone DEEP CLONE events TIMESTAMP AS OF '2024-01-01'";
+        let d = Dialect::get(DialectType::Databricks);
+        let ast = d.parse(sql).expect("Parse failed");
+        let output = d.generate(&ast[0]).expect("Generate failed");
+
+        assert_eq!(output, sql);
+    }
+
+    #[test]
+    fn test_shallow_clone_still_roundtrips() {
+        let sql = "CREATE TABLE events_clone SHALLOW CLONE events";
+        let d = Dialect::get(DialectType::Databricks);
+        let ast = d.parse(sql).expect("Parse failed");
+        let output = d.generate(&ast[0]).expect("Generate failed");
+
+        assert_eq!(output, sql);
+    }
+
+    #[test]
+    fn test_repair_table_commands_roundtrip() {
+        let d = Dialect::get(DialectType::Databricks);
+        let cases = [
+            "REPAIR TABLE events",
+            "MSCK REPAIR TABLE events",
+            "REPAIR TABLE events ADD PARTITIONS",
+            "REPAIR TABLE events DROP PARTITIONS",
+            "REPAIR TABLE events SYNC PARTITIONS",
+            "REPAIR TABLE events SYNC METADATA",
+        ];
+
+        for sql in cases {
+            let ast = d.parse(sql).expect("Parse failed");
+            let output = d.generate(&ast[0]).expect("Generate failed");
+            assert_eq!(output, sql);
+        }
+    }
+
+    #[test]
+    fn test_apply_changes_commands_roundtrip() {
+        let d = Dialect::get(DialectType::Databricks);
+        let cases = [
+            "APPLY CHANGES INTO silver.orders FROM STREAM(bronze.orders) KEYS (id) SEQUENCE BY ts",
+            "APPLY CHANGES INTO LIVE.silver_orders FROM STREAM(LIVE.bronze_orders) KEYS (id) IGNORE NULL UPDATES SEQUENCE BY ts",
+            "APPLY CHANGES INTO LIVE.silver_orders FROM STREAM(LIVE.bronze_orders) KEYS (id) APPLY AS DELETE WHEN operation = 'DELETE' SEQUENCE BY ts COLUMNS * EXCEPT (operation) STORED AS SCD TYPE 1",
+            "APPLY CHANGES INTO LIVE.silver_orders FROM STREAM(LIVE.bronze_orders) KEYS (id) SEQUENCE BY ts STORED AS SCD TYPE 2 TRACK HISTORY ON * EXCEPT (updated_at)",
+            "AUTO CDC INTO silver.orders FROM STREAM(bronze.orders) KEYS (id) SEQUENCE BY ts",
+            "CREATE FLOW apply_cdc AS AUTO CDC INTO silver.orders FROM STREAM(bronze.orders) KEYS (id) SEQUENCE BY ts",
+        ];
+
+        for sql in cases {
+            let ast = d.parse(sql).expect("Parse failed");
+            let output = d.generate(&ast[0]).expect("Generate failed");
+            assert_eq!(output, sql);
+        }
+    }
+
+    #[test]
+    fn test_generate_symlink_format_manifest_roundtrip() {
+        let d = Dialect::get(DialectType::Databricks);
+        let cases = [
+            "GENERATE symlink_format_manifest FOR TABLE events",
+            "GENERATE symlink_format_manifest FOR TABLE catalog.schema.events",
+        ];
+
+        for sql in cases {
+            let ast = d.parse(sql).expect("Parse failed");
+            let output = d.generate(&ast[0]).expect("Generate failed");
+            assert_eq!(output, sql);
+        }
+    }
+
+    #[test]
+    fn test_convert_to_delta_roundtrip() {
+        let d = Dialect::get(DialectType::Databricks);
+        let cases = [
+            "CONVERT TO DELTA parquet.`/mnt/data/events`",
+            "CONVERT TO DELTA database_name.table_name",
+            "CONVERT TO DELTA parquet.`s3://my-bucket/path/to/table` PARTITIONED BY (date DATE)",
+            "CONVERT TO DELTA database_name.table_name NO STATISTICS",
+        ];
+
+        for sql in cases {
+            let ast = d.parse(sql).expect("Parse failed");
+            let output = d.generate(&ast[0]).expect("Generate failed");
+            assert_eq!(output, sql);
+        }
+    }
 }

{polyglot_sql-0.3.2 → polyglot_sql-0.3.4}/crates/polyglot-sql/src/dialects/doris.rs

@@ -5,7 +5,9 @@
 
 use super::{DialectImpl, DialectType};
 use crate::error::Result;
-use crate::expressions::{AggFunc, Case, Cast, Expression, Function, VarArgFunc};
+use crate::expressions::{
+    AggFunc, Case, Cast, Expression, Function, Interval, IntervalUnit, IntervalUnitSpec, VarArgFunc,
+};
 use crate::generator::GeneratorConfig;
 use crate::tokens::TokenizerConfig;
 
@@ -106,6 +108,16 @@ impl DialectImpl for DorisDialect {
 }
 
 impl DorisDialect {
+    fn wrap_day_interval(expr: Expression) -> Expression {
+        Expression::Interval(Box::new(Interval {
+            this: Some(expr),
+            unit: Some(IntervalUnitSpec::Simple {
+                unit: IntervalUnit::Day,
+                use_plural: false,
+            }),
+        }))
+    }
+
     fn transform_function(&self, f: Function) -> Result<Expression> {
         let name_upper = f.name.to_uppercase();
         match name_upper.as_str() {
@@ -207,6 +219,44 @@ impl DorisDialect {
             // DATE_TRUNC is native in Doris
             "DATE_TRUNC" => Ok(Expression::Function(Box::new(f))),
 
+            // Doris normalizes MySQL-style day shorthand to INTERVAL syntax.
+            "DATE_ADD" if f.args.len() == 2 && !matches!(f.args[1], Expression::Interval(_)) => {
+                let mut args = f.args;
+                let date = args.remove(0);
+                let days = args.remove(0);
+                Ok(Expression::Function(Box::new(Function::new(
+                    "DATE_ADD".to_string(),
+                    vec![date, Self::wrap_day_interval(days)],
+                ))))
+            }
+            "DATE_SUB" if f.args.len() == 2 && !matches!(f.args[1], Expression::Interval(_)) => {
+                let mut args = f.args;
+                let date = args.remove(0);
+                let days = args.remove(0);
+                Ok(Expression::Function(Box::new(Function::new(
+                    "DATE_SUB".to_string(),
+                    vec![date, Self::wrap_day_interval(days)],
+                ))))
+            }
+            "ADDDATE" if f.args.len() == 2 && !matches!(f.args[1], Expression::Interval(_)) => {
+                let mut args = f.args;
+                let date = args.remove(0);
+                let days = args.remove(0);
+                Ok(Expression::Function(Box::new(Function::new(
+                    "DATE_ADD".to_string(),
+                    vec![date, Self::wrap_day_interval(days)],
+                ))))
+            }
+            "SUBDATE" if f.args.len() == 2 && !matches!(f.args[1], Expression::Interval(_)) => {
+                let mut args = f.args;
+                let date = args.remove(0);
+                let days = args.remove(0);
+                Ok(Expression::Function(Box::new(Function::new(
+                    "DATE_SUB".to_string(),
+                    vec![date, Self::wrap_day_interval(days)],
+                ))))
+            }
+
             // COLLECT_LIST is native in Doris
             "COLLECT_LIST" => Ok(Expression::Function(Box::new(f))),
 

{polyglot_sql-0.3.2 → polyglot_sql-0.3.4}/crates/polyglot-sql/src/dialects/fabric.rs

@@ -42,6 +42,7 @@ impl DialectImpl for FabricDialect {
             identifier_quote: '[',
             identifier_quote_style: IdentifierQuoteStyle::BRACKET,
             dialect: Some(DialectType::Fabric),
+            null_ordering_supported: false,
             ..Default::default()
         }
     }
@@ -63,7 +64,14 @@ impl DialectImpl for FabricDialect {
                     }
                     _ => {}
                 }
-                // Also transform column data types through Fabric's type mappings
+                // Also transform column data types through Fabric's type mappings.
+                // Apply TSQL normalisation first (e.g. BPCHAR → Char), then Fabric-specific.
+                let tsql = TSQLDialect;
+                if let Ok(Expression::DataType(tsql_dt)) =
+                    tsql.transform_data_type(col.data_type.clone())
+                {
+                    col.data_type = tsql_dt;
+                }
                 if let Expression::DataType(new_dt) =
                     self.transform_fabric_data_type(col.data_type.clone())?
                 {
@@ -361,7 +369,8 @@ impl FabricDialect {
                 let upper = name.to_uppercase();
 
                 // Parse out precision and scale if present: "TYPENAME(n)" or "TYPENAME(n, m)"
-                let (base_name, precision, scale) = Self::parse_type_precision_and_scale(&upper);
+                let (base_name, precision, scale) =
+                    TSQLDialect::parse_type_precision_and_scale(&upper);
 
                 match base_name.as_str() {
                     // DATETIME -> DATETIME2(6)
@@ -518,24 +527,4 @@ impl FabricDialect {
             None => max, // Default to max if not specified
         }
     }
-
-    /// Parse type name and optional precision/scale from strings like "DATETIME2(7)" or "NUMERIC(10, 2)"
-    fn parse_type_precision_and_scale(name: &str) -> (String, Option<u32>, Option<u32>) {
-        if let Some(paren_pos) = name.find('(') {
-            let base = name[..paren_pos].to_string();
-            let rest = &name[paren_pos + 1..];
-            if let Some(close_pos) = rest.find(')') {
-                let args = &rest[..close_pos];
-                let parts: Vec<&str> = args.split(',').map(|s| s.trim()).collect();
-
-                let precision = parts.first().and_then(|s| s.parse::<u32>().ok());
-                let scale = parts.get(1).and_then(|s| s.parse::<u32>().ok());
-
-                return (base, precision, scale);
-            }
-            (base, None, None)
-        } else {
-            (name.to_string(), None, None)
-        }
-    }
 }

{polyglot_sql-0.3.2 → polyglot_sql-0.3.4}/crates/polyglot-sql/src/dialects/hive.rs

@@ -42,6 +42,12 @@ impl DialectImpl for HiveDialect {
         config
             .keywords
             .insert("DIV".to_string(), crate::tokens::TokenType::Div);
+        config
+            .keywords
+            .insert("REPAIR".to_string(), crate::tokens::TokenType::Command);
+        config
+            .keywords
+            .insert("MSCK".to_string(), crate::tokens::TokenType::Command);
         // Hive numeric literal suffixes: 1L -> BIGINT, 1S -> SMALLINT, etc.
         config
             .numeric_literals