fix(bigquery): support LOAD DATA FROM FILES syntax

sqlglot/expressions/dml.py

@@ -282,7 +282,8 @@ class LoadData(Expression):
         "this": True,
         "local": False,
         "overwrite": False,
-        "inpath": True,
+        "inpath": False,
+        "files": False,
         "partition": False,
         "input_format": False,
         "serde": False,

sqlglot/generators/bigquery.py

@@ -560,6 +560,17 @@ def mod_sql(self, expression: exp.Mod) -> str:
             expr.unnest() if isinstance(expr, exp.Paren) else expr,
         )
 
+    def loaddata_sql(self, expression: exp.LoadData) -> str:
+        files = expression.args.get("files")
+        if files:
+            overwrite = " OVERWRITE" if expression.args.get("overwrite") else ""
+            this = self.sql(expression, "this")
+            table = f" {this}" if expression.args.get("overwrite") else f" INTO TABLE {this}"
+            files_sql = self.func("FILES", *files.expressions)
+            return f"LOAD DATA{overwrite}{table} FROM {files_sql}"
+
+        return super().loaddata_sql(expression)
+
     def column_parts(self, expression: exp.Column) -> str:
         if expression.meta.get("quoted_column"):
             # If a column reference is of the form `dataset.table`.name, we need

sqlglot/parsers/bigquery.py

@@ -706,6 +706,27 @@ def _parse_export_data(self) -> exp.Export:
             )
         )
 
+    def _parse_load(self) -> exp.LoadData | exp.Command:
+        index = self._index
+
+        if self._match_text_seq("DATA"):
+            overwrite = self._match(TokenType.OVERWRITE)
+            if not overwrite:
+                self._match_pair(TokenType.INTO, TokenType.TABLE)
+
+            this = self._parse_table(schema=True)
+            if this and self._match_text_seq("FROM", "FILES"):
+                return self.expression(
+                    exp.LoadData(
+                        this=this,
+                        overwrite=overwrite,
+                        files=exp.Properties(expressions=self._parse_wrapped_properties()),
+                    )
+                )
+
+        self._retreat(index)
+        return super()._parse_load()
+
     def _parse_column_ops(self, this: exp.Expr | None) -> exp.Expr | None:
         func_index = self._index + 1
         this = super()._parse_column_ops(this)

tests/dialects/test_bigquery.py

@@ -206,6 +206,14 @@ def test_bigquery(self):
         self.validate_identity("BEGIN TRANSACTION")
         self.validate_identity("COMMIT TRANSACTION")
         self.validate_identity("ROLLBACK TRANSACTION")
+        self.validate_identity(
+            "LOAD DATA OVERWRITE mydataset.table1 FROM FILES(format='AVRO', uris=['gs://bucket/path/file.avro'])",
+            "LOAD DATA OVERWRITE mydataset.table1 FROM FILES(FORMAT='AVRO', uris=['gs://bucket/path/file.avro'])",
+        ).assert_is(exp.LoadData)
+        self.validate_identity(
+            "LOAD DATA INTO TABLE mydataset.table1 FROM FILES(format='AVRO', uris=['gs://bucket/path/file.avro'])",
+            "LOAD DATA INTO TABLE mydataset.table1 FROM FILES(FORMAT='AVRO', uris=['gs://bucket/path/file.avro'])",
+        )
         self.validate_identity("CAST(x AS BIGNUMERIC)")
         self.validate_identity("SELECT y + 1 FROM x GROUP BY y + 1 ORDER BY 1")
         self.validate_identity("SELECT TIMESTAMP_SECONDS(2) AS t")