fix: correctly handle unbound varchar case instead of defaulting to varchar(256)

.changes/unreleased/Fixes-20221209-114403.yaml

@@ -0,0 +1,7 @@
+kind: Fixes
+body: Handle unbound varchar data_type instead of defaulting to varchar(256)
+time: 2022-12-09T11:44:03.689108-08:00
+custom:
+  Author: kalvinnchau
+  Issue: "35"
+  PR: "194"

dbt/adapters/trino/column.py

@@ -5,6 +5,10 @@
 from dbt.adapters.base.column import Column
 from dbt.exceptions import RuntimeException
 
+# Taken from the MAX_LENGTH variable in
+# https://github.com/trinodb/trino/blob/master/core/trino-spi/src/main/java/io/trino/spi/type/VarcharType.java
+TRINO_VARCHAR_MAX_LENGTH = 2147483646
+
 
 @dataclass
 class TrinoColumn(Column):
@@ -15,10 +19,26 @@ class TrinoColumn(Column):
         "INTEGER": "INT",
     }
 
+    @property
+    def data_type(self):
+        # when varchar has no defined size, default to unbound varchar
+        # the super().data_type defaults to varchar(256)
+        if self.dtype.lower() == "varchar" and self.char_size is None:
+            return self.dtype
+
+        return super().data_type
+
     @classmethod
     def string_type(cls, size: int) -> str:
         return "varchar({})".format(size)
 
+    def string_size(self) -> int:
+        # override the string_size function to handle the unbound varchar case
+        if self.dtype.lower() == "varchar" and self.char_size is None:
+            return TRINO_VARCHAR_MAX_LENGTH
+
+        return super().string_size()
+
     @classmethod
     def from_description(cls, name: str, raw_data_type: str) -> "Column":
         # some of the Trino data types specify a type and not a precision

tests/unit/test_adapter.py

@@ -10,6 +10,7 @@
 from dbt.exceptions import DatabaseException, FailedToConnectException, RuntimeException
 
 from dbt.adapters.trino import TrinoAdapter
+from dbt.adapters.trino.column import TRINO_VARCHAR_MAX_LENGTH, TrinoColumn
 from dbt.adapters.trino.connections import (
     HttpScheme,
     TrinoCertificateCredentials,
@@ -530,3 +531,28 @@ def test_convert_date_type(self):
         expected = ["DATE", "DATE", "DATE"]
         for col_idx, expect in enumerate(expected):
             assert TrinoAdapter.convert_date_type(agate_table, col_idx) == expect
+
+
+class TestTrinoColumn(unittest.TestCase):
+    def test_bound_varchar(self):
+        col = TrinoColumn.from_description("my_col", "VARCHAR(100)")
+        assert col.column == "my_col"
+        assert col.dtype == "VARCHAR"
+        assert col.char_size == 100
+        # bounded varchars get formatted to lowercase
+        assert col.data_type == "varchar(100)"
+        assert col.string_size() == 100
+        assert col.is_string() is True
+        assert col.is_number() is False
+        assert col.is_numeric() is False
+
+    def test_unbound_varchar(self):
+        col = TrinoColumn.from_description("my_col", "VARCHAR")
+        assert col.column == "my_col"
+        assert col.dtype == "VARCHAR"
+        assert col.char_size is None
+        assert col.data_type == "VARCHAR"
+        assert col.string_size() == TRINO_VARCHAR_MAX_LENGTH
+        assert col.is_string() is True
+        assert col.is_number() is False
+        assert col.is_numeric() is False