feat: Add null_markers property to LoadJobConfig and CSVOptions

Author: google-labs-jules[bot]

google/cloud/bigquery/external_config.py

@@ -474,6 +474,19 @@ def skip_leading_rows(self):
     def skip_leading_rows(self, value):
         self._properties["skipLeadingRows"] = str(value)
 
+    @property
+    def null_markers(self) -> Optional[Iterable[str]]:
+        """Optional[Iterable[str]]: The strings that are interpreted as NULL values.
+
+        See
+        https://cloud.google.com/bigquery/docs/reference/rest/v2/tables#CsvOptions.FIELDS.null_markers
+        """
+        return self._properties.get("nullMarkers")
+
+    @null_markers.setter
+    def null_markers(self, value: Optional[Iterable[str]]):
+        self._properties["nullMarkers"] = value
+
     def to_api_repr(self) -> dict:
         """Build an API representation of this object.
 

google/cloud/bigquery/job/load.py

@@ -386,6 +386,19 @@ def null_marker(self):
     def null_marker(self, value):
         self._set_sub_prop("nullMarker", value)
 
+    @property
+    def null_markers(self) -> Optional[List[str]]:
+        """Optional[List[str]]: Represents a null value (CSV only).
+
+        See:
+        https://cloud.google.com/bigquery/docs/reference/rest/v2/Job#JobConfigurationLoad.FIELDS.null_markers
+        """
+        return self._get_sub_prop("nullMarkers")
+
+    @null_markers.setter
+    def null_markers(self, value: Optional[List[str]]):
+        self._set_sub_prop("nullMarkers", value)
+
     @property
     def preserve_ascii_control_characters(self):
         """Optional[bool]: Preserves the embedded ASCII control characters when sourceFormat is set to CSV.
@@ -854,6 +867,13 @@ def null_marker(self):
         """
         return self.configuration.null_marker
 
+    @property
+    def null_markers(self):
+        """See
+        :attr:`google.cloud.bigquery.job.LoadJobConfig.null_markers`.
+        """
+        return self.configuration.null_markers
+
     @property
     def quote_character(self):
         """See

tests/unit/job/test_load.py

@@ -140,6 +140,10 @@ def _verifyResourceProperties(self, job, resource):
             self.assertEqual(job.null_marker, config["nullMarker"])
         else:
             self.assertIsNone(job.null_marker)
+        if "nullMarkers" in config:
+            self.assertEqual(job.null_markers, config["nullMarkers"])
+        else:
+            self.assertIsNone(job.null_markers)
         if "quote" in config:
             self.assertEqual(job.quote_character, config["quote"])
         else:

tests/unit/job/test_load_config.py

@@ -469,6 +469,22 @@ def test_null_marker_setter(self):
         config.null_marker = null_marker
         self.assertEqual(config._properties["load"]["nullMarker"], null_marker)
 
+    def test_null_markers_missing(self):
+        config = self._get_target_class()()
+        self.assertIsNone(config.null_markers)
+
+    def test_null_markers_hit(self):
+        null_markers = ["", "NA"]
+        config = self._get_target_class()()
+        config._properties["load"]["nullMarkers"] = null_markers
+        self.assertEqual(config.null_markers, null_markers)
+
+    def test_null_markers_setter(self):
+        null_markers = ["", "NA"]
+        config = self._get_target_class()()
+        config.null_markers = null_markers
+        self.assertEqual(config._properties["load"]["nullMarkers"], null_markers)
+
     def test_preserve_ascii_control_characters_missing(self):
         config = self._get_target_class()()
         self.assertIsNone(config.preserve_ascii_control_characters)

tests/unit/test_external_config.py

@@ -277,6 +277,7 @@ def test_from_api_repr_csv(self):
                     "allowJaggedRows": False,
                     "encoding": "encoding",
                     "preserveAsciiControlCharacters": False,
+                    "nullMarkers": ["", "NA"],
                 },
             },
         )
@@ -293,6 +294,7 @@ def test_from_api_repr_csv(self):
         self.assertEqual(ec.options.allow_jagged_rows, False)
         self.assertEqual(ec.options.encoding, "encoding")
         self.assertEqual(ec.options.preserve_ascii_control_characters, False)
+        self.assertEqual(ec.options.null_markers, ["", "NA"])
 
         got_resource = ec.to_api_repr()
 
@@ -314,6 +316,7 @@ def test_to_api_repr_csv(self):
         options.skip_leading_rows = 123
         options.allow_jagged_rows = False
         options.preserve_ascii_control_characters = False
+        options.null_markers = ["", "NA"]
         ec.csv_options = options
 
         exp_resource = {
@@ -326,6 +329,7 @@ def test_to_api_repr_csv(self):
                 "allowJaggedRows": False,
                 "encoding": "encoding",
                 "preserveAsciiControlCharacters": False,
+                "nullMarkers": ["", "NA"],
             },
         }