Fix/bigquery job label length (#3703)

* add blueprints to resolve issue

* revert to previous version

* intentionally failing test

* add imports

* add validation in existing function

* add passing test for length validation

* add current sanitized label

* remove duplicate var

* Make logging output 2 lines

Co-authored-by: Jeremy Cohen <jeremy@fishtownanalytics.com>

* Raise RuntimeException to better handle error

Co-authored-by: Jeremy Cohen <jeremy@fishtownanalytics.com>

* update test

* fix flake8 errors

* update changelog

Co-authored-by: Jeremy Cohen <jeremy@fishtownanalytics.com>

CHANGELOG.md

@@ -6,6 +6,7 @@
 ### Under the hood
 - Switch to full reparse on partial parsing exceptions. Log and report exception information. ([#3725](https://github.com/dbt-labs/dbt/issues/3725), [#3733](https://github.com/dbt-labs/dbt/pull/3733))
 - Check for existence of test node when removing. ([#3711](https://github.com/dbt-labs/dbt/issues/3711), [#3750](https://github.com/dbt-labs/dbt/pull/3750))
+- Better error handling for BigQuery job labels that are too long. [#3703](https://github.com/dbt-labs/dbt/pull/3703)
 
 
 ## dbt 0.20.1 (August 11, 2021)

plugins/bigquery/dbt/adapters/bigquery/connections.py

@@ -595,9 +595,20 @@ def _is_retryable(error):
 
 _SANITIZE_LABEL_PATTERN = re.compile(r"[^a-z0-9_-]")
 
+_VALIDATE_LABEL_LENGTH_LIMIT = 63
+
 
 def _sanitize_label(value: str) -> str:
     """Return a legal value for a BigQuery label."""
     value = value.strip().lower()
     value = _SANITIZE_LABEL_PATTERN.sub("_", value)
-    return value
+    value_length = len(value)
+    if value_length > _VALIDATE_LABEL_LENGTH_LIMIT:
+        error_msg = (
+            f"Job label length {value_length} is greater than length limit: "
+            f"{_VALIDATE_LABEL_LENGTH_LIMIT}\n"
+            f"Current sanitized label: {value}"
+        )
+        raise RuntimeException(error_msg)
+    else:
+        return value

test/unit/test_bigquery_adapter.py

@@ -1,6 +1,8 @@
 import agate
 import decimal
 import json
+import string
+import random
 import re
 import pytest
 import unittest
@@ -17,7 +19,7 @@
 from dbt.adapters.bigquery import BigQueryRelation
 from dbt.adapters.bigquery import Plugin as BigQueryPlugin
 from dbt.adapters.bigquery.connections import BigQueryConnectionManager
-from dbt.adapters.bigquery.connections import _sanitize_label
+from dbt.adapters.bigquery.connections import _sanitize_label, _VALIDATE_LABEL_LENGTH_LIMIT
 from dbt.adapters.base.query_headers import MacroQueryStringSetter
 from dbt.clients import agate_helper
 import dbt.exceptions
@@ -972,3 +974,22 @@ def test_convert_time_type(self):
 )
 def test_sanitize_label(input, output):
     assert _sanitize_label(input) == output
+
+
+@pytest.mark.parametrize(
+    "label_length",
+    [64, 65, 100],
+)
+def test_sanitize_label_length(label_length):
+    random_string = "".join(
+        random.choice(string.ascii_uppercase + string.digits)
+        for i in range(label_length)
+    )
+    test_error_msg = (
+            f"Job label length {label_length} is greater than length limit: "
+            f"{_VALIDATE_LABEL_LENGTH_LIMIT}\n"
+            f"Current sanitized label: {random_string.lower()}"
+        )
+    with pytest.raises(dbt.exceptions.RuntimeException) as error_info:
+        _sanitize_label(random_string)
+    assert error_info.value.args[0] == test_error_msg