CLN: Warn that default dialect is changing to "standard" (#196)

* CLN: Warn that default dialect is changing to "standard"

Towards issue #195

* Increment stacklevel so line where read_gbq is called is highlighted.

* Update pyenv

* Try Python 2.7.15 via Travis build instead of pyenv

* Add to changelog

* Remove unused TRAVIS_PYTHON_VERSION env var.

Author: tswast

.travis.yml

@@ -1,22 +1,26 @@
-sudo: false
-
 language: python
 
+matrix:
+  include:
+  - os: linux
+    python: 2.7
+    env: PYTHON=2.7 PANDAS=0.19.2 COVERAGE='false' LINT='true'
+  - os: linux
+    python: 3.5
+    env: PYTHON=3.5 PANDAS=0.18.1 COVERAGE='true' LINT='false'
+  - os: linux
+    python: 3.6
+    env: PYTHON=3.6 PANDAS=0.20.1 COVERAGE='false' LINT='false'
+  - os: linux
+    python: 3.6
+    env: PYTHON=3.6 PANDAS=MASTER COVERAGE='false' LINT='true'
 env:
-  - PYTHON=2.7 PANDAS=0.19.2 COVERAGE='false' LINT='true' PYENV_VERSION=2.7.14
-  - PYTHON=3.5 PANDAS=0.18.1 COVERAGE='true' LINT='false' PYENV_VERSION=3.5.4
-  - PYTHON=3.6 PANDAS=0.20.1 COVERAGE='false' LINT='false' PYENV_VERSION=3.6.1
-  - PYTHON=3.6 PANDAS=MASTER COVERAGE='false' LINT='true' PYENV_VERSION=3.6.1
 
 before_install:
   - echo "before_install"
   - source ci/travis_process_gbq_encryption.sh
 
 install:
-  # work around https://github.com/travis-ci/travis-ci/issues/8363
-  # https://github.com/pre-commit/pre-commit/commit/e3ab8902692e896da9ded42bd4d76ea4e1de359d
-  - pyenv install -s $PYENV_VERSION
-  - pyenv global system $PYENV_VERSION
   # Upgrade setuptools and pip to work around
   # https://github.com/pypa/setuptools/issues/885
   - pip install --upgrade setuptools

docs/source/changelog.rst

@@ -1,11 +1,14 @@
 Changelog
 =========
 
-.. _changelog-0.5.1:
+.. _changelog-0.6.0:
 
-0.5.1 / (Unreleased)
+0.6.0 / 2018-08-15
 --------------------
 
+- Warn when ``dialect`` is not passed in to ``read_gbq``. The default dialect
+  will be changing from 'legacy' to 'standard' in a future version.
+  (:issue:`195`)
 - Use general float with 15 decimal digit precision when writing to local
   CSV buffer in ``to_gbq``. This prevents numerical overflow in certain
   edge cases. (:issue:`192`)

pandas_gbq/gbq.py

@@ -471,7 +471,7 @@ def _parse_data(schema, rows):
 
 def read_gbq(query, project_id=None, index_col=None, col_order=None,
              reauth=False, private_key=None, auth_local_webserver=False,
-             dialect='legacy', location=None, configuration=None,
+             dialect=None, location=None, configuration=None,
              verbose=None):
     r"""Load data from Google BigQuery using google-cloud-python
 
@@ -515,6 +515,8 @@ def read_gbq(query, project_id=None, index_col=None, col_order=None,
 
         .. versionadded:: 0.2.0
     dialect : str, default 'legacy'
+        Note: The default value is changing to 'standard' in a future verion.
+
         SQL syntax dialect to use. Value can be one of:
 
         ``'legacy'``
@@ -552,14 +554,21 @@ def read_gbq(query, project_id=None, index_col=None, col_order=None,
     df: DataFrame
         DataFrame representing results of query.
     """
+    if dialect is None:
+        dialect = 'legacy'
+        warnings.warn(
+            'The default value for dialect is changing to "standard" in a '
+            'future version. Pass in dialect="legacy" to disable this '
+            'warning.',
+            FutureWarning, stacklevel=2)
 
     _test_google_api_imports()
 
     if verbose is not None and SHOW_VERBOSE_DEPRECATION:
         warnings.warn(
             "verbose is deprecated and will be removed in "
             "a future version. Set logging level in order to vary "
-            "verbosity", FutureWarning, stacklevel=1)
+            "verbosity", FutureWarning, stacklevel=2)
 
     if dialect not in ('legacy', 'standard'):
         raise ValueError("'{0}' is not valid for dialect".format(dialect))