BUG: resolve divide by 0 error when uploading empty dataframe (#252)

* resolve divide by 0 error when uploading empty dataframe

* reformat with black

* add unit test when uploading empty dataframe

* add empty data upload system test

* remove empty df unit test

* update empty df

* add 0.10.0 release note

* update release note version number to 0.11.0

* update empty dataframe bug fix in change log

Author: bwanglzu

docs/source/changelog.rst

@@ -6,6 +6,8 @@ Changelog
 0.10.0 / TBD
 ------------
 
+- This fixes a bug where pandas-gbq could not upload an empty database. (:issue:`237`)
+
 Dependency updates
 ~~~~~~~~~~~~~~~~~~
 
@@ -235,4 +237,4 @@ Initial release of transfered code from `pandas <https://github.com/pandas-dev/p
 Includes patches since the 0.19.2 release on pandas with the following:
 
 - :func:`read_gbq` now allows query configuration preferences `pandas-GH#14742 <https://github.com/pandas-dev/pandas/pull/14742>`__
-- :func:`read_gbq` now stores ``INTEGER`` columns as ``dtype=object`` if they contain ``NULL`` values. Otherwise they are stored as ``int64``. This prevents precision lost for integers greather than 2**53. Furthermore ``FLOAT`` columns with values above 10**4 are no longer casted to ``int64`` which also caused precision loss `pandas-GH#14064 <https://github.com/pandas-dev/pandas/pull/14064>`__, and `pandas-GH#14305 <https://github.com/pandas-dev/pandas/pull/14305>`__
+- :func:`read_gbq` now stores ``INTEGER`` columns as ``dtype=object`` if they contain ``NULL`` values. Otherwise they are stored as ``int64``. This prevents precision lost for integers greather than 2**53. Furthermore ``FLOAT`` columns with values above 10**4 are no longer casted to ``int64`` which also caused precision loss `pandas-GH#14064 <https://github.com/pandas-dev/pandas/pull/14064>`__, and `pandas-GH#14305 <https://github.com/pandas-dev/pandas/pull/14305>`__

pandas_gbq/gbq.py

@@ -518,8 +518,8 @@ def load_data(
                 chunks = tqdm.tqdm(chunks)
             for remaining_rows in chunks:
                 logger.info(
-                    "\rLoad is {0}% Complete".format(
-                        ((total_rows - remaining_rows) * 100) / total_rows
+                    "\r{} out of {} rows loaded.".format(
+                        total_rows - remaining_rows, total_rows
                     )
                 )
         except self.http_error as ex:

tests/system/test_gbq.py

@@ -924,6 +924,28 @@ def test_upload_data(self, project_id):
         )
         assert result["num_rows"][0] == test_size
 
+    def test_upload_empty_data(self, project_id):
+        test_id = "data_with_0_rows"
+        test_size = 0
+        df = DataFrame()
+
+        gbq.to_gbq(
+            df,
+            self.destination_table + test_id,
+            project_id,
+            credentials=self.credentials,
+        )
+
+        result = gbq.read_gbq(
+            "SELECT COUNT(*) AS num_rows FROM {0}".format(
+                self.destination_table + test_id
+            ),
+            project_id=project_id,
+            credentials=self.credentials,
+            dialect="legacy",
+        )
+        assert result["num_rows"][0] == test_size
+
     def test_upload_data_if_table_exists_fail(self, project_id):
         test_id = "2"
         test_size = 10