Parse column names in a dataframe to avoid breaking the line protocol (influxdata#584)

dmuiruri · xginn8 · commit ed4561b8d31a · 2018-05-07T20:48:49.000-04:00
* Remove comment as issues have been resolved

* Parse column names in a dataframe to handle spaces in tag of field keys

* Patch for ERROR: Test failed write points from df with series

* Patch for flake8 issues(E231, E501)

* Test case for spaces in column names

* flake8 issues

* flake8 issues: trailing wspace

* Testing if test case would catch a regression

* Test catches a regressed build

* Re-run the build
diff --git a/influxdb/_dataframe_client.py b/influxdb/_dataframe_client.py
@@ -288,6 +288,8 @@ def _convert_dataframe_to_lines(self,
             raise TypeError('Must be DataFrame with DatetimeIndex or '
                             'PeriodIndex.')
 
+        dataframe = dataframe.rename(
+            columns={item: _escape_tag(item) for item in dataframe.columns})
         # Create a Series of columns for easier indexing
         column_series = pd.Series(dataframe.columns)
 
diff --git a/influxdb/tests/dataframe_client_test.py b/influxdb/tests/dataframe_client_test.py
@@ -81,6 +81,26 @@ def test_dataframe_write_points_with_whitespace_measurement(self):
             cli.write_points(dataframe, 'meas with space')
             self.assertEqual(m.last_request.body, expected)
 
+    def test_dataframe_write_points_with_whitespace_in_column_names(self):
+        """write_points should escape white space in column names."""
+        now = pd.Timestamp('1970-01-01 00:00+00:00')
+        dataframe = pd.DataFrame(data=[["1", 1, 1.0], ["2", 2, 2.0]],
+                                 index=[now, now + timedelta(hours=1)],
+                                 columns=["column one", "column two",
+                                          "column three"])
+        expected = (
+            b"foo column\\ one=\"1\",column\\ two=1i,column\\ three=1.0 0\n"
+            b"foo column\\ one=\"2\",column\\ two=2i,column\\ three=2.0 "
+            b"3600000000000\n"
+        )
+        with requests_mock.Mocker() as m:
+            m.register_uri(requests_mock.POST,
+                           "http://localhost:8086/write",
+                           status_code=204)
+            cli = DataFrameClient(database='db')
+            cli.write_points(dataframe, 'foo')
+            self.assertEqual(m.last_request.body, expected)
+
     def test_write_points_from_dataframe_with_none(self):
         """Test write points from df in TestDataFrameClient object."""
         now = pd.Timestamp('1970-01-01 00:00+00:00')
