fix: rename columns and tz localization for empty dataframes

Author: silvanmurre

src/enappsys/__init__.py

@@ -3,7 +3,7 @@
 to EnAppSys' API services.
 """
 
-__version__ = "0.1.7"
+__version__ = "0.1.8"
 
 from enappsys.client import EnAppSys as EnAppSys
 from enappsys.client_async import EnAppSysAsync as EnAppSysAsync

src/enappsys/services/bulk.py

@@ -97,34 +97,34 @@ def to_df(
             return pd.DataFrame()
 
         df.index.name = "dateTime"
+        step_size = 1
+        if self.min_avg_max:
+            step_size = 3
+
         if df.empty:
             _warn_empty_response("csv", url=self.url, params=self.params)
-            if remove_units_column and len(df.columns) > 0:
-                df = df.iloc[:, 1:]
-            return df
-
+            df.index = pd.DatetimeIndex([], name="dateTime")
+        
         if tz_localize:
             df.index = df.index.tz_localize(self.time_zone, ambiguous="infer")
 
-        step_size = 1
-        if self.min_avg_max:
-            step_size = 3
-
         if rename_columns or data_type_in_columns or unit_in_columns:
             columns = list(df.columns)
 
             if isinstance(rename_columns, list):
                 # [:1] to not include Units when comparing with rename_columns
                 validate_rename_columns_length(rename_columns, columns[1:], step_size)
 
-            # Units are in first column, for every row, so take first
-            units_colon_delimited = df.iloc[0, 0]
-            if not pd.isnull(units_colon_delimited):
-                units = units_colon_delimited.split(":")
-                if len(units) == 1:
-                    units = units * ((len(columns) - 1) // step_size)
-            else:
-                units = [""] * ((len(columns) - 1) // step_size)
+            units = None
+            if not df.empty:
+                # Units are in first column, for every row, so take first
+                units_colon_delimited = df.iloc[0, 0]
+                if not pd.isnull(units_colon_delimited):
+                    units = units_colon_delimited.split(":")
+                    if len(units) == 1:
+                        units = units * ((len(columns) - 1) // step_size)
+                else:
+                    units = [""] * ((len(columns) - 1) // step_size)
 
             # Add datatype to "Units" column
             if data_type_in_columns:
@@ -140,7 +140,7 @@ def to_df(
                     column_name = rename_columns[column_name]
                 if data_type_in_columns:
                     column_name = f"{self.data_type}.{column_name}"
-                if unit_in_columns:
+                if unit_in_columns and units is not None:
                     column_name = f"{column_name} ({units[(idx - 1) // step_size]})"
 
                 if self.min_avg_max:

src/enappsys/services/chart.py

@@ -90,7 +90,8 @@ def to_df(
         df.index.name = "dateTime"
         if df.empty:
             _warn_empty_response("csv", url=self.url, params=self.params)
-        elif tz_localize:
+            df.index = pd.DatetimeIndex([], name="dateTime")
+        if tz_localize:
             df.index = df.index.tz_localize(self.time_zone, ambiguous="infer")
 
         step_size = 1

tests/unit/test_bulk_api.py

@@ -4,6 +4,8 @@
 
 import logging
 
+import pandas as pd
+
 from enappsys.services.bulk import BulkCSV, BulkJSON, BulkJSONMap
 from enappsys.services.chart import ChartCSV, ChartJSON, ChartJSONMap
 
@@ -77,6 +79,8 @@ def test_empty_bulk_csv_to_df_returns_empty_dataframe(caplog):
         df = out.to_df()
 
     assert df.empty
+    assert isinstance(df.index, pd.DatetimeIndex)
+    assert str(df.index.tz) == "UTC"
     assert list(df.columns) == ["BE.BELGIUM"]
     assert "returning empty DataFrame" in caplog.text
     assert "url=bulk" in caplog.text
@@ -87,6 +91,33 @@ def test_empty_bulk_csv_to_df_returns_empty_dataframe(caplog):
     assert "test-secret" not in caplog.text
 
 
+def test_empty_bulk_csv_to_df_applies_rename_columns(caplog):
+    out = BulkCSV(
+        "Datetime,Units,BE.BELGIUM\n",
+        "bulk",
+        {},
+        "csv",
+        "ENTSOE_DAY_AHEAD_PRICES",
+        ["BE.BELGIUM"],
+        None,
+        None,
+        "hourly",
+        "UTC",
+        False,
+    )
+
+    with caplog.at_level(logging.WARNING):
+        df = out.to_df(
+            data_type_in_columns=True,
+            rename_columns=["be_dap"],
+        )
+
+    assert df.empty
+    assert isinstance(df.index, pd.DatetimeIndex)
+    assert str(df.index.tz) == "UTC"
+    assert list(df.columns) == ["ENTSOE_DAY_AHEAD_PRICES.be_dap"]
+
+
 def test_bulk_empty_warning_uses_full_url(client: "EnAppSys", monkeypatch, caplog):
     def fake_get(url, params):
         assert url == "csvapi"
@@ -135,13 +166,42 @@ def test_empty_chart_csv_to_df_returns_empty_dataframe(caplog):
         df = out.to_df()
 
     assert df.empty
+    assert isinstance(df.index, pd.DatetimeIndex)
+    assert str(df.index.tz) == "UTC"
     assert list(df.columns) == ["BE.BELGIUM"]
     assert "returning empty DataFrame" in caplog.text
     assert "url=datadownload" in caplog.text
     assert "day-ahead-prices" in caplog.text
     assert "202301010300" in caplog.text
 
 
+def test_empty_chart_csv_to_df_applies_rename_columns(caplog):
+    out = ChartCSV(
+        "Datetime,BE.BELGIUM\n,EUR/MWh\n",
+        "datadownload",
+        {},
+        "csv",
+        "day-ahead-prices",
+        None,
+        None,
+        "hourly",
+        "UTC",
+        "EUR",
+        False,
+    )
+
+    with caplog.at_level(logging.WARNING):
+        df = out.to_df(
+            rename_columns=["be_dap"],
+            unit_in_columns=True,
+        )
+
+    assert df.empty
+    assert isinstance(df.index, pd.DatetimeIndex)
+    assert str(df.index.tz) == "UTC"
+    assert list(df.columns) == ["be_dap (EUR/MWh)"]
+
+
 def test_chart_empty_warning_uses_full_url(client: "EnAppSys", monkeypatch, caplog):
     def fake_get(url, params):
         assert url == "datadownload"