fix(waterdata): keep the parent statistic in statistics results

NEWS.md

@@ -1,3 +1,5 @@
+**11/14/2026:** **Bug fix:** `waterdata.get_stats_por()` and `waterdata.get_stats_date_range()` now return the `parent_statistic_id` and `parent_statistic_name` columns that the Statistics API sends with each series. A parameter can have several daily series at one monitoring location, such as the daily maximum, minimum and mean of discharge, and each has its own statistics. The two columns were dropped, so the rows computed from the different daily series could not be told apart without looking up each `parent_time_series_id`.
+
 **10/29/2026:** `waterdata.get_latest_continuous()` accepts `method_category`, a column USGS added to the `latest-continuous` collection in October 2026; `get_continuous()` has accepted it since 09/22/2026. It is the RLMS method category code (`STNRD`, `LMTUS`, `EXPER` or `UNKWN`) for the method in effect over an observation's interval, returned on every record and null for time series that have not been categorized. It could already be passed through `**queryables`; it is now a documented, keyword-only parameter.
 
 **10/16/2026:** **Bug fix:** the time-zone table behind the `*DateTime` columns that `waterdata.get_samples()` and `wqp.get_results()` append, and behind `utils.format_datetime()`, now follows the code lists those services use: the USGS NWIS time-zone list (Samples, and the USGS records that the Water Quality Portal passes through) and the WQX TimeZone domain (Water Quality Portal). USGS records write American Samoa as `ZP-11`, which was missing, so every timed American Samoa row from `get_samples()`, and every timed USGS American Samoa row from `get_results()`, had a `NaT` datetime; it is now UTC-11. 39 other codes from the two lists that used to become `NaT` now resolve, among them the remaining `ZP*` offsets, `IDLW`, `IDLE`, `NZT`, `SWT`, and the retired WQX codes `AHST` and `YST`. **Behavior change:** codes that the table read with a different meaning from the USGS list now take the USGS meaning: `CAST` (+9:30), `CCT` (+8), `EASST` (+11), `EAST` (+10), `NFT` (-3:30), `NZDT` (+13), `WAST` (+7), and `WAT` (-1). `FJST` is now UTC+13. Where the two lists disagree, WQX wins: `SST` stays American Samoa (UTC-11), the meaning the Water Quality Portal returns, and `BST` is UTC-11.

dataretrieval/waterdata/stats.py

@@ -82,6 +82,8 @@ def _handle_nesting(
             ["features", "properties", "data", "parameter_code"],
             ["features", "properties", "data", "unit_of_measure"],
             ["features", "properties", "data", "parent_time_series_id"],
+            ["features", "properties", "data", "parent_statistic_id"],
+            ["features", "properties", "data", "parent_statistic_name"],
         ],
         meta_prefix="",
         errors="ignore",

tests/waterdata_test.py

@@ -1845,6 +1845,9 @@ def test_get_stats_por(httpx_mock):
     assert len(df) == 11
     assert df.loc[df["computation"] == "minimum", "percentile"].tolist() == [0.0]
     assert df.loc[df["computation"] == "arithmetic_mean", "percentile"].isna().all()
+    # Every row names the daily statistic its normals are computed from.
+    assert (df["parent_statistic_id"] == "00003").all()
+    assert (df["parent_statistic_name"] == "Mean").all()
 
 
 def test_get_stats_por_expanded_false(httpx_mock):

tests/waterdata_utils_test.py

@@ -582,6 +582,54 @@ def test_handle_nesting_tolerates_missing_drop_columns():
     assert df["monitoring_location_id"].iloc[0] == "USGS-12345"
 
 
+def test_handle_nesting_keeps_parent_statistic():
+    """A parameter can have several daily series (max, min, mean), each with
+    its own normals. The parent statistic is what tells those rows apart, so
+    it is kept beside ``parent_time_series_id``.
+    """
+    body = {
+        "next": None,
+        "features": [
+            {
+                "properties": {
+                    "monitoring_location_id": "USGS-12345",
+                    "data": [
+                        {
+                            "parameter_code": "00060",
+                            "unit_of_measure": "ft^3/s",
+                            "parent_time_series_id": ts,
+                            "parent_statistic_id": stat_id,
+                            "parent_statistic_name": stat_name,
+                            "values": [
+                                {
+                                    "time_of_year": "07-01",
+                                    "computation": "median",
+                                    "value": value,
+                                }
+                            ],
+                        }
+                        for ts, stat_id, stat_name, value in [
+                            ("ts-max", "00001", "Max", "373.5"),
+                            ("ts-min", "00002", "Min", "267.0"),
+                            ("ts-mean", "00003", "Mean", "320.0"),
+                        ]
+                    ],
+                },
+            }
+        ],
+    }
+
+    df = _handle_nesting(body, geopd=False)
+
+    assert df[
+        ["parent_statistic_id", "parent_statistic_name", "value"]
+    ].to_numpy().tolist() == [
+        ["00001", "Max", "373.5"],
+        ["00002", "Min", "267.0"],
+        ["00003", "Mean", "320.0"],
+    ]
+
+
 def test_handle_nesting_returns_empty_on_empty_features():
     """A mid-pagination empty page ({\"features\": [], \"next\": <tok>})
     must not fail the downstream merge with