Show combined patient series (#2318)

* Upgrade related_patient_columns_to_records to do a "full join" * Add failing test * fixes #2301 --------- Co-authored-by: Dave Evans <[email protected]>
opensafely-core · Dec 13, 2024 · 932e6d5 · 932e6d5
1 parent 9ca5a7e
commit 932e6d5
Show file tree

Hide file tree

Showing 2 changed files with 37 additions and 4 deletions.
diff --git a/ehrql/debugger.py b/ehrql/debugger.py
@@ -145,7 +145,10 @@ def related_columns_to_records(columns):
 
 
 def related_patient_columns_to_records(columns):
-    for patient_id in columns[0].patient_to_value.keys():
+    patients_all = set()
+    for c in columns:
+        patients_all.update(c.patient_to_value.keys())
+    for patient_id in sorted(patients_all):
         record = {"patient_id": patient_id}
         for i, column in enumerate(columns, start=1):
             record[f"series_{i}"] = render_value(column[patient_id], convert_null=True)

diff --git a/tests/unit/test_debugger.py b/tests/unit/test_debugger.py
@@ -5,7 +5,12 @@
 import pytest
 
 from ehrql import create_dataset, show
-from ehrql.debugger import activate_debug_context, elements_are_related_series, render
+from ehrql.debugger import (
+    activate_debug_context,
+    elements_are_related_series,
+    related_patient_columns_to_records,
+    render,
+)
 from ehrql.query_engines.in_memory_database import PatientColumn
 from ehrql.tables import EventFrame, PatientFrame, Series, table
 
@@ -19,7 +24,7 @@ def date_serializer(obj):
 def test_show(capsys):
     expected_output = textwrap.dedent(
         """
-        Show line 27:
+        Show line 32:
         'Hello'
         """
     ).strip()
@@ -32,7 +37,7 @@ def test_show(capsys):
 def test_show_with_label(capsys):
     expected_output = textwrap.dedent(
         """
-        Show line 40: Number
+        Show line 45: Number
         14
         """
     ).strip()
@@ -61,6 +66,31 @@ def test_render_multiple_variables():
     assert render(12, "Hello") == expected_output
 
 
+def test_related_patient_columns_to_records_full_join():
+    c1 = PatientColumn.parse(
+        """
+        1 | 101
+        2 | 102
+        4 | 104
+        """
+    )
+    c2 = PatientColumn.parse(
+        """
+        1 | 201
+        2 | 202
+        3 | 203
+        """
+    )
+    r = list(related_patient_columns_to_records([c1, c2]))
+    r_expected = [
+        {"patient_id": 1, "series_1": 101, "series_2": 201},
+        {"patient_id": 2, "series_1": 102, "series_2": 202},
+        {"patient_id": 3, "series_1": "", "series_2": 203},
+        {"patient_id": 4, "series_1": 104, "series_2": ""},
+    ]
+    assert r == r_expected
+
+
 def test_render_formatted_table():
     expected_output = textwrap.dedent(
         """