fix(clickhouse): fix truncating to date from a timestamp (#10220)

ibis/backends/clickhouse/tests/snapshots/test_functions/test_timestamp_truncate/d/out.sql

@@ -1,2 +1,2 @@
 SELECT
-  toStartOfDay(parseDateTimeBestEffort('2009-05-17T12:34:56')) AS "TimestampTruncate(datetime.datetime(2009, 5, 17, 12, 34, 56), DAY)"
+  CAST(toStartOfDay(parseDateTimeBestEffort('2009-05-17T12:34:56')) AS Nullable(DateTime)) AS "TimestampTruncate(datetime.datetime(2009, 5, 17, 12, 34, 56), DAY)"

ibis/backends/clickhouse/tests/snapshots/test_functions/test_timestamp_truncate/w/out.sql

@@ -1,2 +1,2 @@
 SELECT
-  toMonday(parseDateTimeBestEffort('2009-05-17T12:34:56')) AS "TimestampTruncate(datetime.datetime(2009, 5, 17, 12, 34, 56), WEEK)"
+  CAST(toMonday(parseDateTimeBestEffort('2009-05-17T12:34:56')) AS Nullable(DateTime)) AS "TimestampTruncate(datetime.datetime(2009, 5, 17, 12, 34, 56), WEEK)"

ibis/backends/clickhouse/tests/snapshots/test_functions/test_timestamp_truncate/y/out.sql

@@ -1,2 +1,2 @@
 SELECT
-  toStartOfYear(parseDateTimeBestEffort('2009-05-17T12:34:56')) AS "TimestampTruncate(datetime.datetime(2009, 5, 17, 12, 34, 56), YEAR)"
+  CAST(toStartOfYear(parseDateTimeBestEffort('2009-05-17T12:34:56')) AS Nullable(DateTime)) AS "TimestampTruncate(datetime.datetime(2009, 5, 17, 12, 34, 56), YEAR)"

ibis/backends/clickhouse/tests/test_functions.py

@@ -492,3 +492,19 @@ def my_eq(a: int, b: int) -> bool: ...
     expr = alltypes.int_col.collect().filter(lambda x: my_eq(x, 1))
     result = expr.execute()
     assert set(result) == {1}
+
+
+def test_timestamp_to_start_of_week(con):
+    pytest.importorskip("pyarrow")
+
+    expr = ibis.timestamp("2024-02-03 00:00:00").truncate("W")
+    result = con.to_pyarrow(expr).as_py()
+    assert result == datetime(2024, 1, 29, 0, 0, 0)
+
+
+def test_date_to_start_of_day(con):
+    pytest.importorskip("pyarrow")
+
+    expr = ibis.date("2024-02-03")
+    expr1 = expr.truncate("D")
+    assert con.to_pyarrow(expr1) == con.to_pyarrow(expr)

ibis/backends/sql/compilers/clickhouse.py

@@ -377,21 +377,35 @@ def visit_TimestampFromUNIX(self, op, *, arg, unit):
 
     def visit_TimestampTruncate(self, op, *, arg, unit):
         if (short := unit.short) == "W":
-            func = "toMonday"
+            funcname = "toMonday"
         else:
-            func = f"toStartOf{unit.singular.capitalize()}"
+            funcname = f"toStartOf{unit.singular.capitalize()}"
 
-        if short in ("s", "ms", "us", "ns"):
-            arg = self.f.toDateTime64(arg, op.arg.dtype.scale or 0)
-        return self.f[func](arg)
+        func = self.f[funcname]
+
+        if short in ("Y", "Q", "M", "W", "D"):
+            # these units return `Date` so we have to cast back to the
+            # corresponding Ibis type
+            return self.cast(func(arg), op.dtype)
+        elif short in ("s", "ms", "us", "ns"):
+            return func(self.f.toDateTime64(arg, op.arg.dtype.scale or 0))
+        else:
+            assert short in ("h", "m"), short
+            return func(arg)
 
     visit_TimeTruncate = visit_TimestampTruncate
 
     def visit_DateTruncate(self, op, *, arg, unit):
-        if unit.short == "W":
+        if unit.short == "D":
+            # no op because truncating a date to a date has no effect
+            return arg
+        elif unit.short == "W":
             func = "toMonday"
         else:
             func = f"toStartOf{unit.singular.capitalize()}"
+
+        # no cast needed here because all of the allowed units return `Date`
+        # values
         return self.f[func](arg)
 
     def visit_TimestampBucket(self, op, *, arg, interval, offset):