fix(db-engine-specs): handle bare-year python_date_format columns in time-grain expressions

get_timestamp_expr() only knew how to convert epoch_s/epoch_ms raw values
into a proper timestamp before applying a time-grain truncation function;
any other python_date_format (e.g. "%Y") fell through and passed the raw
value straight into the grain function. On SQLite that means
DATETIME(year, 'start of year') interprets a bare integer year as a Julian
day number rather than a calendar year, silently returning NULL for every
row. This exact pattern ships in Superset's own video_game_sales example
dataset (year: BIGINT, is_dttm=true, python_date_format: '%Y'), so any
time-grain-truncated chart against it (Nightingale Rose included) returns
no data.

Add a year_to_dttm() hook (mirroring epoch_to_dttm/epoch_ms_to_dttm) that
converts a bare year into a proper date before truncation, implemented for
SQLite via printf()-based date construction with an explicit NULL guard
(printf() otherwise treats a NULL argument as 0).

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
This commit is contained in:
Claude Code
2026-07-25 23:30:38 -07:00
parent 9a025267bd
commit 73ca8fb727
3 changed files with 62 additions and 0 deletions

View File

@@ -131,3 +131,34 @@ def test_time_grain_expressions(dttm: str, grain: str, expected: str) -> None:
with engine.connect() as connection:
result = connection.execute(text(sql)).scalar()
assert result == expected
@pytest.mark.parametrize(
"year,expected",
[
(2013, "2013-01-01 00:00:00"),
(2013.0, "2013-01-01 00:00:00"),
(None, None),
],
)
def test_year_pdf_time_grain(year: Optional[float], expected: Optional[str]) -> None:
"""A bare four-digit year (e.g. the `year` column on the `video_game_sales`
example dataset) has no native date type; without `year_to_dttm` the raw
value is passed straight into the grain function, which SQLite reads as a
Julian day number rather than a calendar year, silently producing NULL."""
from sqlalchemy import column
from superset.db_engine_specs.sqlite import SqliteEngineSpec
engine = create_engine("sqlite://", future=True)
with engine.begin() as connection:
connection.execute(text("CREATE TABLE t (year REAL)"))
connection.execute(text("INSERT INTO t VALUES (:year)"), {"year": year})
expression = SqliteEngineSpec.get_timestamp_expr(
col=column("year"), pdf="%Y", time_grain=TimeGrain.YEAR
)
sql = f"SELECT {expression} FROM t" # noqa: S608
with engine.connect() as connection:
result = connection.execute(text(sql)).scalar()
assert result == expected