18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421 | class SnapshotSqlMixin:
"""
Shared SQL snapshot materialization (timestamp + check strategies).
Engines provide small hooks for identifier qualification, expressions,
staging, and execution. All column names come from BaseExecutor constants.
"""
def run_snapshot_sql(self, node: Node, env: Environment) -> None:
ex = cast("BaseExecutor[Any]", self)
meta = self._snapshot_validate_node(node)
cfg = resolve_snapshot_config(node, meta)
body = self._snapshot_render_body(node, env)
rel_name = relation_for(node.name)
target = self._snapshot_target_identifier(rel_name)
if not cfg.unique_key:
raise ValueError(f"{node.path}: snapshot models require a non-empty unique_key list.")
vf = self.SNAPSHOT_VALID_FROM_COL # type: ignore[attr-defined]
vt = self.SNAPSHOT_VALID_TO_COL # type: ignore[attr-defined]
is_cur = self.SNAPSHOT_IS_CURRENT_COL # type: ignore[attr-defined]
hash_col = self.SNAPSHOT_HASH_COL # type: ignore[attr-defined]
upd_meta = self.SNAPSHOT_UPDATED_AT_COL # type: ignore[attr-defined]
self._snapshot_prepare_target()
# First run: create snapshot table
if not ex.exists_relation(rel_name):
sql = self._snapshot_first_run_sql(
body=body,
strategy=cfg.strategy,
unique_key=cfg.unique_key,
updated_at=cfg.updated_at,
check_cols=cfg.check_cols,
target=target,
vf=vf,
vt=vt,
is_cur=is_cur,
hash_col=hash_col,
upd_meta=upd_meta,
)
self._snapshot_exec_and_wait(sql)
return
# Incremental update
src_ref, cleanup = self._snapshot_source_ref(rel_name, body)
try:
keys_pred = " AND ".join([f"t.{k} = s.{k}" for k in cfg.unique_key]) or "FALSE"
if cfg.strategy == "timestamp":
if not cfg.updated_at:
raise ValueError(
f"{node.path}: strategy='timestamp' snapshots require an updated_at column."
)
change_condition = f"s.{cfg.updated_at} > t.{upd_meta}"
new_upd_expr = self._snapshot_updated_at_expr(cfg.updated_at, "s")
new_valid_from_expr = self._snapshot_updated_at_expr(cfg.updated_at, "s")
new_hash_expr = self._snapshot_null_hash()
else:
hash_expr_s = self._snapshot_hash_expr(cfg.check_cols, "s")
change_condition = f"COALESCE({hash_expr_s}, '') <> COALESCE(t.{hash_col}, '')"
new_upd_expr = (
self._snapshot_updated_at_expr(cfg.updated_at, "s")
if cfg.updated_at
else self._snapshot_current_timestamp()
)
new_valid_from_expr = self._snapshot_current_timestamp()
new_hash_expr = hash_expr_s
close_sql = self._snapshot_close_sql(
target=target,
src_ref=src_ref,
keys_pred=keys_pred,
change_condition=change_condition,
vt=vt,
is_cur=is_cur,
)
self._snapshot_exec_and_wait(close_sql)
insert_sql = self._snapshot_insert_sql(
target=target,
src_ref=src_ref,
keys_pred=keys_pred,
first_key=cfg.unique_key[0],
new_upd_expr=new_upd_expr,
new_valid_from_expr=new_valid_from_expr,
new_hash_expr=new_hash_expr,
change_condition=change_condition,
vf=vf,
vt=vt,
is_cur=is_cur,
hash_col=hash_col,
upd_meta=upd_meta,
)
self._snapshot_exec_and_wait(insert_sql)
finally:
with suppress(Exception):
cleanup()
# ---- Core SQL builders -------------------------------------------------
def _snapshot_first_run_sql(
self,
*,
body: str,
strategy: str,
unique_key: list[str],
updated_at: str | None,
check_cols: list[str],
target: str,
vf: str,
vt: str,
is_cur: str,
hash_col: str,
upd_meta: str,
) -> str:
if not unique_key:
raise ValueError("Snapshot models require a non-empty unique_key list.")
if strategy == "timestamp":
if not updated_at:
raise ValueError("strategy='timestamp' snapshots require an updated_at column.")
return f"""
{self._snapshot_create_keyword()} {target} AS
SELECT
s.*,
{self._snapshot_updated_at_expr(updated_at, "s")} AS {upd_meta},
{self._snapshot_updated_at_expr(updated_at, "s")} AS {vf},
{self._snapshot_null_timestamp()} AS {vt},
TRUE AS {is_cur},
{self._snapshot_null_hash()} AS {hash_col}
FROM ({body}) AS s
"""
if not check_cols:
raise ValueError("strategy='check' snapshots require non-empty check_cols.")
hash_expr = self._snapshot_hash_expr(check_cols, "s")
upd_expr = (
self._snapshot_updated_at_expr(updated_at, "s")
if updated_at
else self._snapshot_current_timestamp()
)
return f"""
{self._snapshot_create_keyword()} {target} AS
SELECT
s.*,
{upd_expr} AS {upd_meta},
{self._snapshot_current_timestamp()} AS {vf},
{self._snapshot_null_timestamp()} AS {vt},
TRUE AS {is_cur},
{hash_expr} AS {hash_col}
FROM ({body}) AS s
"""
def _snapshot_close_sql(
self,
*,
target: str,
src_ref: str,
keys_pred: str,
change_condition: str,
vt: str,
is_cur: str,
) -> str:
return f"""
UPDATE {target} AS t
SET
{vt} = {self._snapshot_current_timestamp()},
{is_cur} = FALSE
FROM {src_ref} AS s
WHERE
{keys_pred}
AND t.{is_cur} = TRUE
AND {change_condition}
"""
def _snapshot_insert_sql(
self,
*,
target: str,
src_ref: str,
keys_pred: str,
first_key: str,
new_upd_expr: str,
new_valid_from_expr: str,
new_hash_expr: str,
change_condition: str,
vf: str,
vt: str,
is_cur: str,
hash_col: str,
upd_meta: str,
) -> str:
return f"""
INSERT INTO {target}
SELECT
s.*,
{new_upd_expr} AS {upd_meta},
{new_valid_from_expr} AS {vf},
{self._snapshot_null_timestamp()} AS {vt},
TRUE AS {is_cur},
{new_hash_expr} AS {hash_col}
FROM {src_ref} AS s
LEFT JOIN {target} AS t
ON {keys_pred}
AND t.{is_cur} = TRUE
WHERE
t.{first_key} IS NULL
OR {change_condition}
"""
# ---- Pruning -----------------------------------------------------------
def snapshot_prune(
self,
relation: str,
unique_key: list[str],
keep_last: int,
*,
dry_run: bool = False,
) -> None:
"""
Delete older snapshot versions while keeping the most recent `keep_last`
rows per business key (including the current row).
"""
ex = cast("BaseExecutor[Any]", self)
if keep_last <= 0:
return
keys = [k for k in unique_key if k]
if not keys:
return
target = self._snapshot_target_identifier(relation)
vf = self.SNAPSHOT_VALID_FROM_COL # type: ignore[attr-defined]
key_select = ", ".join(keys)
part_by = ", ".join(keys)
ranked_sql = f"""
SELECT
{key_select},
{vf},
ROW_NUMBER() OVER (
PARTITION BY {part_by}
ORDER BY {vf} DESC
) AS rn
FROM {target}
"""
if dry_run:
sql = f"""
WITH ranked AS (
{ranked_sql}
)
SELECT COUNT(*) AS rows_to_delete
FROM ranked
WHERE rn > {int(keep_last)}
"""
res = ex._execute_sql(sql)
count = self._snapshot_fetch_count(res)
echo(
f"[DRY-RUN] snapshot_prune({relation}): would delete {count} row(s) "
f"(keep_last={keep_last})"
)
return
join_pred = " AND ".join([f"t.{k} = r.{k}" for k in keys])
delete_sql = f"""
DELETE FROM {target} t
USING (
{ranked_sql}
) r
WHERE
r.rn > {int(keep_last)}
AND {join_pred}
AND t.{vf} = r.{vf}
"""
ex._execute_sql(delete_sql)
# ---- Rendering helpers -------------------------------------------------
def _snapshot_render_body(self, node: Node, env: Environment) -> str:
ex = cast("BaseExecutor[Any]", self)
sql_rendered = ex.render_sql(
node,
env,
ref_resolver=lambda name: ex._resolve_ref(name, env),
source_resolver=ex._resolve_source,
)
sql_clean = ex._strip_leading_config(sql_rendered).strip()
return ex._selectable_body(sql_clean).rstrip(" ;\n\t")
def _snapshot_validate_node(self, node: Node) -> dict[str, Any]:
if node.kind != "sql":
raise TypeError(
f"Snapshot materialization is only supported for SQL models, "
f"got kind={node.kind!r} for {node.name}."
)
meta = getattr(node, "meta", {}) or {}
if not self._meta_is_snapshot(meta): # type: ignore[attr-defined]
raise ValueError(f"Node {node.name} is not configured with materialized='snapshot'.")
return meta
# ---- Staging -----------------------------------------------------------
def _snapshot_source_ref(
self, rel_name: str, select_body: str
) -> tuple[str, Callable[[], None]]:
"""
Return (source_ref, cleanup). Default: inline subquery.
Engines can override to use temp views/tables and cleanup afterward.
"""
return f"({select_body})", lambda: None
# ---- Hooks (must be provided by engines) -------------------------------
def _snapshot_target_identifier(self, rel_name: str) -> str: # pragma: no cover - abstract
raise NotImplementedError
def _snapshot_current_timestamp(self) -> str: # pragma: no cover - abstract
raise NotImplementedError
def _snapshot_null_timestamp(self) -> str: # pragma: no cover - abstract
raise NotImplementedError
def _snapshot_null_hash(self) -> str: # pragma: no cover - abstract
raise NotImplementedError
def _snapshot_hash_expr(self, check_cols: list[str], src_alias: str) -> str: # pragma: no cover
raise NotImplementedError
def _snapshot_updated_at_expr(self, updated_at: str, src_alias: str) -> str:
return f"{src_alias}.{updated_at}"
def _snapshot_prepare_target(self) -> None:
"""Hook for engines that need to ensure dataset/schema before writes."""
return None
def _snapshot_exec_and_wait(self, sql: str) -> None:
"""
Execute SQL and, if necessary, wait for completion (jobs, lazy DataFrames).
"""
res = self._execute_sql(sql) # type: ignore[attr-defined]
if res is None:
return
for attr in ("result", "collect"):
fn = getattr(res, attr, None)
if callable(fn):
with suppress(Exception):
fn()
break
# ---- Helpers -----------------------------------------------------------
def _snapshot_concat_expr(self, columns: list[str], src_alias: str) -> str:
parts = [
self._snapshot_coalesce(self._snapshot_cast_as_string(f"{src_alias}.{col}"), "''")
for col in columns
]
return " || '||' || ".join(parts) if parts else "''"
def _snapshot_cast_as_string(self, expr: str) -> str:
return f"CAST({expr} AS STRING)"
def _snapshot_coalesce(self, expr: str, default: str) -> str:
return f"COALESCE({expr}, {default})"
def _snapshot_create_keyword(self) -> str:
"""Hook to allow engines to override CREATE vs CREATE OR REPLACE."""
return "CREATE TABLE"
def _snapshot_fetch_count(self, res: Any) -> int:
"""
Best-effort extraction of a single COUNT(*) value from various result shapes.
"""
try:
if hasattr(res, "fetchone"):
row = res.fetchone()
if row is not None:
return int(row[0])
if hasattr(res, "fetchall"):
rows = res.fetchall()
if rows:
return int(rows[0][0])
result_fn = getattr(res, "result", None)
if callable(result_fn):
rows_obj = result_fn()
if isinstance(rows_obj, Iterable):
rows = list(rows_obj)
if rows:
return int(rows[0][0])
collect_fn = getattr(res, "collect", None)
if callable(collect_fn):
rows_obj = collect_fn()
if isinstance(rows_obj, Iterable):
rows = list(rows_obj)
if rows:
return int(rows[0][0])
if isinstance(res, list) and res:
return int(res[0][0])
except Exception:
return 0
return 0
|