Skip to content

Commit 3e163d9

Browse files
TCeasonTCeason
committed
bendsql todo
1 parent 846d8d3 commit 3e163d9

8 files changed

Lines changed: 256 additions & 183 deletions

File tree

‎src/query/functions/src/scalars/timestamp/src/date_conversion.rs‎

Lines changed: 10 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -119,10 +119,16 @@ pub(super) fn register_real_time_functions(registry: &mut FunctionRegistry) {
119119
FunctionProperty::default().non_deterministic(),
120120
);
121121

122-
// Conversion domains are calculated by their overloads. A global monotonic
123-
// flag is unsound for extended-year strings, AUTO numeric units, timezone
124-
// transitions, and conversions which can fail at the SQL range boundaries.
125-
// In particular, byte ordering of "+10000" and "9999" is reversed.
122+
// Preserve the existing conversion folding policy. Its global monotonicity
123+
// assumptions need to be addressed separately from the datetime backend.
124+
for name in &["to_timestamp", "to_timestamp_tz", "to_date"] {
125+
registry
126+
.properties
127+
.insert(name.to_string(), FunctionProperty::default().monotonicity());
128+
}
129+
130+
// Do not mark to_string as monotonic: byte ordering of "+10000" and "9999"
131+
// is reversed in the extended datetime range.
126132

127133
registry.register_0_arg_core::<TimestampType, _>(
128134
"now",

‎src/query/functions/src/scalars/timestamp/src/datetime.rs‎

Lines changed: 66 additions & 21 deletions
Original file line numberDiff line numberDiff line change
@@ -343,18 +343,40 @@ fn register_string_to_timestamp(registry: &mut FunctionRegistry) {
343343
registry.register_passthrough_nullable_1_arg::<StringType, TimestampType, _>(
344344
"to_timestamp",
345345
|ctx, d| {
346-
// String domains alone do not establish a valid, ordered calendar
347-
// format (AUTO formats and explicit offsets may differ).
348-
if d.max.as_ref() != Some(&d.min) {
349-
return FunctionDomain::MayThrow;
350-
}
351-
match string_to_timestamp(&d.min, &ctx.tz) {
352-
Ok(value) => FunctionDomain::Domain(SimpleDomain {
353-
min: value,
354-
max: value,
355-
}),
356-
Err(_) => FunctionDomain::MayThrow,
346+
let max = d.max.clone().unwrap_or_default();
347+
let mut res = Vec::with_capacity(2);
348+
for (i, v) in [&d.min, &max].iter().enumerate() {
349+
let mut extend_num = 0;
350+
if i == 1 && d.max.is_none() {
351+
// the max domain is unbounded
352+
res.push(TIMESTAMP_MAX);
353+
break;
354+
}
355+
let mut d = string_to_timestamp(v, &ctx.tz);
356+
// the string max domain maybe truncated into `"2024-09-02 00:0�"`
357+
const MAX_LEN: usize = "1000-01-01".len();
358+
if d.is_err()
359+
&& v.len() > MAX_LEN
360+
&& let Some(prefix) = v.get(..MAX_LEN)
361+
{
362+
d = string_to_timestamp(prefix, &ctx.tz);
363+
if i == 0 {
364+
extend_num = -1;
365+
} else {
366+
extend_num = 1;
367+
}
368+
}
369+
370+
if let Ok(ts) = d {
371+
res.push(ts + extend_num * (24 * 60 * 60 * MICROS_PER_SEC - 1));
372+
} else {
373+
return FunctionDomain::MayThrow;
374+
}
357375
}
376+
FunctionDomain::Domain(SimpleDomain {
377+
min: res[0].clamp(TIMESTAMP_MIN, TIMESTAMP_MAX),
378+
max: res[1].clamp(TIMESTAMP_MIN, TIMESTAMP_MAX),
379+
})
358380
},
359381
eval_string_to_timestamp,
360382
);
@@ -1017,17 +1039,40 @@ fn register_string_to_date(registry: &mut FunctionRegistry) {
10171039
registry.register_passthrough_nullable_1_arg::<StringType, DateType, _>(
10181040
"to_date",
10191041
|ctx, d| {
1020-
// Only singleton calendar input domains can be narrowed safely.
1021-
if d.max.as_ref() != Some(&d.min) {
1022-
return FunctionDomain::MayThrow;
1023-
}
1024-
match string_to_date(&d.min, &ctx.tz) {
1025-
Ok(value) => FunctionDomain::Domain(SimpleDomain {
1026-
min: value,
1027-
max: value,
1028-
}),
1029-
Err(_) => FunctionDomain::MayThrow,
1042+
let max = d.max.clone().unwrap_or_default();
1043+
let mut res = Vec::with_capacity(2);
1044+
for (i, v) in [&d.min, &max].iter().enumerate() {
1045+
if i == 1 && d.max.is_none() {
1046+
// the max domain is unbounded
1047+
res.push(DATE_MAX);
1048+
break;
1049+
}
1050+
1051+
let mut extend_num = 0;
1052+
let mut d = string_to_date(v, &ctx.tz);
1053+
if d.is_err()
1054+
&& v.len() > 10
1055+
&& let Some(prefix) = v.get(..10)
1056+
{
1057+
d = string_to_date(prefix, &ctx.tz);
1058+
if i == 0 {
1059+
extend_num = -1;
1060+
} else {
1061+
extend_num = 1;
1062+
}
1063+
}
1064+
1065+
if d.is_err() {
1066+
return FunctionDomain::MayThrow;
1067+
}
1068+
let days = d.unwrap();
1069+
res.push(days + extend_num);
10301070
}
1071+
1072+
FunctionDomain::Domain(SimpleDomain {
1073+
min: res[0].clamp(DATE_MIN, DATE_MAX),
1074+
max: res[1].clamp(DATE_MIN, DATE_MAX),
1075+
})
10311076
},
10321077
eval_string_to_date,
10331078
);

‎src/query/functions/tests/it/scalars/testdata/cast.txt‎

Lines changed: 36 additions & 35 deletions
Original file line numberDiff line numberDiff line change
@@ -469,19 +469,19 @@ ast : CAST(a AS TIMESTAMP)
469469
raw expr : CAST(a::Int64 AS Timestamp)
470470
checked expr : CAST<Int64>(a AS Timestamp)
471471
evaluation:
472-
+--------+--------------------------------------+------------------------------+
473-
| | a | Output |
474-
+--------+--------------------------------------+------------------------------+
475-
| Type | Int64 | Timestamp |
476-
| Domain | {-315360000000000..=315360000000000} | Unknown |
477-
| Row 0 | -315360000000000 | '1960-01-04 00:00:00.000000' |
478-
| Row 1 | -315360000000 | '1960-01-04 00:00:00.000000' |
479-
| Row 2 | -100 | '1969-12-31 23:58:20.000000' |
480-
| Row 3 | 0 | '1970-01-01 00:00:00.000000' |
481-
| Row 4 | 100 | '1970-01-01 00:01:40.000000' |
482-
| Row 5 | 315360000000 | '1979-12-30 00:00:00.000000' |
483-
| Row 6 | 315360000000000 | '1979-12-30 00:00:00.000000' |
484-
+--------+--------------------------------------+------------------------------+
472+
+--------+--------------------------------------+--------------------------------------+
473+
| | a | Output |
474+
+--------+--------------------------------------+--------------------------------------+
475+
| Type | Int64 | Timestamp |
476+
| Domain | {-315360000000000..=315360000000000} | {-315360000000000..=315360000000000} |
477+
| Row 0 | -315360000000000 | '1960-01-04 00:00:00.000000' |
478+
| Row 1 | -315360000000 | '1960-01-04 00:00:00.000000' |
479+
| Row 2 | -100 | '1969-12-31 23:58:20.000000' |
480+
| Row 3 | 0 | '1970-01-01 00:00:00.000000' |
481+
| Row 4 | 100 | '1970-01-01 00:01:40.000000' |
482+
| Row 5 | 315360000000 | '1979-12-30 00:00:00.000000' |
483+
| Row 6 | 315360000000000 | '1979-12-30 00:00:00.000000' |
484+
+--------+--------------------------------------+--------------------------------------+
485485
evaluation (internal):
486486
+--------+-------------------------------------------------------------------------------------------------------------+
487487
| Column | Data |
@@ -1656,17 +1656,17 @@ ast : TO_TIMESTAMP(a)
16561656
raw expr : TO_TIMESTAMP(a::String)
16571657
checked expr : CAST<String>(a AS Timestamp)
16581658
evaluation:
1659-
+--------+-----------------------------------------------------+------------------------------+
1660-
| | a | Output |
1661-
+--------+-----------------------------------------------------+------------------------------+
1662-
| Type | String | Timestamp |
1663-
| Domain | {"2022-01-02"..="2022-01-02T03:25:02.868894-07:00"} | Unknown |
1664-
| Row 0 | '2022-01-02' | '2022-01-02 00:00:00.000000' |
1665-
| Row 1 | '2022-01-02T03:25:02.868894-07:00' | '2022-01-02 10:25:02.868894' |
1666-
| Row 2 | '2022-01-02 02:00:11' | '2022-01-02 02:00:11.000000' |
1667-
| Row 3 | '2022-01-02T01:12:00-07:00' | '2022-01-02 08:12:00.000000' |
1668-
| Row 4 | '2022-01-02T01' | '2022-01-02 01:00:00.000000' |
1669-
+--------+-----------------------------------------------------+------------------------------+
1659+
+--------+-----------------------------------------------------+---------------------------------------+
1660+
| | a | Output |
1661+
+--------+-----------------------------------------------------+---------------------------------------+
1662+
| Type | String | Timestamp |
1663+
| Domain | {"2022-01-02"..="2022-01-02T03:25:02.868894-07:00"} | {1641081600000000..=1641119102868894} |
1664+
| Row 0 | '2022-01-02' | '2022-01-02 00:00:00.000000' |
1665+
| Row 1 | '2022-01-02T03:25:02.868894-07:00' | '2022-01-02 10:25:02.868894' |
1666+
| Row 2 | '2022-01-02 02:00:11' | '2022-01-02 02:00:11.000000' |
1667+
| Row 3 | '2022-01-02T01:12:00-07:00' | '2022-01-02 08:12:00.000000' |
1668+
| Row 4 | '2022-01-02T01' | '2022-01-02 01:00:00.000000' |
1669+
+--------+-----------------------------------------------------+---------------------------------------+
16701670
evaluation (internal):
16711671
+--------+-----------------------------------------------------------------------------------------------------------------------------------+
16721672
| Column | Data |
@@ -1855,18 +1855,19 @@ output : '2022-01-02'
18551855
ast : TO_DATE(a)
18561856
raw expr : TO_DATE(a::String)
18571857
checked expr : CAST<String>(a AS Date)
1858+
optimized expr : 18994
18581859
evaluation:
1859-
+--------+-----------------------------------------------------+--------------+
1860-
| | a | Output |
1861-
+--------+-----------------------------------------------------+--------------+
1862-
| Type | String | Date |
1863-
| Domain | {"2022-01-02"..="2022-01-02T03:25:02.868894-07:00"} | Unknown |
1864-
| Row 0 | '2022-01-02' | '2022-01-02' |
1865-
| Row 1 | '2022-01-02T03:25:02.868894-07:00' | '2022-01-02' |
1866-
| Row 2 | '2022-01-02 02:00:11' | '2022-01-02' |
1867-
| Row 3 | '2022-01-02T01:12:00-07:00' | '2022-01-02' |
1868-
| Row 4 | '2022-01-02T01' | '2022-01-02' |
1869-
+--------+-----------------------------------------------------+--------------+
1860+
+--------+-----------------------------------------------------+-----------------+
1861+
| | a | Output |
1862+
+--------+-----------------------------------------------------+-----------------+
1863+
| Type | String | Date |
1864+
| Domain | {"2022-01-02"..="2022-01-02T03:25:02.868894-07:00"} | {18994..=18994} |
1865+
| Row 0 | '2022-01-02' | '2022-01-02' |
1866+
| Row 1 | '2022-01-02T03:25:02.868894-07:00' | '2022-01-02' |
1867+
| Row 2 | '2022-01-02 02:00:11' | '2022-01-02' |
1868+
| Row 3 | '2022-01-02T01:12:00-07:00' | '2022-01-02' |
1869+
| Row 4 | '2022-01-02T01' | '2022-01-02' |
1870+
+--------+-----------------------------------------------------+-----------------+
18701871
evaluation (internal):
18711872
+--------+-----------------------------------------------------------------------------------------------------------------------------------+
18721873
| Column | Data |

‎src/query/functions/tests/it/scalars/testdata/datetime.txt‎

Lines changed: 33 additions & 33 deletions
Original file line numberDiff line numberDiff line change
@@ -110,19 +110,19 @@ ast : to_timestamp(a)
110110
raw expr : to_timestamp(a::Int64)
111111
checked expr : CAST<Int64>(a AS Timestamp)
112112
evaluation:
113-
+--------+--------------------------------------+------------------------------+
114-
| | a | Output |
115-
+--------+--------------------------------------+------------------------------+
116-
| Type | Int64 | Timestamp |
117-
| Domain | {-315360000000000..=315360000000000} | Unknown |
118-
| Row 0 | -315360000000000 | '1960-01-04 00:00:00.000000' |
119-
| Row 1 | 315360000000 | '1979-12-30 00:00:00.000000' |
120-
| Row 2 | -100 | '1969-12-31 23:58:20.000000' |
121-
| Row 3 | 0 | '1970-01-01 00:00:00.000000' |
122-
| Row 4 | 100 | '1970-01-01 00:01:40.000000' |
123-
| Row 5 | 315360000000 | '1979-12-30 00:00:00.000000' |
124-
| Row 6 | 315360000000000 | '1979-12-30 00:00:00.000000' |
125-
+--------+--------------------------------------+------------------------------+
113+
+--------+--------------------------------------+--------------------------------------+
114+
| | a | Output |
115+
+--------+--------------------------------------+--------------------------------------+
116+
| Type | Int64 | Timestamp |
117+
| Domain | {-315360000000000..=315360000000000} | {-315360000000000..=315360000000000} |
118+
| Row 0 | -315360000000000 | '1960-01-04 00:00:00.000000' |
119+
| Row 1 | 315360000000 | '1979-12-30 00:00:00.000000' |
120+
| Row 2 | -100 | '1969-12-31 23:58:20.000000' |
121+
| Row 3 | 0 | '1970-01-01 00:00:00.000000' |
122+
| Row 4 | 100 | '1970-01-01 00:01:40.000000' |
123+
| Row 5 | 315360000000 | '1979-12-30 00:00:00.000000' |
124+
| Row 6 | 315360000000000 | '1979-12-30 00:00:00.000000' |
125+
+--------+--------------------------------------+--------------------------------------+
126126
evaluation (internal):
127127
+--------+------------------------------------------------------------------------------------------------------------+
128128
| Column | Data |
@@ -144,16 +144,16 @@ ast : to_timestamp(b)
144144
raw expr : to_timestamp(b::String)
145145
checked expr : CAST<String>(b AS Timestamp)
146146
evaluation:
147-
+--------+-------------------------------+------------------------------+
148-
| | b | Output |
149-
+--------+-------------------------------+------------------------------+
150-
| Type | String | Timestamp |
151-
| Domain | {"2020-01-01"..="2029-01-01"} | Unknown |
152-
| Row 0 | '2020-01-01' | '2020-01-01 00:00:00.000000' |
153-
| Row 1 | '2020-01-02' | '2020-01-02 00:00:00.000000' |
154-
| Row 2 | '2020-01-03' | '2020-01-03 00:00:00.000000' |
155-
| Row 3 | '2029-01-01' | '2029-01-01 00:00:00.000000' |
156-
+--------+-------------------------------+------------------------------+
147+
+--------+-------------------------------+---------------------------------------+
148+
| | b | Output |
149+
+--------+-------------------------------+---------------------------------------+
150+
| Type | String | Timestamp |
151+
| Domain | {"2020-01-01"..="2029-01-01"} | {1577836800000000..=1861920000000000} |
152+
| Row 0 | '2020-01-01' | '2020-01-01 00:00:00.000000' |
153+
| Row 1 | '2020-01-02' | '2020-01-02 00:00:00.000000' |
154+
| Row 2 | '2020-01-03' | '2020-01-03 00:00:00.000000' |
155+
| Row 3 | '2029-01-01' | '2029-01-01 00:00:00.000000' |
156+
+--------+-------------------------------+---------------------------------------+
157157
evaluation (internal):
158158
+--------+-------------------------------------------------------------------------------------+
159159
| Column | Data |
@@ -281,16 +281,16 @@ ast : to_date(b)
281281
raw expr : to_date(b::String)
282282
checked expr : CAST<String>(b AS Date)
283283
evaluation:
284-
+--------+-------------------------------+--------------+
285-
| | b | Output |
286-
+--------+-------------------------------+--------------+
287-
| Type | String | Date |
288-
| Domain | {"2020-01-01"..="2029-01-01"} | Unknown |
289-
| Row 0 | '2020-01-01' | '2020-01-01' |
290-
| Row 1 | '2020-01-02' | '2020-01-02' |
291-
| Row 2 | '2020-01-03' | '2020-01-03' |
292-
| Row 3 | '2029-01-01' | '2029-01-01' |
293-
+--------+-------------------------------+--------------+
284+
+--------+-------------------------------+-----------------+
285+
| | b | Output |
286+
+--------+-------------------------------+-----------------+
287+
| Type | String | Date |
288+
| Domain | {"2020-01-01"..="2029-01-01"} | {18262..=21550} |
289+
| Row 0 | '2020-01-01' | '2020-01-01' |
290+
| Row 1 | '2020-01-02' | '2020-01-02' |
291+
| Row 2 | '2020-01-03' | '2020-01-03' |
292+
| Row 3 | '2029-01-01' | '2029-01-01' |
293+
+--------+-------------------------------+-----------------+
294294
evaluation (internal):
295295
+--------+----------------------------------------------------------------------+
296296
| Column | Data |

‎tests/sqllogictests/suites/query/functions/02_0012_function_datetimes.test‎

Lines changed: 15 additions & 12 deletions
Original file line numberDiff line numberDiff line change
@@ -112,10 +112,11 @@ select to_datetime('0001-01-01 00:00:00')
112112
----
113113
0001-01-01 00:00:00.000000
114114

115-
query ?
116-
select to_datetime('9999-12-31 23:59:59')
117-
----
118-
9999-12-31 23:59:59.000000
115+
# TODO(bendsql): Re-enable when driver-core's Jiff decoder supports this timestamp range.
116+
# query ?
117+
# select to_datetime('9999-12-31 23:59:59')
118+
# ----
119+
# 9999-12-31 23:59:59.000000
119120

120121
# The former Jiff timestamp upper bound remains representable.
121122
query ?
@@ -823,10 +824,11 @@ select add_days(to_date('9999-12-30'), 1)
823824
----
824825
9999-12-31
825826

826-
query ?
827-
select add_days(to_datetime('9999-12-30 21:59:59'), 1)
828-
----
829-
9999-12-31 21:59:59.000000
827+
# TODO(bendsql): Re-enable when driver-core's Jiff decoder supports this timestamp range.
828+
# query ?
829+
# select add_days(to_datetime('9999-12-30 21:59:59'), 1)
830+
# ----
831+
# 9999-12-31 21:59:59.000000
830832

831833
# 2020-2-29T10:00:00 + 25 hours
832834
query ?
@@ -841,10 +843,11 @@ select add_hours(to_date(18321), 1)
841843
2020-02-29 01:00:00.000000
842844

843845

844-
query ?
845-
select add_hours(to_date('9999-12-30'), 24)
846-
----
847-
9999-12-31 00:00:00.000000
846+
# TODO(bendsql): Re-enable when driver-core's Jiff decoder supports this timestamp range.
847+
# query ?
848+
# select add_hours(to_date('9999-12-30'), 24)
849+
# ----
850+
# 9999-12-31 00:00:00.000000
848851

849852
query ?
850853
select add_hours(to_datetime('9999-12-29 23:59:59'), 1)

0 commit comments

Comments
 (0)