Pekka Enberg
fdf3b0d16b
Merge '(core): Primary key index scans and single-column secondary index scans' from Jussi Saurio
...
This PR adds an index on `users.age` to `testing.db`, and support for
indexed lookups. Only single-column ascending indexes are currently
supported.
This PR also gets rid of `Operator::Seekrowid` in favor of
`Operator::Search` which handles all non-full-table-scan searches: 1.
integer primary key (rowid) point queries 2. integer primary key index
scans, and 3. secondary index scans.
examples:
```
limbo> select first_name, age from users where age > 90 limit 10;
Miranda|90
Sarah|90
Justin|90
Justin|90
John|90
Jeremy|90
Stephanie|90
Joshua|90
Jenny|90
Jennifer|90
limbo> explain query plan select first_name, age from users where age > 90 limit 10;
QUERY PLAN
`--TAKE 10
`--PROJECT first_name, age
| `--SEARCH users USING INDEX age_idx
limbo> explain select first_name, age from users where age > 90 limit 10;
addr opcode p1 p2 p3 p4 p5 comment
---- ----------------- ---- ---- ---- ------------- -- -------
0 Init 0 15 0 0 Start at 15
1 OpenReadAsync 0 2 0 0 table=users, root=2
2 OpenReadAwait 0 0 0 0
3 OpenReadAsync 1 274 0 0 table=age_idx, root=274
4 OpenReadAwait 0 0 0 0
5 Integer 90 1 0 0 r[1]=90
6 SeekGT 1 14 1 0
7 DeferredSeek 1 0 0 0
8 Column 0 1 2 0 r[2]=users.first_name
9 Column 0 9 3 0 r[3]=users.age
10 ResultRow 2 2 0 0 output=r[2..3]
11 DecrJumpZero 4 14 0 0 if (--r[4]==0) goto 14
12 NextAsync 1 0 0 0
13 NextAwait 1 7 0 0
14 Halt 0 0 0 0
15 Transaction 0 0 0 0
16 Integer 10 4 0 0 r[4]=10
17 Goto 0 1 0 0
```
Sqlite version:
```
sqlite> explain select first_name, age from users where age > 90 limit 10;
addr opcode p1 p2 p3 p4 p5 comment
---- ------------- ---- ---- ---- ------------- -- -------------
0 Init 0 13 0 0 Start at 13
1 Integer 10 1 0 0 r[1]=10; LIMIT counter
2 OpenRead 0 2 0 10 0 root=2 iDb=0; users
3 OpenRead 1 274 0 k(2,,) 0 root=274 iDb=0; age_idx
4 Integer 90 2 0 0 r[2]=90
5 SeekGT 1 12 2 1 0 key=r[2]
6 DeferredSeek 1 0 0 0 Move 0 to 1.rowid if needed
7 Column 0 1 3 0 r[3]= cursor 0 column 1
8 Column 1 0 4 0 r[4]= cursor 1 column 0
9 ResultRow 3 2 0 0 output=r[3..4]
10 DecrJumpZero 1 12 0 0 if (--r[1])==0 goto 12
11 Next 1 6 0 0
12 Halt 0 0 0 0
13 Transaction 0 0 3 0 1 usesStmtJournal=0
14 Goto 0 1 0 0
```
---
´Seek` instructions are also now supported for primary key rowid
searches:
```
limbo> select id, first_name from users where id > 9995;
9996|Donald
9997|Ruth
9998|Dorothy
9999|Gina
10000|Nicole
limbo> explain query plan select id, first_name from users where id > 9995;
QUERY PLAN
`--PROJECT id, first_name
`--SEARCH users USING INTEGER PRIMARY KEY (rowid=?)
limbo> explain select id, first_name from users where id > 9995;
addr opcode p1 p2 p3 p4 p5 comment
---- ----------------- ---- ---- ---- ------------- -- -------
0 Init 0 11 0 0 Start at 11
1 OpenReadAsync 0 2 0 0 table=users, root=2
2 OpenReadAwait 0 0 0 0
3 Integer 9995 1 0 0 r[1]=9995
4 SeekGT 0 10 1 0
5 RowId 0 2 0 0 r[2]=users.rowid
6 Column 0 1 3 0 r[3]=users.first_name
7 ResultRow 2 2 0 0 output=r[2..3]
8 NextAsync 0 0 0 0
9 NextAwait 0 5 0 0
10 Halt 0 0 0 0
11 Transaction 0 0 0 0
12 Goto 0 1 0 0
```
sqlite:
```
sqlite> explain select id, first_name from users where id > 9995;
addr opcode p1 p2 p3 p4 p5 comment
---- ------------- ---- ---- ---- ------------- -- -------------
0 Init 0 8 0 0 Start at 8
1 OpenRead 0 2 0 2 0 root=2 iDb=0; users
2 SeekGT 0 7 1 0 key=r[1]; pk
3 Rowid 0 2 0 0 r[2]=users.rowid
4 Column 0 1 3 0 r[3]= cursor 0 column 1
5 ResultRow 2 2 0 0 output=r[2..3]
6 Next 0 3 0 0
7 Halt 0 0 0 0
8 Transaction 0 0 3 0 1 usesStmtJournal=0
9 Integer 9995 1 0 0 r[1]=9995
10 Goto 0 1 0 0
```
---
More complex example with a join that uses both a rowid lookup and a
secondary index scan:
```
limbo> explain query plan select u.first_name, p.name from users u join products p on u.id = p.id and u.age > 70;
QUERY PLAN
`--PROJECT u.first_name, p.name
`--JOIN
| |--SEARCH u USING INDEX age_idx
| `--SEARCH p USING INTEGER PRIMARY KEY (rowid=?)
limbo> explain select u.first_name, p.name from users u join products p on u.id = p.id and u.age > 70;
addr opcode p1 p2 p3 p4 p5 comment
---- ----------------- ---- ---- ---- ------------- -- -------
0 Init 0 18 0 0 Start at 18
1 OpenReadAsync 0 2 0 0 table=u, root=2
2 OpenReadAwait 0 0 0 0
3 OpenReadAsync 1 274 0 0 table=age_idx, root=274
4 OpenReadAwait 0 0 0 0
5 OpenReadAsync 2 3 0 0 table=p, root=3
6 OpenReadAwait 0 0 0 0
7 Integer 70 1 0 0 r[1]=70
8 SeekGT 1 17 1 0
9 DeferredSeek 1 0 0 0
10 RowId 0 2 0 0 r[2]=u.rowid
11 SeekRowid 2 2 15 0 if (r[2]!=p.rowid) goto 15
12 Column 0 1 3 0 r[3]=u.first_name
13 Column 2 1 4 0 r[4]=p.name
14 ResultRow 3 2 0 0 output=r[3..4]
15 NextAsync 1 0 0 0
16 NextAwait 1 9 0 0
17 Halt 0 0 0 0
18 Transaction 0 0 0 0
19 Goto 0 1 0 0
```
sqlite:
```
sqlite> explain select u.first_name, p.name from users u join products p on u.id = p.id and u.age > 70;
addr opcode p1 p2 p3 p4 p5 comment
---- ------------- ---- ---- ---- ------------- -- -------------
0 Init 0 14 0 0 Start at 14
1 OpenRead 0 2 0 10 0 root=2 iDb=0; users
2 OpenRead 2 274 0 k(2,,) 0 root=274 iDb=0; age_idx
3 OpenRead 1 3 0 2 0 root=3 iDb=0; products
4 Integer 70 1 0 0 r[1]=70
5 SeekGT 2 13 1 1 0 key=r[1]
6 DeferredSeek 2 0 0 0 Move 0 to 2.rowid if needed
7 IdxRowid 2 2 0 0 r[2]=rowid; users.rowid
8 SeekRowid 1 12 2 0 intkey=r[2]
9 Column 0 1 3 0 r[3]= cursor 0 column 1
10 Column 1 1 4 0 r[4]= cursor 1 column 1
11 ResultRow 3 2 0 0 output=r[3..4]
12 Next 2 6 0 0
13 Halt 0 0 0 0
14 Transaction 0 0 3 0 1 usesStmtJournal=0
15 Goto 0 1 0 0
```
Closes #350
2024-10-13 10:11:23 +03:00
Arpit Saxena
28a603f56a
[sorter] Hold records in Vec instead of a BTreeMap
...
We now just insert them one after the other in the vector. When rewind
is called, the vector is sorted. Iterating is just taking elements from
the vector
2024-10-09 17:58:28 +05:30
jussisaurio
556f4b73c9
Refine edge case handling: add optional predicate to get_next_record()
2024-10-08 08:23:30 +03:00
jussisaurio
43038cb6aa
Handle seek() edge case with index seek
2024-10-08 07:45:21 +03:00
jussisaurio
572db69b5e
Add TODO comment about index corner case
2024-10-07 17:12:19 +03:00
jussisaurio
93a8110773
dont assume index key has rowid in the second column: its the last
2024-10-07 17:05:38 +03:00
jussisaurio
8563d620af
renaming
2024-10-07 17:03:50 +03:00
jussisaurio
fc71f2b32f
traverse index properly
...
interior index cells have values that are not in the leaves, e.g.
(interior: 3)
/ \
(leaf: 2) (leaf: 4)
so their values need to be emitted after the left subtree is emitted.
2024-10-07 13:04:03 +03:00
jussisaurio
e5cf052f07
Why do sqlite btree child keys have <= keys and not < keys
2024-10-06 23:48:59 +03:00
Lauri Virtanen
0ae1412193
Add instr(X,Y) scalar function
...
Relates to issue #144
2024-10-06 20:19:37 +03:00
jussisaurio
15a66ea662
single seek function in cursor trait
2024-10-06 09:21:15 +03:00
jussisaurio
6e7db36121
reorder
2024-10-06 00:58:32 +03:00
jussisaurio
af9a751d36
Single seek function
2024-10-06 00:56:18 +03:00
jussisaurio
1ae8d28669
Use same move_to() function for tables and indexes
2024-10-06 00:51:14 +03:00
jussisaurio
37f877109e
Reduce duplication in btree.rs
2024-10-06 00:39:50 +03:00
jussisaurio
bb1c8b65e8
fmt
2024-10-06 00:22:12 +03:00
jussisaurio
dde10d2dd7
Better EXPLAIN QUERY PLAN for Operator::Search
2024-10-06 00:19:56 +03:00
jussisaurio
47534cb8df
Get rid of Seekrowid operator in favor of a unified Search operator
2024-10-06 00:11:38 +03:00
jussisaurio
d3e797f59e
rewind_labels was renamed to scan_loop_body_labels
2024-10-05 18:27:18 +03:00
jussisaurio
d22dbe9840
remove garbage comment
2024-10-05 18:25:04 +03:00
jussisaurio
3826d4e1ff
Add comment about code duplication
2024-10-05 18:25:04 +03:00
jussisaurio
d2233d69d3
Dont assume the rowid is the second column - it's the last
2024-10-05 18:25:04 +03:00
jussisaurio
9169f6e39b
same thing
2024-10-05 18:25:04 +03:00
jussisaurio
fe90aacd35
Handle CursorResult in deferred seek
2024-10-05 18:25:04 +03:00
jussisaurio
3a11887122
fixerinos
2024-10-05 18:25:04 +03:00
jussisaurio
ed19f47762
fix
2024-10-05 18:25:04 +03:00
jussisaurio
ff236c7781
Fix not advancing the cell index of pages
2024-10-05 18:25:04 +03:00
jussisaurio
99871bbeea
yield on io
2024-10-05 18:25:04 +03:00
jussisaurio
e118b70127
fmt
2024-10-05 18:25:04 +03:00
jussisaurio
3d56fbd91c
stuff
2024-10-05 18:25:04 +03:00
jussisaurio
f02da18acd
index scan wip foo doesnt work yet
2024-10-05 18:25:04 +03:00
jussisaurio
d3015ad854
Merge 'Fix NextAwait's next instruction; rename rewind_labels' from Arpit Saxena
...
Closes #351
NextAwait's next instruction was pointing to RewindAwait earlier which
is not current. The instruction now points to the instruction after
RewindAwait, which will be the loop body. Hence, rename rewind_labels to
scan_loop_body_labels to make it clearer
Reviewed-by: Jussi Saurio <jussi.saurio@gmail.com >
Closes #354
2024-10-05 18:23:37 +03:00
Lauri Virtanen
9e80a0c4a8
Add randomblob(N) scalar function
...
Relates to issue #144
2024-10-03 00:05:23 +03:00
Pekka Enberg
b7656c265c
Fix source formatting with cargo fmt...
2024-10-02 11:06:38 +03:00
Pekka Enberg
6fcd818160
Merge 'Add unhex(X) scalar function' from Lauri Virtanen
...
This commit adds `unhex(X)` scalar function. Function with `unhex(X,Y)`
two arguments is not implemented yet.
Relates to issue #144
Closes #353
2024-10-02 11:01:15 +03:00
Arpit Saxena
47c8fd1964
Fix NextAwait's next instruction; rename rewind_labels
...
Closes #351
NextAwait's next instruction was pointing to RewindAwait earlier which
is not current. The instruction now points to the instruction after
RewindAwait, which will be the loop body. Hence, rename rewind_labels to
scan_loop_body_labels to make it clearer
2024-10-02 11:58:42 +05:30
Lauri Virtanen
adc6f9b6cd
Add unhex(X) scalar function
...
This commit adds `unhex(X)` scalar function. Function with `unhex(X,Y)`
two arguments is not implemented yet.
Relates to issue #144
2024-09-30 00:13:43 +03:00
Lauri Virtanen
f612ead8a3
Add zeroblob(N) scalar function
...
Relates to issue #144
2024-09-29 23:39:53 +03:00
Arpit Saxena
b7debabd81
Wrap IoUring to ensure pending_ops is always correctly updated
...
Adds a struct WrappedIOUring which contains a IoUring and a pending_ops
field. Entry submission and popping from the queue is done through
functions operating on it, which also maintains pending_ops count
NOTE: This is a bit weird/hacky since in get_completion we create a
CompletionQueue and just call its next(). If it were a normal iterator
it would always return the same first item. However it is a queue posing
as an iterator which makes this work.
2024-09-28 00:10:05 +05:30
Arpit Saxena
a7d735d368
Maintain pending io_uring ops count for early return on run_once
...
Add a `pending_ops` field to `InnerLinuxIO` struct which is incremented
for each operation submitted to the ring and decremented when they are
taken off the completion queue. With this, we can exit from run_once if
there are no pending operations. Otherwise, in that case, it would hang
indefinitely due to call of `ring.submit_and_wait(1)`
2024-09-27 23:32:19 +05:30
JeanArhancet
7b5d0fcabd
refactor: split min/max func
2024-09-25 17:51:49 +02:00
JeanArhancet
222fea9cc8
fix: cmp number/text
2024-09-25 17:39:37 +02:00
Lauri Virtanen
0597c048fc
Better support for BLOBs
...
- Limbo command line shell supports e.g. `SELECT x'616263';`
- `EXPLAIN SELECT x'616263';` lists the opcode
Missing:
- Command line shell not entirely compatible with SQLite when blobs have
non-printable characters in the middle (e.g. `SELECT x'610062';`)
- Python bindings not supported (incoming soon)
2024-09-22 17:55:00 +03:00
Pekka Enberg
5aeb18f474
Fix source formatting with cargo fmt
2024-09-22 08:52:32 -04:00
baishen
5c00c576a5
Add support for hex scalar function
2024-09-22 08:50:08 -04:00
jussisaurio
c738674c5d
Merge 'fix: cmp integer/float' from Jean Arhancet
...
Closes https://github.com/penberg/limbo/issues/339
Reviewed-by: Lauri Virtanen <pere-altea@hotmail.com >
Closes #340
2024-09-21 18:06:04 +03:00
JeanArhancet
41c1708ee7
fix: cmp integer/float
2024-09-21 16:10:42 +02:00
Lauri Virtanen
67573e12e5
Add typeof(X) scalar function
2024-09-21 15:56:29 +03:00
김선우
bbd1eea9ae
Add support for parenthesized conditions
2024-09-21 16:03:49 +09:00
Pekka Enberg
b67640c437
Merge 'core: constraint check uniqueness' from Pere Diaz Bou
...
```bash
limbo> insert into products values (1, 'asdf', 432);
Runtime error: UNIQUE constraint failed: products.id (19)
```
Closes #336
2024-09-20 13:27:10 +03:00