Pere Diaz Bou
39a75147d4
Page cache by page_number and frame_number
...
Since page cache is now shared by default, we need to cache pages by
page number and something else. I chose to go with max_frame of
connection, because this connection will have a max_frame set until from
the start of a transaction until the end of it.
With key pairs of (pgno, max_frame) we make sure each connection is
caching based on the snapshot it is at as two different connections
might have the same pageno being using but a different frame. If both
have same max_frame then they will share same page.
2024-12-13 21:57:27 +01:00
Pekka Enberg
138b3a00e8
Merge 'use correct min/max frames' from Pere Diaz Bou
...
Reviewed-by: Jussi Saurio <jussi.saurio@gmail.com >
Closes #465
2024-12-13 18:48:26 +02:00
Pere Diaz Bou
1e5239a164
use correct min/max frames
2024-12-13 14:06:29 +01:00
Pere Diaz Bou
352b3540ee
loop writing header while initializing
...
if somehow this doesn't finish in a single loop, do more just in case
2024-12-13 13:25:56 +01:00
Pere Diaz Bou
1a663a6ed7
cargo stuff rm
2024-12-13 13:09:13 +01:00
Pere Diaz Bou
b43e8e46f6
impl sync/send for cache
2024-12-13 13:09:13 +01:00
Pere Diaz Bou
97dd95abea
core: change Rc<RefCell<Page>> to Arc<Page>
...
This includes an inner struct in Page wrapped with Unsafe cell to access
it. This is done intentionally because concurrency control of pages is
handled by pages and not by the page itself.
2024-12-13 13:09:13 +01:00
Pere Diaz Bou
c816186326
fmt
2024-12-13 13:09:13 +01:00
Pere Diaz Bou
3fda2d09b9
Extract multi threaded part from WalFile to WalFileShared
...
Since we expect to ensure thread safety between multiple threads in the
future, we extract what is important to be shared between multiple
connections with regards to WAL.
This is WIP so I just put whatever feels like important behind a RwLock
but expect this to change to Atomics in the future as needed. Maybe even
these locks might disappear because they will be better served with
transaction locks.
2024-12-13 13:09:13 +01:00
Pere Diaz Bou
a4297702bd
extract page cache to be multi threaded
2024-12-13 13:09:13 +01:00
Pekka Enberg
2b85d2a600
Merge 'Add implementation and tests for replace scalar function' from Alperen Keleş
...
Adds `replace` scalar function.
Reviewed-by: Jussi Saurio <jussi.saurio@gmail.com >
Closes #446
2024-12-13 11:02:08 +02:00
JeanArhancet
8bf6572e9e
feat: support unary positive
2024-12-13 02:07:34 +01:00
alpaylan
4f395623ab
fix the comment on replace panic in case of text cast failure
2024-12-12 15:03:35 -05:00
alpaylan
5742c51c5c
improve error messages on replace function call
2024-12-12 14:14:04 -05:00
Alperen Keleş
841a4fe2f8
Merge branch 'tursodatabase:main' into main
2024-12-12 13:13:41 -05:00
Pekka Enberg
5796b418b9
Merge 'Add support for soundex() function' from flaneur
...
add [soundex](https://www.sqlite.org/lang_corefunc.html#soundex ) scalar
function.
it seems that sqlite did not enable `soundex()` function by default
unless build it with `SQLITE_SOUNDEX`, while the sqlite in the ci
workflow did not enable it. this pr skipped the test over `soundex()`
temporarily in the `scalar-function.test` file.
Closes #453
2024-12-12 18:10:46 +02:00
Pekka Enberg
91764b85e6
Merge 'Add bitwise vdbe ops' from Preston Thorpe
...
Love the project, been following your blog posts for quite a while now.
I asked on Discord prior to submitting this, just because I didn't see a
specific issue for this feature... but if this PR is out of scope for
contributors, feel free to close it as I just had a good time hacking on
it.
This PR adds support for `BitAnd`, `BitOr`, and `BitNot` operators in
the vdbe, as well as unary expressions applied to aggregate functions;
which was needed in order to have `BitNot` support the same tests that
the other operators had.
*Also added unary negation of function calls, because since unary ops
were added, I figured adding support for the other existing unary
operator might be in scope, but lmk if not.
Let me know if there is any more tests or documentation to add/improve.
Closes #445
2024-12-12 17:35:12 +02:00
Li Yazhou
03288e5170
add impl about scalar function soundex with test
2024-12-12 21:48:05 +08:00
Alex Miller
c4d4569dc9
Merge remote-tracking branch 'upstream/main' into expr-iif
2024-12-11 20:13:54 -08:00
alpaylan
021456326e
change 0.1+0.2 test into 0.3 as limbo does not yet support decimals
2024-12-11 16:32:06 -05:00
alpaylan
da28ed51ca
add implementation and tests for replace scalar function
2024-12-11 16:23:13 -05:00
PThorpe92
16595f39f5
Add support for unary op negation of aggregates
2024-12-11 15:38:21 -05:00
PThorpe92
d5391dc716
Add vdbe bitwise operators: and, or, not
2024-12-11 11:06:22 -05:00
jussisaurio
eb9374aebf
Merge 'Add support for CASE expressions.' from Alex Miller
...
There's two forms of case:
CASE (WHEN [bool expr] THEN [value])+ (ELSE [value])? END
which checks a series of boolean conditions, and:
CASE expr (WHEN [expr] THEN [value})+ (ELSE [value])? END
Which checks a series of equality conditions.
This implements support for both. Note that the ELSE is optional, and
will be equivalent to `ELSE null` if not specified.
sqlite3 gives the implementation as:
```
sqlite> explain select case a WHEN a THEN b WHEN c THEN d ELSE 0 END from casetest;
addr opcode p1 p2 p3 p4 p5 comment
---- ------------- ---- ---- ---- ------------- -- -------------
0 Init 0 16 0 0 Start at 16
1 OpenRead 0 3 0 4 0 root=3 iDb=0; casetest
2 Rewind 0 15 0 0
3 Column 0 0 2 0 r[2]= cursor 0 column 0
4 Column 0 0 3 0 r[3]= cursor 0 column 0
5 Ne 3 8 2 BINARY-8 83 if r[2]!=r[3] goto 8
6 Column 0 1 1 0 r[1]= cursor 0 column 1
7 Goto 0 13 0 0
8 Column 0 2 3 0 r[3]= cursor 0 column 2
9 Ne 3 12 2 BINARY-8 83 if r[2]!=r[3] goto 12
10 Column 0 3 1 0 r[1]= cursor 0 column 3
11 Goto 0 13 0 0
12 Integer 0 1 0 0 r[1]=0
13 ResultRow 1 1 0 0 output=r[1]
14 Next 0 3 0 1
15 Halt 0 0 0 0
16 Transaction 0 0 2 0 1 usesStmtJournal=0
17 Goto 0 1 0 0
```
and after this patch, limbo gives:
```
addr opcode p1 p2 p3 p4 p5 comment
---- ----------------- ---- ---- ---- ------------- -- -------
0 Init 0 19 0 0 Start at 19
1 OpenReadAsync 0 4 0 0 table=casetest, root=4
2 OpenReadAwait 0 0 0 0
3 RewindAsync 0 0 0 0
4 RewindAwait 0 18 0 0 Rewind table casetest
5 Column 0 0 2 0 r[2]=casetest.a
6 Column 0 0 3 0 r[3]=casetest.a
7 Ne 2 3 10 0 if r[2]!=r[3] goto 10
8 Column 0 1 1 0 r[1]=casetest.b
9 Goto 0 15 0 0
10 Column 0 2 3 0 r[3]=casetest.c
11 Ne 2 3 14 0 if r[2]!=r[3] goto 14
12 Column 0 3 1 0 r[1]=casetest.d
13 Goto 0 15 0 0
14 Integer 0 1 0 0 r[1]=0
15 ResultRow 1 1 0 0 output=r[1]
16 NextAsync 0 0 0 0
17 NextAwait 0 5 0 0
18 Halt 0 0 0 0
19 Transaction 0 0 0 0
20 Goto 0 1 0 0
```
And then as there's nowhere to annotate this new support in COMPAT.md, I
added a corresponding heading for SELECT expressions and what is/isn't
supported.
Reviewed-by: Jussi Saurio <jussi.saurio@gmail.com >
Closes #425
2024-12-11 17:05:41 +02:00
Pekka Enberg
ab07c77036
Upgrade pprof to 0.14
...
Github's dependabot complains that the current version has an unsoudness
issue so let's bump to a newer version:
https://github.com/tursodatabase/limbo/security/dependabot/10
2024-12-11 11:21:09 +02:00
Pekka Enberg
04f196113a
Merge 'Add last_insert_rowid() function' from Krishna Vishal
...
- Changed `Cursor` trait to be able to get access to `root_page`
- SQLite only updates last_insert_rowid for non-schema inserts. So we
check if the `InsertAwait` is not for `root_page` before updating
rowid
In SQLite it looks like this:
```
sqlite> EXPLAIN SELECT last_insert_rowid();
addr opcode p1 p2 p3 p4 p5 comment
---- ------------- ---- ---- ---- ------------- -- -------------
0 Init 0 4 0 0
1 Function 0 0 1 last_insert_rowid(0) 0
2 ResultRow 1 1 0 0
3 Halt 0 0 0 0
4 Goto 0 1 0 0
```
In limbo it will look like this:
```
limbo> EXPLAIN SELECT last_insert_rowid();
addr opcode p1 p2 p3 p4 p5 comment
---- ----------------- ---- ---- ---- ------------- -- -------
0 Init 0 4 0 0 Start at 4
1 Function 0 2 1 last_insert_rowid 0 r[1]=func()
2 ResultRow 1 1 0 0 output=r[1]
3 Halt 0 0 0 0
4 Transaction 0 0 0 0
5 Goto 0 1 0 0
```
Closes #427
2024-12-11 10:44:34 +02:00
Pekka Enberg
ca272ba937
Merge 'Support JOIN USING and NATURAL JOIN' from Jussi Saurio
...
Closes #360
Closes #361
Closes #422
2024-12-11 09:17:51 +02:00
Pekka Enberg
eda1f5396c
Merge 'Add octet_length scalar function' from Kacper Kołodziej
...
Adds `octet_length` scalar function.
Part of solution for: #144
Closes #430
2024-12-11 07:44:04 +02:00
Alex Miller
88c862ce4d
Comments, resolve label better, make tests more fun
2024-12-10 19:59:54 -08:00
Alex Miller
e85df1c895
resolve labels to current offset. make test clearer.
2024-12-10 19:36:54 -08:00
Kacper Kołodziej
d4bff2c93e
add octet_length scalar function
2024-12-10 22:56:38 +01:00
Kacper Kołodziej
660d3e8d07
fix: count characters in string in length function
...
`length` function should count characters, not bytes.
https://www.sqlite.org/lang_corefunc.html#length
2024-12-10 22:48:50 +01:00
krishvishal
134b5576ad
Ran cargo fmt
2024-12-09 22:55:54 +05:30
krishvishal
7e2928a5f1
Feature: last_insert_rowid()
...
- Changed `Cursor` trait to be able to get access to `root_page`
- SQLite only updates last_insert_rowid for non-schema inserts. So we check if the `InsertAwait` is not for `root_page` before
updating rowid
2024-12-09 22:48:42 +05:30
jussisaurio
fe88d45e5e
Add more comments to push_predicate/push_predicates
2024-12-09 17:50:29 +02:00
jussisaurio
840caed2f7
Fix bug with multiway joins that include the same table multiple times
2024-12-09 17:50:29 +02:00
jussisaurio
7924f9b64d
consider all joined tables instead of just previous in natural/using
2024-12-09 17:50:29 +02:00
jussisaurio
81b6605453
support NATURAL JOIN
2024-12-09 17:50:29 +02:00
jussisaurio
bed932c186
Support join USING
2024-12-09 17:50:29 +02:00
Pekka Enberg
ba1f7cd16f
Merge 'feat(core/translate): support HAVING' from Jussi Saurio
...
support the HAVING clause.
note that sqlite (and i think standard sql?) supports HAVING even
without GROUP BY, but `sqlite3-parser` doesn't.
also fixes some issues with the PartialOrd implementation of OwnedValue
and the implementations of `concat` and `round` which i discovered due
to my HAVING tcl tests failing
Closes #420
2024-12-09 17:30:40 +02:00
krishvishal
1e23af7d24
Added last_insert_rowid() function.
...
Need to fix its behavior. Problem is probably with `Cursor` implementation.
2024-12-09 17:41:28 +05:30
Alex Miller
f7bb7f8dee
Fix typo and improve comment
2024-12-08 14:20:23 -08:00
Alex Miller
c2e3957d73
I misunderstood what a constant instruction was
2024-12-08 14:12:45 -08:00
Alex Miller
eb00226cfe
Add support for CASE expressions.
...
There's two forms of case:
CASE (WHEN [bool expr] THEN [value])+ (ELSE [value])? END
which checks a series of boolean conditions, and:
CASE expr (WHEN [expr] THEN [value})+ (ELSE [value])? END
Which checks a series of equality conditions.
This implements support for both. Note that the ELSE is optional, and
will be equivalent to `ELSE null` if not specified.
sqlite3 gives the implementation as:
sqlite> explain select case a WHEN a THEN b WHEN c THEN d ELSE 0 END from casetest;
addr opcode p1 p2 p3 p4 p5 comment
---- ------------- ---- ---- ---- ------------- -- -------------
0 Init 0 16 0 0 Start at 16
1 OpenRead 0 3 0 4 0 root=3 iDb=0; casetest
2 Rewind 0 15 0 0
3 Column 0 0 2 0 r[2]= cursor 0 column 0
4 Column 0 0 3 0 r[3]= cursor 0 column 0
5 Ne 3 8 2 BINARY-8 83 if r[2]!=r[3] goto 8
6 Column 0 1 1 0 r[1]= cursor 0 column 1
7 Goto 0 13 0 0
8 Column 0 2 3 0 r[3]= cursor 0 column 2
9 Ne 3 12 2 BINARY-8 83 if r[2]!=r[3] goto 12
10 Column 0 3 1 0 r[1]= cursor 0 column 3
11 Goto 0 13 0 0
12 Integer 0 1 0 0 r[1]=0
13 ResultRow 1 1 0 0 output=r[1]
14 Next 0 3 0 1
15 Halt 0 0 0 0
16 Transaction 0 0 2 0 1 usesStmtJournal=0
17 Goto 0 1 0 0
and after this patch, limbo gives:
addr opcode p1 p2 p3 p4 p5 comment
---- ----------------- ---- ---- ---- ------------- -- -------
0 Init 0 18 0 0 Start at 18
1 OpenReadAsync 0 4 0 0 table=casetest, root=4
2 OpenReadAwait 0 0 0 0
3 RewindAsync 0 0 0 0
4 RewindAwait 0 17 0 0 Rewind table casetest
5 Column 0 0 2 0 r[2]=casetest.a
6 Column 0 0 3 0 r[3]=casetest.a
7 Ne 2 3 10 0 if r[2]!=r[3] goto 10
8 Column 0 1 1 0 r[1]=casetest.b
9 Goto 0 14 0 0
10 Column 0 2 3 0 r[3]=casetest.c
11 Ne 2 3 14 0 if r[2]!=r[3] goto 14
12 Column 0 3 1 0 r[1]=casetest.d
13 Goto 0 14 0 0
14 ResultRow 1 1 0 0 output=r[1]
15 NextAsync 0 0 0 0
16 NextAwait 0 5 0 0
17 Halt 0 0 0 0
18 Transaction 0 0 0 0
19 Integer 0 1 0 0 r[1]=0
20 Goto 0 1 0 0
And then as there's nowhere to annotate this new support in COMPAT.md, I
added a corresponding heading for SELECT expressions and what is/isn't
supported.
2024-12-08 14:09:03 -08:00
Alex Miller
27ef95e0ab
LLM recommended a better way to write args matching
2024-12-07 21:15:04 -08:00
Alex Miller
183ea8e362
Implement support for iif().
...
In sqlite, iif() looks like:
sqlite> create table iiftest(a int, b int, c int);
sqlite> explain select iif(a,b,c) from iiftest;
addr opcode p1 p2 p3 p4 p5 comment
---- ------------- ---- ---- ---- ------------- -- -------------
0 Init 0 11 0 0 Start at 11
1 OpenRead 0 2 0 3 0 root=2 iDb=0; iiftest
2 Rewind 0 10 0 0
3 Column 0 0 2 0 r[2]= cursor 0 column 0
4 IfNot 2 7 1 0
5 Column 0 1 1 0 r[1]= cursor 0 column 1
6 Goto 0 8 0 0
7 Column 0 2 1 0 r[1]= cursor 0 column 2
8 ResultRow 1 1 0 0 output=r[1]
9 Next 0 3 0 1
10 Halt 0 0 0 0
11 Transaction 0 0 1 0 1 usesStmtJournal=0
12 Goto 0 1 0 0
And with this change, in limbo it looks like:
addr opcode p1 p2 p3 p4 p5 comment
---- ----------------- ---- ---- ---- ------------- -- -------
0 Init 0 14 0 0 Start at 14
1 OpenReadAsync 0 2 0 0 table=iiftest, root=2
2 OpenReadAwait 0 0 0 0
3 RewindAsync 0 0 0 0
4 RewindAwait 0 13 0 0 Rewind table iiftest
5 Column 0 0 2 0 r[2]=iiftest.a
6 IfNot 2 9 1 0 if !r[2] goto 9
7 Column 0 1 1 0 r[1]=iiftest.b
8 Goto 0 10 0 0
9 Column 0 2 1 0 r[1]=iiftest.c
10 ResultRow 1 1 0 0 output=r[1]
11 NextAsync 0 0 0 0
12 NextAwait 0 5 0 0
13 Halt 0 0 0 0
14 Transaction 0 0 0 0
15 Goto 0 1 0 0
2024-12-07 21:04:03 -08:00
jussisaurio
ca25a73c95
a little bit more explanation about left join handling
2024-11-30 20:54:22 +02:00
jussisaurio
83f8ea1b13
Fix bug with multiway joins and clean up left join implementation
2024-11-30 20:47:48 +02:00
jussisaurio
3f80e41e7a
support HAVING
2024-11-30 10:05:13 +02:00
jussisaurio
84742b81fa
Obsolete comment
2024-11-27 22:43:36 +02:00
