This optimization reuses an existing cursor when op_open_write() is
called on the same table/index (same root_page). This is safe because
the cursor position doesn't matter - op_rewind() is always called after
op_open_write() to position the cursor at the beginning of the
table/index before any operations are performed.
This change speeds up op_open_write() by avoiding unnecessary cursor re-
initialization.
Closes#3815
Trying to return integer sometimes to match SQLite led to more problems
that I anticipated. The reason being, we can't *really* match SQLite's
behavior unless we know the type of *every* element in the sum. This is
not impossible, but it is very hard, for very little gain.
Fixes#3831Closes#3832
The SQLite varint specification states that the varint is guaranteed to
be a maximum of 9 bytes, but our version of write_varint initializes a
buffer of 10 bytes. Changing the size to match the specification.
We could put an assert to check that the 10th byte is never written to
but, reducing the size seems like the correct thing to do based on the
specification.
Varint constraints are mentioned in this specification:
https://www.sqlite.org/fileformat.htmlCloses#3835
The SQLite varint specification states that the varint is guaranteed to be a maximum of 9 bytes, but our version of write_varint initializes a buffer of 10 bytes. Changing the size to match the specification.
Trying to return integer sometimes to match SQLite led to more problems
that I anticipated. The reason being, we can't *really* match SQLite's
behavior unless we know the type of *every* element in the sum. This is
not impossible, but it is very hard, for very little gain.
Fixes#3831
Implements COUNT/SUM/AVG(DISTINCT) and SELECT DISTINCT for materialized
views. To do this we have to keep a list of the actual distinct values
(similarly to how we do for min/max). We then update the operator (and
issue deltas) only when there is a state transition (for example, if we
already count the value x = 1, and we see an insert for x = 1, we do
nothing).
SELECT DISTINCT (with no aggregator) is similar. We already have to keep
a list of the values anyway to power the aggregates. So we just issue
new deltas based on the transition, without updating the aggregator.
Closes#3808
Our foreign key constraint checks were checking for changes in PRIMARY
KEYs, but not unique indexes - which are in practice the same thing,
apart from the `INTEGER PRIMARY KEY` special case, where the PRIMARY KEY
is an alias for the rowid of the table.
Closes#3648Closes#3652 (reimplements)
Reviewed-by: Preston Thorpe <preston@turso.tech>
Closes#3825
The `julian_day_converter` crate is GPL, which is problematic for apps
embedding Turso. Switch to SQLite's Julian date logic by porting the C
code to Rust.
Open a connection per transaction in the rusqlite benchmark so that
we're comparing the same workload with Turso.
Reviewed-by: Jussi Saurio <jussi.saurio@gmail.com>
Closes#3816
If WAL is already enabled, let's just continue execution instead of
erroring out.
Reviewed-by: Jussi Saurio <jussi.saurio@gmail.com>
Reviewed-by: Pere Diaz Bou <pere-altea@homail.com>
Closes#3819
closes#3811
adds `text_cache` which owns the null terminated bytes, which get cached
if a subsequent call to `sqlite3_column_text` is made.
#3809 depends on this fix
Reviewed-by: Jussi Saurio <jussi.saurio@gmail.com>
Closes#3817
This optimization reuses an existing cursor when op_open_write() is
called on the same table/index (same root_page). This is safe because
the cursor position doesn't matter - op_rewind() is always called
after op_open_write() to position the cursor at the beginning of the
table/index before any operations are performed.
This change speeds up op_open_write() by avoiding unnecessary cursor
re-initialization.
