Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Allow using vector index when the queried vector is provided in a user variable. #2816

Merged
merged 2 commits into from
Jan 15, 2025
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
12 changes: 12 additions & 0 deletions enginetest/queries/vector_index_queries.go
Original file line number Diff line number Diff line change
Expand Up @@ -26,6 +26,7 @@ var VectorIndexQueries = []ScriptTest{
"create table vectors (id int primary key, v json);",
`insert into vectors values (1, '[4.0,3.0]'), (2, '[0.0,0.0]'), (3, '[-1.0,1.0]'), (4, '[0.0,-2.0]');`,
`create vector index v_idx on vectors(v);`,
`set @query_vec = '[0.0,0.0]';`,
},
Assertions: []ScriptTestAssertion{
{
Expand All @@ -44,6 +45,17 @@ var VectorIndexQueries = []ScriptTest{
},
ExpectedIndexes: []string{"v_idx"},
},
{
// Queries against a user var can be optimized.
Query: "select * from vectors order by VEC_DISTANCE(@query_vec, v) limit 4",
Expected: []sql.Row{
{2, types.MustJSON(`[0.0, 0.0]`)},
{3, types.MustJSON(`[-1.0, 1.0]`)},
{4, types.MustJSON(`[0.0, -2.0]`)},
{1, types.MustJSON(`[4.0, 3.0]`)},
},
ExpectedIndexes: []string{"v_idx"},
},
{
// Only queries with a limit can use a vector index.
Query: "select * from vectors order by VEC_DISTANCE('[0.0,0.0]', v)",
Expand Down
17 changes: 13 additions & 4 deletions sql/analyzer/replace_order_by_distance.go
Original file line number Diff line number Diff line change
Expand Up @@ -59,15 +59,24 @@ func replaceIdxOrderByDistanceHelper(ctx *sql.Context, scope *plan.Scope, node s
if !isDistance {
return n, transform.SameTree, nil
}

// We currently require that the query vector to the distance function is a constant value that does not
// depend on the row. Right now that can be a Literal or a UserVar.
isLiteral := func(expr sql.Expression) bool {
switch expr.(type) {
case *expression.Literal, *expression.UserVar:
return true
}
return false
}

var column sql.Expression
var literal sql.Expression
_, leftIsLiteral := distance.LeftChild.(*expression.Literal)
if leftIsLiteral {
if isLiteral(distance.LeftChild) {
column = distance.RightChild
literal = distance.LeftChild
} else {
_, rightIsLiteral := distance.RightChild.(*expression.Literal)
if rightIsLiteral {
if isLiteral(distance.RightChild) {
column = distance.LeftChild
literal = distance.RightChild
} else {
Expand Down
Loading