[NO MERGE]fix: expression ownership in the temp table of subquery #168
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
9 tasks
Member
|
Have you tried this before? CREATE TABLE t1(pk int primary key, a int, b int);
CREATE TABLE t2(pk int primary key, a int, c int);
explain SELECT b FROM t1 WHERE a in (SELECT a FROM t2 where a > 1);
PLAN
-------------------------------------------------------------------
Projection [t1.b] [Project] +
LeftSemi Join On t1.a = (t1.a) as (_temp_table_1_.a) [HashJoin]+
Scan t1 -> [a, b] [SeqScan] +
Projection [t1.a] [Project] +
Filter (t1.a > 1), Is Having: false [Filter] +
Scan t2 -> [] [SeqScan]The problem is here: First bind t1 and t1.b, t1.a, then the subquery. When binding the subquery, it hit these code, and get t1.a but not t2.a. |
Member
Author
|
You are a genius! Found so many issues and I fixed them, please check them out sir! let _ = fnck_sql
.run("CREATE TABLE t1(pk int primary key, a int, b int);")
.await?;
let _ = fnck_sql
.run("CREATE TABLE t2(pk int primary key, a int, c int);")
.await?;
let (schema, tuples) = fnck_sql
.run("explain SELECT b FROM t1 WHERE a in (SELECT a FROM t2 where a > 1);")
.await?;+-------------------------------------------------------------------+
| PLAN |
+===================================================================+
| Projection [t1.b] [Project] |
| LeftSemi Join On t1.a = (t2.a) as (_temp_table_1_.a) [HashJoin] |
| Scan t1 -> [a, b] [SeqScan] |
| Projection [t2.a] [Project] |
| Filter (t2.a > 1), Is Having: false [Filter] |
| Scan t2 -> [a] [SeqScan] |
+-------------------------------------------------------------------+ |
Member
Author
@crwen |
Member
|
consider this also: explain select a from t1 where a in (select t2.a from t1 as t2 where t2.pk > 10);
PLAN
-------------------------------------------------------------------
Projection [t1.a] [Project] +
LeftSemi Join On t1.a = (t1.a) as (_temp_table_1_.a) [HashJoin]+
Scan t1 -> [a] [SeqScan] +
Projection [t1.a] [Project] +
Filter (t1.pk > 10), Is Having: false [Filter] +
Scan t1 -> [pk, a] [SeqScan]When calling |
Member
Author
Great, your observation skills are amazing. I fixed the problem. Now the pos in the case you gave seems to be the same as before, but it is correct. I checked it in the new commit. The corresponding relationship of pos. By the way, their pos may be the same, but they actually mean the positions in the respective left tuple and right tuple. This is because expr needs to be used in left_build: https://github.com/KipData/FnckSQL/blob/main/src/execution/volcano/dql/join/hash_join.rs#L119 best_plan plan: LogicalPlan {
operator: Project(
ProjectOperator {
exprs: [
Reference {
expr: ColumnRef(
ColumnCatalog {
summary: ColumnSummary {
id: Some(
1,
),
name: "a",
table_name: Some(
"t1",
),
},
nullable: true,
desc: ColumnDesc {
column_datatype: Integer,
is_primary: false,
is_unique: false,
default: None,
},
},
),
pos: 0,
},
],
},
),
childrens: [
LogicalPlan {
operator: Join(
JoinOperator {
on: On {
on: [
(
Reference {
expr: ColumnRef(
ColumnCatalog {
summary: ColumnSummary {
id: Some(
1,
),
name: "a",
table_name: Some(
"t1",
),
},
nullable: true,
desc: ColumnDesc {
column_datatype: Integer,
is_primary: false,
is_unique: false,
default: None,
},
},
),
pos: 0,
},
Reference {
expr: Alias {
expr: ColumnRef(
ColumnCatalog {
summary: ColumnSummary {
id: Some(
1,
),
name: "a",
table_name: Some(
"t1",
),
},
nullable: true,
desc: ColumnDesc {
column_datatype: Integer,
is_primary: false,
is_unique: false,
default: None,
},
},
),
alias: Expr(
ColumnRef(
ColumnCatalog {
summary: ColumnSummary {
id: Some(
1,
),
name: "a",
table_name: Some(
"_temp_table_1_",
),
},
nullable: true,
desc: ColumnDesc {
column_datatype: Integer,
is_primary: false,
is_unique: false,
default: None,
},
},
),
),
},
pos: 0,
},
),
],
filter: None,
},
join_type: LeftSemi,
},
),
childrens: [
LogicalPlan {
operator: Scan(
ScanOperator {
table_name: "t1",
primary_key: 0,
columns: [
(
1,
ColumnCatalog {
summary: ColumnSummary {
id: Some(
1,
),
name: "a",
table_name: Some(
"t1",
),
},
nullable: true,
desc: ColumnDesc {
column_datatype: Integer,
is_primary: false,
is_unique: false,
default: None,
},
},
),
],
limit: (
None,
None,
),
index_infos: [
IndexInfo {
meta: IndexMeta {
id: 0,
column_ids: [
0,
],
table_name: "t1",
pk_ty: Integer,
name: "pk_pk",
ty: PrimaryKey,
},
range: None,
},
],
},
),
childrens: [],
physical_option: Some(
SeqScan,
),
_output_schema_ref: None,
},
LogicalPlan {
operator: Project(
ProjectOperator {
exprs: [
Alias {
expr: Reference {
expr: ColumnRef(
ColumnCatalog {
summary: ColumnSummary {
id: Some(
1,
),
name: "a",
table_name: Some(
"t1",
),
},
nullable: true,
desc: ColumnDesc {
column_datatype: Integer,
is_primary: false,
is_unique: false,
default: None,
},
},
),
pos: 0,
},
alias: Expr(
ColumnRef(
ColumnCatalog {
summary: ColumnSummary {
id: Some(
1,
),
name: "a",
table_name: Some(
"_temp_table_1_",
),
},
nullable: true,
desc: ColumnDesc {
column_datatype: Integer,
is_primary: false,
is_unique: false,
default: None,
},
},
),
),
},
],
},
),
childrens: [
LogicalPlan {
operator: Project(
ProjectOperator {
exprs: [
Reference {
expr: ColumnRef(
ColumnCatalog {
summary: ColumnSummary {
id: Some(
1,
),
name: "a",
table_name: Some(
"t1",
),
},
nullable: true,
desc: ColumnDesc {
column_datatype: Integer,
is_primary: false,
is_unique: false,
default: None,
},
},
),
pos: 1,
},
],
},
),
childrens: [
LogicalPlan {
operator: Filter(
FilterOperator {
predicate: Binary {
op: Gt,
left_expr: Reference {
expr: ColumnRef(
ColumnCatalog {
summary: ColumnSummary {
id: Some(
0,
),
name: "pk",
table_name: Some(
"t1",
),
},
nullable: false,
desc: ColumnDesc {
column_datatype: Integer,
is_primary: true,
is_unique: false,
default: None,
},
},
),
pos: 0,
},
right_expr: Constant(
Int32(0),
),
ty: Boolean,
},
having: false,
},
),
childrens: [
LogicalPlan {
operator: Scan(
ScanOperator {
table_name: "t1",
primary_key: 0,
columns: [
(
0,
ColumnCatalog {
summary: ColumnSummary {
id: Some(
0,
),
name: "pk",
table_name: Some(
"t1",
),
},
nullable: false,
desc: ColumnDesc {
column_datatype: Integer,
is_primary: true,
is_unique: false,
default: None,
},
},
),
(
1,
ColumnCatalog {
summary: ColumnSummary {
id: Some(
1,
),
name: "a",
table_name: Some(
"t1",
),
},
nullable: true,
desc: ColumnDesc {
column_datatype: Integer,
is_primary: false,
is_unique: false,
default: None,
},
},
),
],
limit: (
None,
None,
),
index_infos: [
IndexInfo {
meta: IndexMeta {
id: 0,
column_ids: [
0,
],
table_name: "t1",
pk_ty: Integer,
name: "pk_pk",
ty: PrimaryKey,
},
range: Some(
Scope {
min: Excluded(
Int32(0),
),
max: Unbounded,
},
),
},
],
},
),
childrens: [],
physical_option: Some(
SeqScan,
),
_output_schema_ref: None,
},
],
physical_option: Some(
Filter,
),
_output_schema_ref: None,
},
],
physical_option: Some(
Project,
),
_output_schema_ref: None,
},
],
physical_option: Some(
Project,
),
_output_schema_ref: None,
},
],
physical_option: Some(
HashJoin,
),
_output_schema_ref: None,
},
],
physical_option: Some(
Project,
),
_output_schema_ref: None,
} |
Member
|
It seems to project twice on temp table. And I don't think it can handle more complex situations. explain SELECT b FROM t1 WHERE a in (SELECT a FROM t2 where (select a from t1));
PLAN
-------------------------------------------------------------------
Projection [t1.b] [Project] +
LeftSemi Join On t1.a = (t2.a) as (_temp_table_2_.a) [HashJoin]+
Scan t1 -> [a, b] [SeqScan] +
Projection [(t2.a) as (_temp_table_2_.a)] [Project] +
Projection [t2.a] [Project] +
Inner Join Where _temp_table_1_.a [HashJoin] +
Scan t2 -> [a] [SeqScan] +
Projection [(t2.a) as (_temp_table_1_.a)] [Project] +
Projection [t2.a] [Project] +
Scan t1 -> [] [SeqScan]I think every query using their own binder context may be a solution, but it may be a big work. |
Member
Author
Yes, let me create an issue for this problem to avoid affecting your PR. |
Member
Author
|
For the two Projects, I referred to Datafusion’s TableAlias and Project. |
Member
Author
|
close by #164 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What problem does this PR solve?
Fix code for：#164
Code changes
Check List
Tests
Side effects
Note for reviewer