apache / apache/datafusion-sqlparser-rs
Keywords are parsed as functions in select statements
- 主要语言
- Rust
- 星标
- 3.5k
- 派生
- 772
- 平均合并
- 4 天 9 小时
- 30 天内合并 PR
- 17
描述
# Abstract
Qualified names like of the form `.` are sometimes parsed as functions, like in:
```sql
SELECT t.user FROM t;
```
# Current Behavior
This is done with `cargo run --example cli -- test.sql --generic`:
```sql
select t.user from t;
```
results in:
```text
[
Query(
Query {
with: None,
body: Select(
Select {
select_token: TokenWithSpan {
token: Word(
Word {
value: "select",
quote_style: None,
keyword: SELECT,
},
),
span: Span(Location(1,1)..Location(1,7)),
},
distinct: None,
top: None,
top_before_distinct: false,
projection: [
UnnamedExpr(
Function(
Function {
name: ObjectName(
[
Identifier(
Ident {
value: "t",
quote_style: None,
span: Span(Location(1,8)..Location(1,9)),
},
),
Identifier(
Ident {
value: "user",
quote_style: None,
span: Span(Location(1,10)..Location(1,14)),
},
),
],
),
uses_odbc_syntax: false,
parameters: None,
args: None,
filter: None,
null_treatment: None,
over: None,
within_group: [],
},
),
),
],
into: None,
from: [
TableWithJoins {
relation: Table {
name: ObjectName(
[
Identifier(
Ident {
value: "t",
quote_style: None,
span: Span(Location(1,20)..Location(1,21)),
},
),
],
),
alias: None,
args: None,
with_hints: [],
version: None,
with_ordinality: false,
partitions: [],
json_path: None,
sample: None,
index_hints: [],
},
joins: [],
},
],
lateral_views: [],
prewhere: None,
selection: None,
group_by: Expressions(
[],
[],
),
cluster_by: [],
distribute_by: [],
sort_by: [],
having: None,
named_window: [],
qualify: None,
window_before_qualify: false,
value_table_mode: None,
connect_by: None,
flavor: Standard,
},
),
order_by: None,
limit_clause: None,
fetch: None,
locks: [],
for_clause: None,
settings: None,
format_clause: None,
pipe_operators: [],
},
),
]
```
and `cargo run --example cli -- test.sql --postgres` is the same:
```text
[
Query(
Query {
with: None,
body: Select(
Select {
select_token: TokenWithSpan {
token: Word(
Word {
value: "select",
quote_style: None,
keyword: SELECT,
},
),
span: Span(Location(1,1)..Location(1,7)),
},
distinct: None,
top: None,
top_before_distinct: false,
projection: [
UnnamedExpr(
Function(
Function {
name: ObjectName(
[
Identifier(
Ident {
value: "t",
quote_style: None,
span: Span(Location(1,8)..Location(1,9)),
},
),
Identifier(
Ident {
value: "user",
quote_style: None,
span: Span(Location(1,10)..Location(1,14)),
},
),
],
),
uses_odbc_syntax: false,
parameters: None,
args: None,
filter: None,
null_treatment: None,
over: None,
within_group: [],
},
),
),
],
into: None,
from: [
TableWithJoins {
relation: Table {
name: ObjectName(
[
Identifier(
Ident {
value: "t",
quote_style: None,
span: Span(Location(1,20)..Location(1,21)),
},
),
],
),
alias: None,
args: None,
with_hints: [],
version: None,
with_ordinality: false,
partitions: [],
json_path: None,
sample: None,
index_hints: [],
},
joins: [],
},
],
lateral_views: [],
prewhere: None,
selection: None,
group_by: Expressions(
[],
[],
),
cluster_by: [],
distribute_by: [],
sort_by: [],
having: None,
named_window: [],
qualify: None,
window_before_qualify: false,
value_table_mode: None,
connect_by: None,
flavor: Standard,
},
),
order_by: None,
limit_clause: None,
fetch: None,
locks: [],
for_clause: None,
settings: None,
format_clause: None,
pipe_operators: [],
},
),
]
```
# Expected Behavior
I think this should parse similar to
```sql
select t.foo from t;
```
with `cargo run --example cli -- test.sql --generic`:
```text
[
Query(
Query {
with: None,
body: Select(
Select {
select_token: TokenWithSpan {
token: Word(
Word {
value: "select",
quote_style: None,
keyword: SELECT,
},
),
span: Span(Location(1,1)..Location(1,7)),
},
distinct: None,
top: None,
top_before_distinct: false,
projection: [
UnnamedExpr(
CompoundIdentifier(
[
Ident {
value: "t",
quote_style: None,
span: Span(Location(1,8)..Location(1,9)),
},
Ident {
value: "foo",
quote_style: None,
span: Span(Location(1,10)..Location(1,13)),
},
],
),
),
],
into: None,
from: [
TableWithJoins {
relation: Table {
name: ObjectName(
[
Identifier(
Ident {
value: "t",
quote_style: None,
span: Span(Location(1,19)..Location(1,20)),
},
),
],
),
alias: None,
args: None,
with_hints: [],
version: None,
with_ordinality: false,
partitions: [],
json_path: None,
sample: None,
index_hints: [],
},
joins: [],
},
],
lateral_views: [],
prewhere: None,
selection: None,
group_by: Expressions(
[],
[],
),
cluster_by: [],
distribute_by: [],
sort_by: [],
having: None,
named_window: [],
qualify: None,
window_before_qualify: false,
value_table_mode: None,
connect_by: None,
flavor: Standard,
},
),
order_by: None,
limit_clause: None,
fetch: None,
locks: [],
for_clause: None,
settings: None,
format_clause: None,
pipe_operators: [],
},
),
]
```
# Others
Also see PostgreSQL:
```text
psql (17.5 (Debian 17.5-1.pgdg120+1))
Type "help" for help.
postgres=# create table t(a int, "user" text);
CREATE TABLE
postgres=# insert into t values (1, 'foo');
INSERT 0 1
postgres=# select t.user from t;
user
------
foo
(1 row)
```
Interestingly, even the plain `user` works:
```text
postgres=# select user from t;
user
----------
postgres
(1 row)
```
# References
- a potentially valid case of treating `user` as a function: #1658
- downstream: https://github.com/apache/datafusion/issues/14141
贡献指南
这个仓库没有索引到贡献指南
调研方向
使用 test.sql 以及 --generic 和 --postgres,复现使用所提到的 cargo run --example cli 命令时出现的问题。跟踪 SQL 表达式解析过程,该过程将 t.user 分类为 Function,而不是 t.foo 所示的 CompoundIdentifier,然后为失败的查询添加覆盖,并验证两种方言都会生成预期的解析形式。
由索引模型根据 Issue 内容生成。
评估
- 技术栈
- rust, sql
- 领域
- compilers
- Issue 类型
- 缺陷
- 难度
- 3/5
- 预计耗时
- 1-2 天
- 活跃度
- 停滞
- 描述清晰度
- 基本清楚
- 新手友好度
- 45/100