Sqlalchemy in statement is incorrectly compiled for the in clause

在 2.9.6 版本中，使用所示的 create_engine、select、statement compilation 和 conn.execute 進入點重現此問題。將 literal_binds 的輸出與參數化 SQL 以及執行期間傳送的參數進行比較。當包含 True 的 IN 子句能夠成功執行，且不存在 boolean 與 integer 之間的型別不相容時，即表示完成。

Version 2.9.6

Repro:

from sqlalchemy import create_engine, MetaData, Table, select, column

# Assuming you have a working database connection string
engine = create_engine('...'). # make sure to specify the catalog

metadata = MetaData()
engine.execute("create table if not exists bogdankyryliuk.bool_test as select 1 as one_val, True as bool_val")
test_table = Table('bool_test', metadata, autoload_with=engine, schema='bogdankyryliuk')
stmt = select(test_table).where(test_table.c.bool_val.in_([True]))

compiled_stmt = stmt.compile(compile_kwargs={"literal_binds": True})
print(compiled_stmt)

with engine.connect() as conn:
    result = conn.execute(stmt)
    for row in result:
        print(row)

Compiled statement is printed as expected:

SELECT bogdankyryliuk.bool_test.one_val, bogdankyryliuk.bool_test.bool_val
FROM bogdankyryliuk.bool_test
WHERE bogdankyryliuk.bool_test.bool_val IN (true)

However conn.execute(stmt) executes different statement:

DatabaseError: (databricks.sql.exc.ServerOperationError) [DATATYPE_MISMATCH.DATA_DIFF_TYPES] Cannot resolve "(bool_val IN (1))" due to data type mismatch: Input to `in` should all be the same type, but it's ["BOOLEAN", "INT"]. SQLSTATE: 42K09; line 3 pos 40
[SQL: SELECT bogdankyryliuk.bool_test.one_val, bogdankyryliuk.bool_test.bool_val
FROM bogdankyryliuk.bool_test
WHERE bogdankyryliuk.bool_test.bool_val IN (%(bool_val_1_1)s)]
[parameters: {'bool_val_1_1': 1}]
(Background on this error at: https://sqlalche.me/e/14/4xp6)

That leads to the exception