Binder error when external data source has a table with uppercase column name

[xzdandy]

Search before asking

• I have searched the EvaDB issues and found no similar bug report.

Bug

Example home_sales table in sqlite3:

    home_sales.saledate home_sales.MA home_sales.type home_sales.bedrooms
0              saledate            MA            type            bedrooms
1            30/09/2007        441854           house                   2
2            31/12/2007        441854           house                   2
3            31/03/2008        441854           house                   2
4            30/06/2008        441854           house                   2
..                  ...           ...             ...                 ...
343          30/09/2018        603281            unit                   3
344          31/12/2018        601167            unit                   3
345          31/03/2019        605637            unit                   3
346          30/06/2019        599339            unit                   3
347          30/09/2019        597884            unit                   3

When we run the following query in EvaDB:

SELECT * FROM sqlite_date.home_sales; 

It crashed with the following error message

Traceback (most recent call last):
  File "/home/zxu330/eva/evadb/executor/plan_executor.py", line 178, in execute_plan
    yield from output
  File "/home/zxu330/eva/evadb/executor/project_executor.py", line 34, in exec
    batch = apply_project(batch, self.target_list, self.catalog())
  File "/home/zxu330/eva/evadb/executor/executor_utils.py", line 42, in apply_project
    batches = [expr.evaluate(batch) for expr in project_list]
  File "/home/zxu330/eva/evadb/executor/executor_utils.py", line 42, in <listcomp>
    batches = [expr.evaluate(batch) for expr in project_list]
  File "/home/zxu330/eva/evadb/expression/tuple_value_expression.py", line 71, in evaluate
    return batch.project([self.col_alias])
  File "/home/zxu330/eva/evadb/models/storage/batch.py", line 246, in project
    assert len(unknown_cols) == 0, unknown_cols
AssertionError: ['home_sales.ma']

As we can see, the binder is looking for column home_sales.ma instead of home_sales.MA.

Environment

No response

Are you willing to submit a PR?

• Yes I'd like to help by submitting a PR!

[xzdandy]

node.col_alias = "{}.{}".format(table_alias, node.name.lower()) at evadb/binder/statement_binder.py makes the column name to lowercase that caused this issue.

@gaurav274 What is the motivation to turn the column name into lower case in alias instead of keeping it as original?

[americast]

@xzdandy @gaurav274 The issue seems to still persists:

In [3]: cursor.query("CREATE TABLE ABC(\                                                                                                              
   ...: HELLO INTEGER);").execute()                                                                                                                   
Out[3]: <evadb.models.storage.batch.Batch at 0x7fe0fc4c3bd0>               
                                                                           
In [4]: print(cursor.query("SELECT HELLO FROM ABC;").df())                                                                                            
09-28-2023 22:30:06 ERROR [statement_binder_context:statement_binder_context.py:raise_error:0145] Found invalid column hello

[xzdandy]

Thanks for reporting.