Pandas read_sql vs dask read_sql issues about dask

[benrutter]

Is this still an issue for you? Does the same thing happen on the latest version of dask or is it just 2023.11?

Also, what versions of pandas + sqlalchemy are in your environment?

[frbelotto]

I´ve just tested against the latests versions :

dask:2024.4.1
pandas:2.2.2
sqla : 2.0.29

import pandas as pd
sql = """SELECT t1.NR_SEQL_SLCT_CPR
        FROM ORANDPOW0000.SLCT_CPR_PRD_PCR t1
        WHERE ROWNUM <= 1000"""
pd.read_sql(sql = sql, con=oracle.uri, index_col = 'nr_seql_slct_cpr')

import dask.dataframe as dd
sql = """SELECT t1.NR_SEQL_SLCT_CPR
        FROM ORANDPOW0000.SLCT_CPR_PRD_PCR t1
        WHERE ROWNUM <= 1000"""
dd.read_sql(sql = sql, con=oracle.uri, index_col = 'nr_seql_slct_cpr')

---------------------------------------------------------------------------
NoSuchTableError                          Traceback (most recent call last)
Cell In[15], [line 5](vscode-notebook-cell:?execution_count=15&line=5)
      [1](vscode-notebook-cell:?execution_count=15&line=1) import dask.dataframe as dd
      [2](vscode-notebook-cell:?execution_count=15&line=2) sql = """SELECT t1.NR_SEQL_SLCT_CPR
      [3](vscode-notebook-cell:?execution_count=15&line=3)         FROM ORANDPOW0000.SLCT_CPR_PRD_PCR t1
      [4](vscode-notebook-cell:?execution_count=15&line=4)         WHERE ROWNUM <= 1000"""
----> [5](vscode-notebook-cell:?execution_count=15&line=5) dd.read_sql(sql = sql, con=oracle.uri, index_col = 'nr_seql_slct_cpr')

File [c:\Users\fabio\AppData\Local\Programs\Python\Python311\Lib\site-packages\dask_expr\io\sql.py:7](file:///C:/Users/fabio/AppData/Local/Programs/Python/Python311/Lib/site-packages/dask_expr/io/sql.py:7), in read_sql(sql, con, index_col, **kwargs)
      [4](file:///C:/Users/fabio/AppData/Local/Programs/Python/Python311/Lib/site-packages/dask_expr/io/sql.py:4) def read_sql(sql, con, index_col, **kwargs):
      [5](file:///C:/Users/fabio/AppData/Local/Programs/Python/Python311/Lib/site-packages/dask_expr/io/sql.py:5)     from dask.dataframe.io.sql import read_sql
----> [7](file:///C:/Users/fabio/AppData/Local/Programs/Python/Python311/Lib/site-packages/dask_expr/io/sql.py:7)     df = read_sql(sql, con, index_col, **kwargs)
      [8](file:///C:/Users/fabio/AppData/Local/Programs/Python/Python311/Lib/site-packages/dask_expr/io/sql.py:8)     return from_legacy_dataframe(df)

File [c:\Users\fabio\AppData\Local\Programs\Python\Python311\Lib\site-packages\dask\dataframe\io\sql.py:394](file:///C:/Users/fabio/AppData/Local/Programs/Python/Python311/Lib/site-packages/dask/dataframe/io/sql.py:394), in read_sql(sql, con, index_col, **kwargs)
    [361](file:///C:/Users/fabio/AppData/Local/Programs/Python/Python311/Lib/site-packages/dask/dataframe/io/sql.py:361) """
    [362](file:///C:/Users/fabio/AppData/Local/Programs/Python/Python311/Lib/site-packages/dask/dataframe/io/sql.py:362) Read SQL query or database table into a DataFrame.
    [363](file:///C:/Users/fabio/AppData/Local/Programs/Python/Python311/Lib/site-packages/dask/dataframe/io/sql.py:363) 
   (...)
    [391](file:///C:/Users/fabio/AppData/Local/Programs/Python/Python311/Lib/site-packages/dask/dataframe/io/sql.py:391) read_sql_query : Read SQL query into a DataFrame.
    [392](file:///C:/Users/fabio/AppData/Local/Programs/Python/Python311/Lib/site-packages/dask/dataframe/io/sql.py:392) """
    [393](file:///C:/Users/fabio/AppData/Local/Programs/Python/Python311/Lib/site-packages/dask/dataframe/io/sql.py:393) if isinstance(sql, str):
--> [394](file:///C:/Users/fabio/AppData/Local/Programs/Python/Python311/Lib/site-packages/dask/dataframe/io/sql.py:394)     return read_sql_table(sql, con, index_col, **kwargs)
    [395](file:///C:/Users/fabio/AppData/Local/Programs/Python/Python311/Lib/site-packages/dask/dataframe/io/sql.py:395) else:
...
   [1541](file:///C:/Users/fabio/AppData/Local/Programs/Python/Python311/Lib/site-packages/sqlalchemy/engine/reflection.py:1541) if _reflect_info.table_options:

NoSuchTableError: SELECT t1.NR_SEQL_SLCT_CPR
        FROM ORANDPOW0000.SLCT_CPR_PRD_PCR t1
        WHERE ROWNUM <= 1000
Output is truncated. View as a [scrollable element](command:cellOutput.enableScrolling?43003cc2-5504-4dc1-91ab-a6d2b9db8d10) or open in a [text editor](command:workbench.action.openLargeOutput?43003cc2-5504-4dc1-91ab-a6d2b9db8d10). Adjust cell output [settings](command:workbench.action.openSettings?%5B%22%40tag%3AnotebookOutputLayout%22%5D)...