Source code for deephaven.dbc.odbc

#
# Copyright (c) 2016-2025 Deephaven Data Labs and Patent Pending
#
"""This module supports ingesting data from external relational databases into Deephaven via the Python DB-API 2.0
(PEP 249) and the Open Database Connectivity (ODBC) interfaces by using the Turbodbc module.

Turbodbc is DB-API 2.0 compliant, provides access to relational databases via the ODBC interface and more
importantly it has optimized, built-in Apache Arrow support when fetching ODBC result sets. This enables Deephaven to
achieve maximum efficiency when ingesting relational data. """
import warnings

from deephaven import DHError
from deephaven import arrow as dharrow
from deephaven.table import Table

try:
    import turbodbc
    # this import is necessary to ensure that the turbodbc module is built with the Apache Arrow support
    import turbodbc.cursor
except ImportError:
    warnings.warn(message="import turbodbc failed, please install turbodbc with arrow support, the ODBC driver manager,"
                          "and the target database driver first.", stacklevel=2)
    turbodbc = None

def read_cursor(cursor: 'turbodbc.cursor.Cursor') -> Table:
    """Converts the result set of the provided cursor into a Deephaven table.

    Args:
        cursor (turbodbc.cursor.Cursor): a Turbodbc cursor. Prior to it being passed in, its execute() method must be
            called to run a query operation that produces a result set

    Returns:
        a new Table

    Raises:
        DHError, TypeError
    """
    if turbodbc is None:
        raise DHError(message="turbodbc is not installed, please install it, the ODBC driver manager, and the target "
                          "database driver first before calling this function.")

    if not isinstance(cursor, turbodbc.cursor.Cursor):
        raise TypeError(f"expect {turbodbc.cursor.Cursor} got {type(cursor)} instead.")
    try:
        pa_table = cursor.fetchallarrow()
    except Exception as e:
        raise DHError(e, message="failed to fetch ODBC result as a Arrow table.") from e

    return dharrow.to_table(pa_table)