Go to market

[yash-chauhan-dev]

What does this PR do?

Fixes #

Changes

How was it tested?

Anything reviewers should know?

To fix the problem, remove the unused timedelta symbol from the import statement so that only the actually used datetime is imported. This removes the unnecessary dependency and makes the code cleaner without changing behavior.

Concretely, in airflow-provider/example_dags/example_schema_dag.py, update line 10 from from datetime import datetime, timedelta to from datetime import datetime. No other changes are needed, since timedelta is not referenced anywhere else in the shown code. This keeps the DAG’s behavior identical while resolving the CodeQL finding.

In general, to fix an empty except you either (a) narrow the exception to the specific expected types and explain why it is safe to ignore them, or (b) handle the exception in a meaningful way, such as logging, then proceed with a safe fallback. Here, _ensure_numeric wants to be resilient: failures in pyarrow/decimal detection should simply mean “don’t convert, just return the original series”. We should keep that behavior but avoid swallowing unexpected errors silently.

The best fix with minimal functional change is:

• Narrow each except to a more specific set where reasonable (e.g. ImportError and AttributeError), but given limited context and to avoid changing behavior, we’ll keep Exception and add a small handling action.
• Log the exception at debug level using the logging module so users can diagnose issues when needed, but default behavior remains unaffected.
• Add an explanatory comment clarifying that on any error we fall back to returning the original series.

Concretely in datacheck/rules/numeric_rules.py:

• Add import logging near the top (without modifying existing imports).
• In the first try block (lines 20–27), replace the empty except Exception: pass with except Exception as exc: and add a comment and a logging call such as logging.getLogger(__name__).debug(...).
• Do the same in the second try block (lines 31–37).

This preserves existing logic: _ensure_numeric still returns the original series when conversion fails, but the exceptions are no longer entirely ignored.

General approach: Avoid bare “do-nothing” except Exception blocks. Either (a) narrow the exception type and document why it is safe to ignore, or (b) log/record the error while preserving the current non-failing behavior.

Best fix here without changing existing functionality:

• Keep the behavior that _ensure_numeric never raises from the decimal-detection logic and instead falls back to returning series.
• Add a comment explaining that failures are intentionally ignored because conversion is best-effort.
• Optionally capture the exception as e so future logging can be added, but to keep behavior strictly identical we will not log or re-raise.
• Apply the same pattern to both except blocks in _ensure_numeric (lines 20–27 and 31–37) for consistency.

Concretely in datacheck/rules/numeric_rules.py:

• Replace except Exception: followed by pass in the Arrow/pyarrow block with except Exception: # best-effort Arrow decimal handling and include a short explanatory comment inside the block.
• Replace the except Exception: in the decimal.Decimal block the same way.

We do not need new imports or helper methods; the changes are limited to these except blocks and comments.

In general, to fix empty except blocks, either (a) narrow the exception type and handle it explicitly, (b) add at least logging or a comment explaining why it is safe to ignore, or (c) re-raise after doing necessary cleanup. Here, the intent is to attempt an optimized PyArrow-based conversion but to fall back silently to pd.to_datetime if PyArrow is unavailable or the cast fails. The best fix without changing functionality is to:

• Add explanatory comments to both except blocks clarifying that the function will fall back to pd.to_datetime.
• Optionally narrow the outer exception to ImportError (the main expected failure), while still keeping behavior identical for callers. However, per instruction to avoid changing functionality, we’ll keep the broad catch but document it.

Because the only snippet we may edit is _to_datetime_fast in datacheck/rules/temporal_rules.py, we will replace the two except Exception: pass blocks with versions that include clear comments explaining the intentional fallbacks. We will not add imports or logging libraries (to avoid new dependencies and behavior changes). No new methods or definitions are needed; we only modify these lines in this file.