fix: use absolute column index for DTypeMap and support deferred selections for tables with use_columns

[lukapeschke]

Follow-up work on #438 , #437 and #434 . Work items:

• Ensures that use_columns can be used with table in excel range form "A:B"
• Ensures that indices for dtypes parameter are also absolute
• Increased test coverage for use_columns & dtypes usage for both Python and Rust APIs
• Fixed a flaky test: sometimes mocking getLogger would not work, as fastexcel's logger could already be instantiated if tests running before test_fallback_infer_dtypes logged a message

[Copilot]

Pull request overview

This PR fixes column selection and dtype handling for Excel tables and sheets with offsets, ensuring that absolute column indices are used consistently throughout the codebase. The main changes include support for deferred selections (range-based column selections like "A:B" or "D:") on tables and using absolute indices for dtype lookups.

Key changes:

• Fixed dtype lookups to use absolute column indices instead of relative indices
• Added support for deferred column selections (range-based) on Excel tables
• Improved test coverage for column selection and dtype handling with offset tables
• Fixed flaky test by using pytest's caplog instead of mocker for logging assertions

Reviewed changes

Copilot reviewed 6 out of 6 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
tests/column_selection.rs	Added comprehensive Rust tests for column selection with offset tables and dtypes
src/types/exceltable/mod.rs	Added helper function and deferred selection support for tables
src/types/excelsheet/mod.rs	Extracted deferred selection resolution logic into reusable function
src/types/excelsheet/column_info/mod.rs	Fixed dtype and column name lookups to use absolute indices
python/tests/test_dtypes.py	Fixed flaky logger test using caplog instead of mocker
python/tests/test_column_selection.py	Updated dtype expectations and added Python tests for column selection with dtypes

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

The function returns a table-relative index when looking up by name, but later compares it with absolute indices. When a column name is found in table_columns at position idx, this idx is relative to the table (0-based), not absolute. However, on line 70, the code checks if selected_column_indices contains (idx + column_offset), which is an absolute index. This mixing of relative and absolute indices will cause columns looked up by name to be incorrectly filtered out. The name lookup should return (idx + column_offset) to convert to an absolute index, matching the behavior of IdxOrName::Idx.

Suggested change

	.find_map(|(idx, col_name)| (col_name.as_str() == name.as_str()).then_some(idx))
	.find_map(|(idx, col_name)| {
	(col_name.as_str() == name.as_str()).then_some(idx + column_offset)
	})