Preserve columns matching the first header during transpose

[Tamako-C]

Dataset.transpose() currently skips every column whose header matches the first header. Only the first column belongs in the new header row; later columns with the same name must still become data rows.

Skip the first column by position. Add regression tests for a repeated first header and for the case where an original cell value matches the first header, causing data loss after two transposes.

This addresses the remaining case reported by punchagan in #199 (comment). The earlier PR #199 is already merged; this is a follow-up fix.

Validation:

• Both regression tests fail against the original source and pass after the fix.
• Full suite on Python 3.12.14 with pandas 3.0.1 and all required format dependencies: 185 passed with process-local TZ=UTC, without skipped tests.
• In the host's UTC+8 timezone: 184 passed and one existing date-conversion test failed. Running that same test against the original HEAD reproduces the failure; it matches Fix timezone-dependent getDate timestamp test #675. No system timezone or test expectations were changed.
• Changed Python files pass local Ruff 0.11.9. After submission, pre-commit.ci passed using the repository's configuration, which pins Ruff 0.15.20.
• git diff --check passes.

AI assistance: OpenAI Codex assisted with the implementation, regression tests, and this PR description.