Map projected columns for ChunkTableSource

[wmTJc9IK0Q]

Fixes #66

This moves the projection mapping logic from Row to DataChunk and handles it consistently in one place for FillRow or FillChunk in table UDFs.

[Copilot]

Pull request overview

This PR refactors the projection mapping logic for table UDFs by moving it from the Row structure to the DataChunk structure. This consolidates projection handling in one place, making it consistent for both row-based (FillRow) and chunk-based (FillChunk) table UDF implementations.

Key changes:

• Moved projection mapping from Row to DataChunk for centralized handling
• Updated Row methods to delegate projection logic to the underlying DataChunk
• Added projection-aware logic to DataChunk's GetValue, SetValue, and SetChunkValue methods
• Separated projection-specific tests into dedicated test functions for better coverage

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 5 comments.

Show a summary per file

File	Description
table_udf.go	Assigns projection mapping to DataChunk instead of Row for consistent projection handling across all table UDF types
row.go	Removes projection field from Row struct and delegates projection logic to the underlying DataChunk
data_chunk.go	Adds projection field and implements projection-aware column access with automatic remapping of column indices
errors.go	Adds new error function for unprojected column access errors
table_udf_test.go	Adds dedicated tests for chunk-based and row-based projection pushdown, separates pushdownTableUDF test into standalone test function

---

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

[Copilot]

Pull request overview

Copilot reviewed 5 out of 5 changed files in this pull request and generated 5 comments.

---

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

[Copilot]

Missing bounds check for negative colIdx values when no projection is set. The code checks colIdx >= len(chunk.columns) but not colIdx < 0, which could lead to a panic. The check should be:

if colIdx < 0 || colIdx >= len(chunk.columns) {
    return getError(errAPI, columnCountError(colIdx, len(chunk.columns)))
}

[VGSML]

@wmTJc9IK0Q Could you try to avoid using nested if and if else construction. It's looks like code can be more readable

[wmTJc9IK0Q]

@VGSML Thanks for the review, I've cleaned this up.

[VGSML]

Looks good! @taniabogatsch could you take a look at

[taniabogatsch]

Thanks for the PR! I've left some comments and a few nits. :)