docs(readme): drop Nullable(Array/Tuple) example that the compiler rejects

[lohanidamodar]

Summary

The Array/Tuple ClickHouse schema section in README.md (current main, post-#10) documented a ->array('scores', ColumnType::String)->nullable() example expected to compile to Nullable(Array(String)). The schema compiler at src/Query/Schema/ClickHouse.php:56-58 actually throws UnsupportedException("Nullable(Array(...)) is not supported in ClickHouse. Use an empty array [] to represent a missing value instead.") for this combination. The trailing prose also claimed "Nullable(...) wraps the whole Array/Tuple" — also wrong for both: Nullable(Tuple(...)) is rejected at line 80-82 with a pointer to Tuple(Nullable(T1), ...) and the experimental allow_experimental_nullable_tuple_type setting.

Changes

• Remove the ->array('scores', ...)->nullable() example from the code block.
• Remove the matching \scores` Nullable(Array(String))` line from the expected-SQL comment.
• Rewrite the trailing prose to accurately document both rejections (Nullable on Array, Nullable on Tuple) and the recommended workarounds.

Test plan

• No source/test changes; documentation-only.
• Verified the two error paths exist at src/Query/Schema/ClickHouse.php:56-58 and :80-82.
• Existing test testArrayRejectsNullableWrap (added in feat(clickhouse): add UUID, Decimal, Array/Tuple, UInt8/Int8, raw ORDER BY, rawColumn passthrough #10 via ebcdaba) pins the Array rejection behavior.

Related

• Discovered while investigating the unit test failure on feat(clickhouse): add UUID, Decimal, Array/Tuple, UInt8/Int8, raw ORDER BY, rawColumn passthrough #10 that was fixed by ebcdaba. The README claim slipped through because docs aren't covered by test snapshots.

[greptile-apps]

Greptile Summary

This documentation-only PR corrects a README example that described a ->array()->nullable() combination which the ClickHouse schema compiler actually rejects with an UnsupportedException. The updated prose now accurately documents both Nullable(Array(...)) and Nullable(Tuple(...)) rejections and their recommended workarounds.

• Removes the ->array('scores', ColumnType::String)->nullable() builder call and its corresponding scores Nullable(Array(String)) expected-SQL line from the code block.
• Rewrites the trailing paragraph to state that Nullable(Array(...)) is rejected (use [] as a sentinel) and Nullable(Tuple(...)) is experimental (use Tuple(Nullable(T1), ...) instead), matching the error messages thrown at src/Query/Schema/ClickHouse.php:56-58 and :80-82.

Confidence Score: 5/5

Documentation-only change that removes a broken example and replaces incorrect prose with accurate descriptions of both rejection paths, verified against the actual compiler source.

No source or test files are touched. Both claims in the new prose — the empty-array workaround for Nullable(Array(...)) and the Tuple(Nullable(T), ...) workaround for Nullable(Tuple(...)) — match the exact error messages thrown at src/Query/Schema/ClickHouse.php:56-58 and :80-82, so the documentation is now correct and consistent with the implementation.

No files require special attention.

Important Files Changed

Filename	Overview
README.md	Removes the invalid ->array('scores', ...)->nullable() example and rewrites the prose to accurately document both Nullable(Array(...)) and Nullable(Tuple(...)) rejections with their correct workarounds — both verified against src/Query/Schema/ClickHouse.php:56-58 and :80-82.

_{Reviews (1): Last reviewed commit: "docs(readme): drop Nullable(Array/Tuple)..." | Re-trigger Greptile}

[github-actions]

📊 Coverage

Metric	PR	Baseline	Δ
Lines	91.80% (7407/8069)	91.80%	+0.00%
Methods	84.42% (1100/1303)	84.42%	+0.00%
Classes	65.85% (135/205)	65.85%	+0.00%

Full per-file breakdown in the job summary.