Remove more unused code for pre-computed hashes

[raunaqmorarka]

Description

This code became unused after HashGenerationOptimizer was removed

Additional context and related issues

Follow-up to #26002

Release notes

(x) This is not user-visible or is docs only, and no release notes are required.
( ) Release notes are required. Please propose a release note for me.
( ) Release notes are required, with the following suggested text:

## Section
* Fix some things. ({issue}`issuenumber`)

[Copilot]

Pull Request Overview

This PR removes unused code related to pre-computed hashes following the deprecation of HashGenerationOptimizer, updating method signatures and eliminating the now‑unused Optional hashChannel parameters.

• Removed unused boolean flag and Optional hashChannel parameters from group-by hash and aggregation-related methods.
• Updated helper methods (e.g. selectGroupByHashMode) to reflect the removal of pre-computed hash support.
• Removed extraneous hash handling code in various operator implementations.

Reviewed Changes

Copilot reviewed 154 out of 154 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
core/trino-main/src/main/java/io/trino/operator/index/UnloadedIndexKeyRecordSet.java	Updated createGroupByHash call by removing an extra boolean parameter.
core/trino-main/src/main/java/io/trino/operator/exchange/LocalExchange.java	Removed unused Optional partitionHashChannel and related references.
core/trino-main/src/main/java/io/trino/operator/aggregation/partial/SkipAggregationBuilder.java	Eliminated the inputHashChannel parameter and its subsequent handling.
core/trino-main/src/main/java/io/trino/operator/GroupByHash.java	Adjusted selectGroupByHashMode signature to remove hasPrecomputedHash.
core/trino-main/src/main/java/io/trino/operator/BigintGroupByHash.java	Removed outputRawHash related calculations and hash writing logic.

Comments suppressed due to low confidence (5)

core/trino-main/src/main/java/io/trino/operator/index/UnloadedIndexKeyRecordSet.java:68

• Ensure that the removal of the second boolean parameter in the createGroupByHash call correctly matches the updated method signature and does not affect the intended behavior.

GroupByHash groupByHash = createGroupByHash(session, distinctChannelTypes, false, false, 10_000, hashStrategyCompiler, NOOP);

core/trino-main/src/main/java/io/trino/operator/GroupByHash.java:52

• Verify that all downstream calls to selectGroupByHashMode have been updated to use the new signature that omits the precomputed hash flag.

static GroupByHashMode selectGroupByHashMode(boolean spillable, List<Type> types)

core/trino-main/src/main/java/io/trino/operator/MarkDistinctOperator.java:80

• Remove the now-unneeded hashChannel parameter from the duplicate method to ensure the factory creation aligns with the updated constructor signature.

return new MarkDistinctOperatorFactory(operatorId, planNodeId, types.subList(0, types.size() - 1), markDistinctChannels, hashChannel, hashStrategyCompiler);

core/trino-main/src/main/java/io/trino/operator/BigintGroupByHash.java:137

• [nitpick] Confirm that the removal of the outputRawHash logic and related hash writing does not affect any consumers that might have depended on raw hash output.

if (outputRawHash) { ... }

core/trino-main/src/main/java/io/trino/operator/FlatHash.java:206

• [nitpick] Check that always computing the hash via flatHashStrategy.hash (after removal of precomputed hash handling) does not introduce performance regressions in high‑throughput scenarios.

long hash = flatHashStrategy.hash(blocks, position);