Try not to re-caclculate indexes in requestReadingInOrder

[vdimir]

Changelog category (leave one):

• Bug Fix

Changelog entry (a user-readable short description of the changes that goes into CHANGELOG.md):

• Skip unnecessary extra index analysis when read-in-order optimization is applied.

Followup for #89815

---

Note

Medium Risk
Changes ReadFromMergeTree::requestReadingInOrder to stop calling selectRangesToRead(), which could affect how previously-computed analysis results are reused and thus impact read-order execution behavior if assumptions are wrong.

Overview
Avoids re-running index/range selection when switching a ReadFromMergeTree step to read-in-order.

Instead of calling selectRangesToRead() to recompute analysis, requestReadingInOrder() now directly updates the existing AnalysisResult::read_type (when present and not reading from a projection) based on the requested sort direction.

^{Written by Cursor Bugbot for commit 1dcc299. This will update automatically on new commits. Configure here.}

[clickhouse-gh]

Workflow [PR], commit [1dcc299]

Summary: ❌

job_name	test_name	status	info	comment
Upgrade check (amd_release)		failure
	Error message in clickhouse-server.log (see upgrade_error_messages.txt)	FAIL	cidb	IGNORED

[clickhouse-gh]

LLVM Coverage Report

Metric	Baseline	Current	Δ
Lines	83.20%	83.80%	+0.60%
Functions	23.60%	23.90%	+0.30%
Branches	75.30%	76.40%	+1.10%

PR changed lines: PR changed-lines coverage: 100.00% (11/11)
Diff coverage report
Uncovered code

[CurtizJ]

Let's backport this change. So, please write a changelog entry.