feat(query): virtual column support materialized cte

[b41sh]

I hereby agree to the terms of the CLA available at: https://docs.databend.com/dev/policies/cla/

Summary

Materialized CTEs can hide the original Variant column path from the virtual column rewrite. For example, after a CTE projects v['message'] AS message, downstream expressions such as message['attribute']['user_id'] are bound against the materialized CTE output instead of the original table column. As a result, the planner can no longer recover the full path v['message']['attribute']['user_id'], so virtual column pushdown is missed.

This also affects large chained CTE queries where a reused CTE is auto-materialized: downstream consumers may repeatedly read large intermediate Variant objects instead of extracting only the required nested fields earlier.

This PR adds a binder-side rewrite for auto-materialized CTEs that preserves static Variant path access across CTE boundaries.

The following examples show the original SQL and the rewritten SQL

CREATE TABLE t(v Variant) enable_virtual_column=true;

# original SQL
WITH logs AS (
    SELECT v['message'] AS message FROM t
)
SELECT message['attribute']['user_id'] FROM logs;

# rewrited SQL
WITH logs AS (
    SELECT v['message']['attribute']['user_id'] AS __databend_virtual_column__0 FROM t
)
SELECT __databend_virtual_column__0 FROM logs;

SETTINGS (
  enable_experimental_virtual_column = 1,
  enable_auto_materialize_cte = 1
)
EXPLAIN WITH logs AS (
    SELECT v['message'] AS message FROM t
)
SELECT message['attribute']['user_id'] FROM logs;
╭─────────────────────────────────────────────────────────────────╮
│                             explain                             │
│                              String                             │
├─────────────────────────────────────────────────────────────────┤
│ TableScan                                                       │
│ ├── table: default.default.t                                    │
│ ├── scan id: 0                                                  │
│ ├── output columns: [v['message']['attribute']['user_id'] (#2)] │
│ ├── read rows: 0                                                │
│ ├── read size: 0                                                │
│ ├── partitions total: 0                                         │
│ ├── partitions scanned: 0                                       │
│ ├── push downs: [filters: [], limit: NONE]                      │
│ ├── virtual columns: [v['message']['attribute']['user_id']]     │
│ └── estimated rows: 0.00                                        │
╰─────────────────────────────────────────────────────────────────╯

other changes

• add_internal_column_into_expr and add_virtual_column_into_expr have been modified to add internal_column and virtual_column in batches, thereby avoiding redundant operations
• The column_mapping of MaterializedCTERef has been changed from HashMap to BTreeMap to ensure the stability of the EXPLAIN results

fixes: #[Link the issue here]

Tests

• Unit Test
• Logic Test
• Benchmark Test
• No Test - Explain why

Type of change

• Bug Fix (non-breaking change which fixes an issue)
• New Feature (non-breaking change which adds functionality)
• Breaking Change (fix or feature that could cause existing functionality not to work as expected)
• Documentation Update
• Refactoring
• Performance Improvement
• Other (please describe):

---

This change is 

[github-actions]

🤖 CI Job Analysis

Workflow: 26674127669

📊 Summary

• Total Jobs: 88
• Failed Jobs: 2
• Retryable: 0
• Code Issues: 2

❌ NO RETRY NEEDED

All failures appear to be code/test issues requiring manual fixes.

🔍 Job Details

• ❌ linux / sqllogic / cluster (tpcds, 4c, hybrid): Not retryable (Code/Test)
• ❌ linux / sqllogic / cluster (tpcds, 4c, http): Not retryable (Code/Test)

---

🤖 About

Automated analysis using job annotations to distinguish infrastructure issues (auto-retried) from code/test issues (manual fixes needed).

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: c79fbf971f

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".