fix flatten regression

[sfc-gh-aling]

1. Which Jira issue is this PR addressing? Make sure that there is an accompanying issue to your PR.

   Fixes SNOW-NNNNNNN

2. Fill out the following pre-review checklist:

   • I am adding a new automated test(s) to verify correctness of my new code
     • If this test skips Local Testing mode, I'm requesting review from @snowflakedb/local-testing
   • I am adding new logging messages
   • I am adding a new telemetry message
   • I am adding new credentials
   • I am adding a new dependency
   • If this is a new feature/behavior, I'm adding the Local Testing parity changes.
   • I acknowledge that I have ensured my changes to be thread-safe. Follow the link for more information: Thread-safe Developer Guidelines
   • If adding any arguments to public Snowpark APIs or creating new public Snowpark APIs, I acknowledge that I have ensured my changes include AST support. Follow the link for more information: AST Support Guidelines

3. Please describe how your code solves the related issue.

Commit 07689e0 ("Loosen flattening rules for sort and filter") loosened can_clause_dependent_columns_flatten so that, when both _is_snowpark_connect_compatible_mode and _snowpark_connect_flatten_select_after_sort are True, a filter depending on a NEW (e.g. aliased) column in a subquery is allowed to be flattened.
That loosening is correct for sort / select / distinct, but it is unsound for filter across LIMIT/OFFSET. Flattening pushes the WHERE in front of the LIMIT, which changes the result set:

-- Input DataFrame (in Snowpark Connect compatible mode):
src.select(col("ID").alias("id_a"))
  .select(col("id_a").alias("id_b"))
  .limit(5)
  .filter(col("id_b") > 3)
-- Before 07689e09c (correct):            -- After 07689e09c (buggy):
SELECT … WHERE "id_b" > 3                 SELECT … WHERE "id_b" > 3 LIMIT 5
FROM (… LIMIT 5)                          --       ^^^^^ WHERE now applied
                                         --             BEFORE LIMIT

Repro SQL on Snowpark Connect: SELECT * FROM (SELECT * FROM range(10) LIMIT 5) WHERE id > 3 used to return [4], now returns [4, 5, 6, 7, 8].

Fix
Narrow the loosening that 07689e0 introduced instead of reverting it. Inside the NEW-column branch of can_clause_dependent_columns_flatten — which is only reachable when both compat-mode flags are True — add a single guard that rejects flattening when clause == "filter" and the subquery has LIMIT or OFFSET:

elif dc_state.change_state == ColumnChangeState.NEW:
  ...
  if not (
      context._is_snowpark_connect_compatible_mode
      and context._snowpark_connect_flatten_select_after_sort
  ):
      return False
  # NEW: WHERE cannot cross LIMIT/OFFSET without changing semantics.
  if clause == "filter" and subquery_has_limit_or_offset:
      return False

The helper gains one optional kwarg, subquery_has_limit_or_offset: bool = False, passed in from SelectStatement.filter. SelectStatement.sort already short-circuits on not self.limit_ / not self.offset at the call site, so no change is needed there.

Is this a BCR?
No. Existing Snowpark customers (legacy mode) observe no change in generated SQL. The only behavioral delta is for Snowpark Connect / compatible-mode callers on the exact pattern the bad commit newly enabled — where we are restoring pre-commit semantics.

[github-actions]

Thank you for your submission, we really appreciate it. Like many open-source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution. You can sign the CLA by just posting a Pull Request Comment same as the below format.

---

I have read the CLA Document and I hereby sign the CLA

---

_{You can retrigger this bot by commenting recheck in this Pull Request. Posted by the CLA Assistant Lite bot.}

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 95.16%. Comparing base (0aeccd9) to head (e2fccaa).

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #4197      +/-   ##
==========================================
- Coverage   95.41%   95.16%   -0.26%     
==========================================
  Files         171      171              
  Lines       43786    43788       +2     
  Branches     7502     7503       +1     
==========================================
- Hits        41778    41670     -108     
- Misses       1227     1295      +68     
- Partials      781      823      +42     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.