Move extend clause after summarize for calculated measures

[AlisonNeedsCopilot]

Description

This PR ensures that the | extend clause is placed after | summarize in the generated KQL query. This ordering is necessary for calculated measures that reference summarized columns.

Changes

dialect_kql.py:
Reordered query construction in visit_select to output | summarize before | extend
tests/unit/test_dialect_kql.py:
Updated existing tests to expect summarize before extend

Why this change?

In KQL, extend operations that reference summarized columns must come after the summarize clause. Previously, the generated query placed extend before summarize, which would fail for calculated measures

UI Changes

Before:

After:

[Copilot]

Pull request overview

This PR fixes the ordering of KQL query clauses by moving the extend clause to come after the summarize clause. This change is necessary for calculated measures that reference columns created in the summarize operation, as KQL requires that extend operations referencing summarized columns must come after the summarize clause.

Changes:

• Modified query generation logic to output summarize before extend
• Updated existing tests to reflect the new query ordering
• Added a new test case to verify calculated measures work correctly

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
sqlalchemy_kusto/dialect_kql.py	Reordered query construction in visit_select to output summarize before extend; minor style improvement removing unnecessary parentheses
tests/unit/test_dialect_kql.py	Updated 4 existing tests to expect summarize before extend; added new test for calculated measures; minor formatting cleanup

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Pull request overview

Copilot reviewed 2 out of 2 changed files in this pull request and generated no new comments.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Pull request overview

Copilot reviewed 2 out of 2 changed files in this pull request and generated no new comments.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Egorrko]

Hello!
Thanks for your work on this PR, supporting post-aggregation calculations is definitely needed.

However, strictly moving the extend clause after summarize breaks standard pre-aggregation calculations (e.g., creating a calculated column and then grouping by it).

In KQL, summarize drops all columns that aren't in the by clause or the aggregations. So, if extend relies on raw columns, it will fail because those columns no longer exist after summarize.

Here is a test case that passes on the main branch but fails with these changes because the extend is placed too late:
from sqlalchemy import select, func, literal_column, text

def test_pre_aggregated_calculations():
    app_col = literal_column("App")
    ns_col = literal_column("Namespace")
    id_col = literal_column("_id")

    # 1. Calculated column before aggregation
    app_namespace = func.strcat(app_col, ns_col).label("App_Namespace")

    # 2. Group by the calculated column
    query = (
        select(app_namespace, func.count(id_col).label("TotalLogs"))
        .select_from(text("Logs"))
        .group_by(app_namespace)
    )

    query_compiled = str(
        query.compile(engine, compile_kwargs={"literal_binds": True})
    ).replace("\n", "")

    # Expected KQL (fails on this PR branch):
    query_expected = (
        '["Logs"]'
        '| extend ["App_Namespace"] = strcat(App, Namespace)'
        '| summarize ["TotalLogs"] = count(["_id"])  by ["App_Namespace"]'
        '| project ["App_Namespace"], ["TotalLogs"]'
    )
    assert query_compiled == query_expected