CHORE: Deduplicate tests while preserving coverage and edge cases

[Jahnvi Thakkar (jahnvi480)]

Work Item / Issue Reference

AB#49061

Not applicable; the ADO work item above tracks this change.

---

Summary

• Remove 29 overwritten test definitions and consolidate equivalent repeated scenarios without changing the surviving cursor/encoding test bodies.
• Consolidate Binary UTF-8 tests while retaining all 104 distinct string inputs and 7 raw-byte inputs, and explicitly test rejection of surrogate code points. Remove checks of Python's decoder that misleadingly claimed native conversion coverage.
• Preserve distinct database paths, NULL/LOB/encoding cases, and parameterized coverage; add a regression guard against same-scope duplicate test names.
• Change tests only: production code, coverage exclusions, pipeline configuration, and reporting configuration are unchanged.

Coverage: The authoritative combined Python/C++ reports retain exactly the same covered-line sets across all 44 production files: 9,419 / 11,068 lines (85.10%), with zero previously covered lines lost. The existing report provides line coverage, not branch coverage.

Validation: Windows native extension build passed; 326 targeted unit tests and 254 retained integration tests passed. python -m black --check --line-length=100 mssql_python tests passed (101 files). Exact comparison confirmed all original Binary string/byte inputs are retained once. Selected-suite collection changed from 1,607 to 1,559 items.

Cross-platform CI: Baseline run 180444 tested 1cf04a1317b07b861f3b42f4147404006b5953ba; cleanup run 180469 tested 69e9c49cef5c9490b90d121555dd2e83869d40d8. All 17 jobs passed after retries. The cleanup run used the unchanged PR-validation pipeline on the feature branch.

OS / configuration	Before	After	Change
Windows x64 - SQL 2022	5:32	4:29	-19.0%
Windows x64 - SQL 2025	7:08	7:04	-1.0%
Windows x64 - LocalDB	4:41	4:13	-9.9%
macOS - SQL 2022 (retry)	11:27	10:05	-11.9%
macOS - SQL 2025 (retry)	17:56	14:21	-20.0%
Ubuntu x64 - SQL 2022	3:14	2:43	-15.7%
Ubuntu x64 - SQL 2025	4:18	3:38	-15.7%
Debian x64 - SQL 2022	3:05	2:40	-13.6%
Debian x64 - SQL 2025	4:05	3:32	-13.5%
RHEL x64	3:15	2:39	-18.2%
Alpine x64	3:19	2:44	-17.3%
Ubuntu ARM64	13:57	13:59	+0.2%
Debian ARM64	18:00	17:13	-4.4%
RHEL ARM64	15:56	15:48	-0.8%
Alpine ARM64 (retry)	23:37	22:41	-4.0%

Times are successful pytest-step wall times, including coverage and command/container overhead but excluding queueing, dependency installation, builds, and failed-attempt costs. Both macOS jobs initially timed out during SQL setup before pytest; Alpine ARM64 initially failed the unchanged pooling-speed threshold. Those three jobs passed on attempt 2 without code changes. These are single-run observations with hosted-runner variability, not a demonstrated repeatable causal speedup.

[github-actions]

PR Performance Report

✅ No regression detected

No consistent slowdowns detected across all 2 environments.

0 IMPROVEMENTS 0 SLOWDOWNS 2/2 ENVIRONMENTS

Coverage: 2 of 2 environments completed. Advisory result; does not block merging.

Performance diagnostics

Phase times are inclusive diagnostics and must not be added together. They identify where measured time changed, not why it changed.

No affected phases or call-count changes were recorded.

All database tasks and timings

Unix / SQL Server 2022

Database task	Before	After	Paired change	Result
Connection opening	10.567 ms	10.677 ms	-1.2%	no signal
SELECT queries	1.095 ms	1.087 ms	+0.8%	no signal
Row insertion	35.301 ms	34.980 ms	+0.1%	no signal
Executemany inserts	156.553 ms	158.787 ms	+0.4%	no signal
Fetch-all queries	122.651 ms	122.038 ms	-0.5%	no signal
Row-by-row fetching	14.116 ms	14.599 ms	+1.9%	no signal
Batched row fetching	121.344 ms	117.190 ms	-1.9%	no signal
Transaction commit and rollback	118.289 ms	116.358 ms	-1.0%	no signal
Arrow row fetching	95.799 ms	95.593 ms	-0.2%	no signal
100,000-row insertion	441.653 ms	451.937 ms	+0.1%	no signal
Row fetching in batches of 100	122.727 ms	122.410 ms	+0.6%	no signal
Row fetching in batches of 10,000	135.773 ms	138.765 ms	+0.9%	no signal
Repeated positional queries	34.100 ms	34.375 ms	+1.5%	no signal
Repeated named-parameter queries	36.872 ms	36.837 ms	+0.1%	no signal
Legacy 100,000-row insertion	353.476 ms	356.769 ms	+1.0%	no signal
Insertion with explicit input sizes	502.133 ms	494.410 ms	-2.7%	no signal
Joined aggregation queries	179.443 ms	177.388 ms	-1.2%	no signal
Large joined-result fetching	192.315 ms	189.296 ms	+2.4%	no signal
1.2-million-row fetching	3530.148 ms	3471.510 ms	-0.7%	no signal
Common table expression queries	5.472 ms	5.509 ms	+2.4%	no signal
256 KiB VARCHAR(MAX) / fetchall()	1.345 ms	1.321 ms	+6.6%	no signal
10,000 scalar values / fetchval() (debug disabled)	107.297 ms	106.960 ms	+0.3%	no signal

Unix / SQL Server 2025

Database task	Before	After	Paired change	Result
Connection opening	99.343 ms	97.487 ms	-2.9%	no signal
SELECT queries	1.192 ms	1.089 ms	-6.8%	no signal
Row insertion	36.027 ms	33.982 ms	-5.7%	no signal
Executemany inserts	155.384 ms	156.106 ms	-0.4%	no signal
Fetch-all queries	127.381 ms	124.489 ms	-0.6%	no signal
Row-by-row fetching	14.247 ms	14.310 ms	+0.1%	no signal
Batched row fetching	119.379 ms	120.328 ms	+1.3%	no signal
Transaction commit and rollback	118.426 ms	114.973 ms	-1.4%	no signal
Arrow row fetching	96.445 ms	94.499 ms	-2.0%	no signal
100,000-row insertion	469.381 ms	448.026 ms	-3.0%	no signal
Row fetching in batches of 100	121.960 ms	123.059 ms	+1.9%	no signal
Row fetching in batches of 10,000	136.178 ms	136.845 ms	-0.7%	no signal
Repeated positional queries	33.629 ms	33.674 ms	-0.8%	no signal
Repeated named-parameter queries	38.804 ms	36.237 ms	-1.8%	no signal
Legacy 100,000-row insertion	375.193 ms	380.107 ms	+0.5%	no signal
Insertion with explicit input sizes	486.823 ms	510.615 ms	+3.0%	no signal
Joined aggregation queries	160.254 ms	163.556 ms	+2.1%	no signal
Large joined-result fetching	179.601 ms	193.693 ms	+1.2%	no signal
1.2-million-row fetching	3602.054 ms	3538.029 ms	-1.8%	no signal
Common table expression queries	5.213 ms	5.367 ms	+3.2%	no signal
256 KiB VARCHAR(MAX) / fetchall()	1.542 ms	1.436 ms	+0.8%	no signal
10,000 scalar values / fetchval() (debug disabled)	119.085 ms	111.734 ms	-9.1%	no signal

Build and measurement details

ADO build 180764

PR head: ef9bc0fe19601be3ced66196dda1124721e342a9
Base: 4c4195d4bb57c947cdab76334175bac4e023a828
Measured merge: 9cb98fbb5c9009076e84b71fb1ec5e91b1830f5b

• Unix / SQL Server 2022: Python 3.12.3, x86_64, SQL 16.0.4295.3; 5 paired comparisons and 1 warmup.
• Unix / SQL Server 2025: Python 3.12.3, x86_64, SQL 17.0.5005.3; 5 paired comparisons and 1 warmup.

A consistent change requires more than 20% median paired movement, at least 1 ms between the median runtimes, and at least 80% of pairs exceeding the relative threshold in the same direction. A slowdown without enough pair agreement is reported as inconsistent.

The displayed change is the median of paired before-and-after ratios. It is not recalculated from the two displayed median runtimes.

Both revisions use profiling-enabled builds on the same agent and database, with alternating order and discarded warmups. Results are diagnostic and do not represent production-wheel latency.

Raw samples and logs are attached to the ADO run as profiler-* artifacts.

[Copilot]

Copilot review overview

🟡 Changes recommended

The duplicate-name guard misses tests inside same-scope control-flow blocks, allowing silent overwrites.

Review effort: Balanced
Findings: 1

Open (1)

• Detect duplicate tests inside control-flow blocks · New

What changed in this PR

Deduplicates test coverage while preserving distinct scenarios and edge cases.

Changes:

• Removes redundant and overwritten tests.
• Consolidates Binary UTF-8 cases and adds surrogate rejection coverage.
• Adds an AST-based duplicate test-name guard.

File	Description
tests/​test_test_definitions.py	Adds duplicate-name detection.
tests/​test_015_pyformat_parameters.py	Removes redundant parameter tests.
tests/​test_014_ddbc_bindings_coverage.py	Deletes misleading conversion tests.
tests/​test_013_encoding_decoding.py	Removes an overwritten EUC-KR test.
tests/​test_012_connection_string_integration.py	Removes parser scenarios covered elsewhere.
tests/​test_010_connection_string_parser.py	Parameterizes unknown-keyword cases.
tests/​test_008_auth.py	Removes duplicate MSI coverage.
tests/​test_003_connection.py	Removes duplicate cursor cleanup coverage.
tests/​test_002_types.py	Consolidates Binary encoding edge cases.

---

💡 Configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[Copilot]

Copilot review overview

🟡 Changes recommended

The new duplicate-name guard misses duplicate Test* classes, which can still silently replace collected tests.

Review effort: Balanced
Findings: 2

Open (2)

• Detect duplicate pytest class names to prevent test loss · New
• Detect duplicate tests inside control-flow blocks

[Copilot]

Copilot review overview

🟡 Changes recommended

The duplicate guard misses overwritten test classes, and a distinct EUC-KR VARCHAR path is removed.

Review effort: Balanced
Findings: 2

Open (2)

• Duplicate test classes overwrite earlier class definitions · New
• Detect duplicate pytest class names to prevent test loss

Resolved since last review (1)

• Detect duplicate tests inside control-flow blocks

Previously missed (1)

In code that hasn't changed since last review

Korean VARCHAR SQL_CHAR encoding regression coverage was removed

tests/​test_013_encoding_decoding.py:3314

This removal drops the only case combining Korean str values with an EUC-KR SQL_CHAR configuration and a VARCHAR column. The retained SQL_CHAR/EUC-KR case at lines 2465-2603 only sends ASCII, while the Korean round-trip at lines 6840-6886 uses NVARCHAR/SQL_WCHAR; therefore the PR's stated preservation of distinct encoding/database paths is not met. Retain a Korean VARCHAR/SQL_CHAR execution case, ideally with a meaningful result assertion.

[Copilot]

Copilot review overview

🟡 Changes recommended

The duplicate-name guard misses collected unittest.TestCase classes not named with the Test* prefix.

Review effort: Balanced
Findings: 1

Open (1)

• Detect duplicate unittest.TestCase classes regardless of class name · New

Resolved since last review (2)

• Duplicate test classes overwrite earlier class definitions
• Detect duplicate pytest class names to prevent test loss

[github-actions]

📊 Code Coverage Report

🔥 Diff Coverage 100%

🎯 Overall Coverage 85%

📈 Total Lines Covered: 9445 out of 11094 📁 Project: mssql-python

---

Diff Coverage

Diff: main...HEAD, staged and unstaged changes

No lines with coverage information in this diff.

---

📋 Files Needing Attention

📉 Files with overall lowest coverage (click to expand)

mssql_python.pybind.performance_counter.hpp: 0.7%
mssql_python.pybind.logger_bridge.cpp: 57.9%
mssql_python.pybind.ddbc_bindings.h: 62.6%
mssql_python.pybind.logger_bridge.hpp: 70.8%
mssql_python.pybind.connection.connection_pool.cpp: 82.3%
mssql_python.pybind.connection.connection.cpp: 83.1%
mssql_python.logging.py: 86.2%
mssql_python.pooling.py: 90.1%
mssql_python.pybind.fetch_temporal.hpp: 92.1%
mssql_python.cursor.py: 92.5%

---

🔗 Quick Links

⚙️ Build Summary	📋 Coverage Details
View Azure DevOps Build	Browse Full Coverage Report

[Gaurav Sharma (bewithgaurav)]

the cleanup broadly addresses the ask, but the newly enabled test breaks the supported rhel runs. requesting changes to resolve that failure and close the remaining duplicate-class gap raised in the existing thread.

[Copilot]

Copilot review overview

🟡 Changes recommended

The PR description materially contradicts the included pipeline and environment configuration changes.

Review effort: Balanced
Findings: 1

Open (1)

• Disclose or revert unintended pipeline dependency changes · New

Resolved since last review (1)

• Detect duplicate unittest.TestCase classes regardless of class name

[Gaurav Sharma (bewithgaurav)]

the cleanup preserves the existing cases, and the follow-up resolves both review findings without weakening the assertions. approving.