Fix | Build ReservedWords table in code

[edwardneal]

Description

This PR populates the ReservedWords collection returned by SqlConnection.GetSchema in code, rather than by deserializing the SqlMetaData.xml embedded resource.

There are currently 393 reserved words in this embedded resource. To make this easier to review, I've broken it down commit-by-commit.

I'm aiming to do nothing more than migrate the data out of the XML file, with one exception: the "NATIONAL " keyword is now corrected to "NATIONAL", removing the trailing space. I believe this was an oversight.

Separately to this, I've noticed two things:

• This list of words hasn't been kept in sync with the publicly documented set of reserved words. I've linked to the documentation and noted all missing words, but it's not clear at present whether they should all be added, or whether there are other criteria. If we add any, we'll need to make sure the minimum version is correctly specified.
• If we're to take the linked documentation as canonical, there's an additional reserved word which is unique to Azure Synapse Analytics. This highlights the requirement for Cleanup | Consolidate connection capability metadata #3862 (because we're no longer basing the list of reserved words strictly on the version of the SQL Server) and raises the question of whether we determine this with a query of SERVERPROPERTY('EngineEdition'), or a simpler check against the DNS name.

I've avoided pulling a response to either of them in this PR in order to prevent it drifting into the weeds and to keep it self-contained.

Issues

Contributes to #3372.

Testing

I've manually verified that reserved words match, and added a test which verifies that the data is correct. With adjustment to account for the "NATIONAL " case, this same test also passes against main.

[edwardneal]

This is the only reserved word I expect to be different between the old and the new structure. I've assumed that it was a typo.

[benrr101]

/azp run

[azure-pipelines]

Azure Pipelines successfully started running 2 pipeline(s).

[benrr101]

I like getting rid of xml files - especially when the list seems completely arbitrary :D

[benrr101]

I'd prefer it if these were separate tests (maybe a theory?) and VerifyReservedWordsTable was embedded into the test. But this isn't mandatory for approval.

[edwardneal]

Thanks @benrr101. It may be worth looking at #4362 before I start looking at changes. That PR is another approach which would also cover this PR's scope (and the rest of the SqlMetaData.xml scope's resources.)

I'd be happy enough to close this PR and tack tests on to the end of #4362, depending on the approach you want to work with.

[benrr101]

Sure, we'll triage that one tomorrow, and we can start looking into it. It sounds like #4362 is more complete, but more complicated to review as a result.

[codecov]

Codecov Report

❌ Patch coverage is 95.20000% with 6 lines in your changes missing coverage. Please review.
✅ Project coverage is 63.48%. Comparing base (c724554) to head (52c8a20).
⚠️ Report is 39 commits behind head on main.

Files with missing lines	Patch %	Lines
...Data/SqlClient/SqlMetaDataFactory.ReservedWords.cs	95.16%	6 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #4328      +/-   ##
==========================================
- Coverage   66.57%   63.48%   -3.10%     
==========================================
  Files         284      281       -3     
  Lines       43301    66371   +23070     
==========================================
+ Hits        28827    42133   +13306     
- Misses      14474    24238    +9764     

Flag	Coverage Δ
CI-SqlClient	?
PR-SqlClient-Project	63.48% <95.20%> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 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.