Fix various places in PerfStatsAnalysis.sql - minor issues

[PiJoCoder]

This PR applies a set of small correctness and robustness fixes to PerfStatsAnalysis.sql, primarily around wait stats calculations and a few stored-procedure logic issues.

Changes:

• Harden wait stats outputs (e.g., clamp negative wait deltas; exclude zero/negative waits from “top” category selection; use BIGINT for available CPU seconds).
• Fix query logic issues (e.g., correct boolean grouping in blocking-chain time filtering; fix cursor loop condition).
• Prevent runtime errors / logic inversions (e.g., NULLIF to avoid divide-by-zero; correct COL_LENGTH null check; remove stray debug SELECT).

[Copilot]

Pull request overview

This PR applies a set of small correctness and robustness fixes to PerfStatsAnalysis.sql, primarily around wait stats calculations and a few stored-procedure logic issues.

Changes:

• Harden wait stats outputs (e.g., clamp negative wait deltas; exclude zero/negative waits from “top” category selection; use BIGINT for available CPU seconds).
• Fix query logic issues (e.g., correct boolean grouping in blocking-chain time filtering; fix cursor loop condition).
• Prevent runtime errors / logic inversions (e.g., NULLIF to avoid divide-by-zero; correct COL_LENGTH null check; remove stray debug SELECT).

---

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

[PiJoCoder]

@copilot apply changes based on the comments in this thread

[Copilot]

@copilot apply changes based on the comments in this thread

Applied in commit 9e57bde — updated the usp_SmallSampledStats description text to use NULLIF([rows], 0) so it matches the safe predicate and users following the guidance won't hit a divide-by-zero.

[PiJoCoder]

@copilot: can you summarize and update the description in the PR of all the changes/commits in this PR

[Copilot]

@copilot: can you summarize and update the description in the PR of all the changes/commits in this PR

Updated the PR description with a full summary of all changes across the three commits. Here's the breakdown:

Wait stats hardening

• Clamp negative wait-time deltas to 0 before computing wait_time_ms_per_sec
• Use NULLIF(DATEDIFF(...)+1, 0) as divisor to avoid divide-by-zero on invalid intervals
• Change @avail_cpu_time_sec from INT → BIGINT (both wait stats procs)
• Exclude zero/negative wait entries and SOS_SCHEDULER_YIELD from the TOP 5 subquery
• Add missing DROP PROC guard before CREATE PROC DataSet_WaitStats_WaitStatsTopCategoriesOther

Query logic fixes

• Fix boolean precedence in blocking-chain time filter (AND (...OR...))
• Fix cursor WHILE condition from <> -1 to = 0

Runtime error / logic-inversion fixes

• usp_SmallSampledStats: use NULLIF([rows], 0) in both the predicate and the recommendation text
• usp_BatchSort: fix inverted COL_LENGTH null check (IS NULL → IS NOT NULL)
• usp_AutoUpdateStats: remove stray debug SELECT * that leaked a result set

[Copilot]

Pull request overview

Copilot reviewed 1 out of 1 changed files in this pull request and generated 1 comment.

---

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