Reading a Plan: Estimated vs Actual Rows
All lessons
- Reading a Plan: Estimated vs Actual RowsFree
- Selectivity for Ranges, NULLs and Combined Predicates Locked
- Histograms: Estimates for Skewed Data Locked
- Join Cardinality: How Many Rows a Join Returns Locked
- Nested Loops vs Hash Join: Counting the Work Locked
- Index Access Cost and the Clustering Factor Locked
- Bind Variables and the Cost of Hard Parsing Locked
- Bind Peeking and Adaptive Cursor Sharing Locked
- PL/SQL Blocks, Variables and IF Logic Locked
- Loops: WHILE, FOR, EXIT and CONTINUE Locked
- Explicit Cursors and the %NOTFOUND Trap Locked
- SQL%ROWCOUNT and SELECT INTO Rules Locked
- Exceptions: Where the Handler Sits Matters Locked
- Transactions, Savepoints and Statement-Level Rollback Locked
- Row by Row vs Set-Based: Counting Context Switches Locked
- BULK COLLECT with LIMIT and the Lost Last Batch Locked
- FORALL and SAVE EXCEPTIONS Locked
- Packages and Per-Session State Locked
- Flashback Query: Reading the Past Locked
- Undo Retention, Sizing and ORA-01555 Locked
- Capstone: Tuning a Nightly PL/SQL Job Locked
You are not signed in — the lab works, but your progress is not saved. Sign up
Read first
The code
SELECT /*+ gather_plan_statistics */ *
FROM shipments
WHERE status = :st
AND region = :rg;
SELECT *
FROM TABLE(DBMS_XPLAN.DISPLAY_CURSOR(
format => 'ALLSTATS LAST'));
-- Id | Operation | E-Rows | A-Rows
-- 0 | SELECT STATEMENT | |
-- 1 | TABLE ACCESS FULL | ??? | 9750Always read E-Rows and A-Rows side by side. If a line is off by a factor of 10 or more, fix that estimate before touching hints or indexes.
Loading the lab…