This lab illustrates how minor changes in a query may have a significant impact on the execution plan.$$

The lab consists of a MySQL Workbench exercise followed by a zyLab exercise. The MySQL Workbench exercise is not graded.

MySQL Workbench exercise

Refer to the film, actor, and film$\_$actor tables of the Sakila database. This exercise is based on the default Sakila installation. If you have altered these tables or their data, your results may be different.

Step $1.$ Open MySQL Workbench and enter the following statements:

USE sakila; SELECT last$\_$name, first$\_$name, ROUND$($AVG$($length$),0)$ AS average FROM actor INNER JOIN film$\_$actor ON film$\_$actor.actor$\_$id $=$ actor.actor$\_$id INNER JOIN film ON film$\_$actor.film$\_$id $=$ film.film$\_$id WHERE title $=$ "ALONE TRIP" GROUP BY last$\_$name, first$\_$name ORDER BY average;

Step $2.$ Highlight the SELECT statement.

Step $3.$ In the main menu, select Query $>$ Explain Current Statement.

Step $4.$ In the Display Info box, highlighted in red below, select Data Read per Join. Workbench displays the execution plan for the SELECT statement:

$($ref the image$)$

The execution plan has seven parts, marked by red numbers on the screenshot:

Access a single film row using the idx$\_$title index on the title column.

Access matching film$\_$actor rows using the idx$\_$fk$\_$film$\_$id index on the film$\_$id foreign key.

Join the results using the nested loop algorithm.

Access actor rows via the index on the primary key.

Join actor rows with the prior join result using the MySQL nested loop algorithm.

Store the result in a temporary table and compute the aggregate function.

Sort and generate the result table.

Step $5.$Replace $=$ in the WHERE clause with $$ and generate a new execution plan. Part $1$ of the execution plan says Index Range Scan. The index scan accesses all films with titles preceding "ALONE TRIP", rather than a single film.

Step $6.$ Replace $$ in the WHERE clause with $>$ and generate a third execution plan. Part $1$ of the execution plan says Full Table Scan and accesses actor rather than film.

zyLab exercise

Step $7.$In Main.sql$,$ write EXPLAIN statements for the three SELECT statements, in the order described above. Run and submit the statements for testing.

Given the same query and data, EXPLAIN generates the same execution plan as Workbench Query $>$ Explain Current Statement. However, this lab uses a subset of Sakila data, so the execution plans of step $7$ and prior steps do not match exactly.

$--$ Initialize database

source Initialize.sql

$--$ Your EXPLAIN statements go here