SELECT

e$.$first$\_$name,

e$.$last$\_$name,

t$.$title,

AVG$($s$.$salary$)$ AS average$\_$salary

FROM

employees e

JOIN

salaries s ON e$.$emp$\_$no $=$ s$.$emp$\_$no

JOIN

titles t ON e$.$emp$\_$no $=$ t$.$emp$\_$no

JOIN

dept$\_$emp de ON e$.$emp$\_$no $=$ de$.$emp$\_$no

WHERE

s$.$from$\_$date BETWEEN $\text{'}1999-01-01\text{'}$ AND $\text{'}2001-12-31\text{'}$

AND t$.$to$\_$date $>\text{'}2001-12-31\text{'}$

AND de$.$dept$\_$no $=\text{'}$d$005\text{'}$

GROUP BY

e$.$first$\_$name,

e$.$last$\_$name,

t$.$title

HAVING

AVG$($s$.$salary$)>($SELECT AVG$($salary$)*1\mathrm{.}2$ FROM salaries WHERE from$\_$date BETWEEN $\text{'}1999-01-01\text{'}$ AND $\text{'}2001-12-31\text{'})$

ORDER BY

average$\_$salary DESC;

a$)$ Without using indexes, create histograms on one $($or many$)$ columns that will improve query performance. Show the query cost

$($and query time$)$ prior to the creation of the histograms and after the creation of the histograms

b$)$ Explain the results. Were you able to show improvement? Why or why not?

c$)$ If the WHERE clause could be modified with a different date range of your choosing, based on the data distribution, could you

achieve better performance?