Hello,

I have a view which is a result of the cross product of three tables, I want to 
test how much time is required to populate this view. Also, I want to test the 
scalability of this view in the future. Since, I have already live data I am 
wondering if I can do that without creating an automatic data generation i.e

time psql  -c 'SELECT * FROM view limit 100'
time psql  -c 'SELECT * FROM view limit 200'
...
time psql  -c 'SELECT * FROM view limit 10000'


Also, I would like to do a hot and cold tests. How Can I do the cold tests on 
Postgres, should I shut down and restart the server ? or there are some 
commands 
to clear the DBMS buffer.

How can I test this using a formula. The time to calculate the view is 
proportional to the cross product of the three tables i.e 


view (t) =  number of rows of table1 * number of rows table2 * number of rows 
table3.  Since I am doing cross product I assume a sequential scan will be used 
for data retrial from the H.D for all tables. What are other parameters I need 
to include in this formula and How I can get it. i.e I know that the row size 
is 
one factor and the time to read a data page on the H.D. is another one.

Regards

Reply via email to