evermuse / Automotive-Index

DB Indexing Exercise in PostgreSQL

Automotive Index

DB Indexing Exercise in PostgreSQL

When connecting to your pg database, connect from this project directory.

To run sql statements from an external file, use the \i [filepath] command.

Setting up

1. Create a new postgres user named indexed_cars_user
2. Create a new database named indexed_cars owned by indexed_cars_user
3. Run the provided scripts/car_models.sql script on the indexed_cars database
4. Run the provided scripts/car_model_data.sql script on the indexed_cars database 10 times
   there should be 223380 rows in car_models

Timing Select Statements

Enable timing queries in Postgres by toggling the \timing command in the psql shell.

1. Run a query to get a list of all make_title values from the car_models table where the make_code is 'LAM', without any duplicate rows, and note the time somewhere. (should have 1 result)
2. Run a query to list all model_title values where the make_code is 'NISSAN', and the model_code is 'GT-R' without any duplicate rows, and note the time somewhere. (should have 1 result)
3. Run a query to list all make_code, model_code, model_title, and year from car_models where the make_code is 'LAM', and note the time somewhere. (should have 1360 rows)
4. Run a query to list all fields from all car_models in years between 2010 and 2015, and note the time somewhere (should have 78840 rows)
5. Run a query to list all fields from all car_models in the year of 2010, and note the time somewhere (should have 13140 rows)

Indexing

Given the current query requirements, "should get all make_titles", "should get a list of all model_titles by the make_code", etc.
Create indexes on the columns that would improve query performance.

To add an index:

CREATE INDEX [index name]
  ON [table name] ([column name(s) index]);

Record your index statements in indexing.sql

Write the following statements in indexing.sql

1. Create a query to get a list of all make_title values from the car_models table where the make_code is 'LAM', without any duplicate rows, and note the time somewhere. (should have 1 result)
2. Create a query to list all model_title values where the make_code is 'NISSAN', and the model_code is 'GT-R' without any duplicate rows, and note the time somewhere. (should have 1 result)
3. Create a query to list all make_code, model_code, model_title, and year from car_models where the make_code is 'LAM', and note the time somewhere. (should have 1360 rows)
4. Create a query to list all fields from all car_models in years between 2010 and 2015, and note the time somewhere (should have 78840 rows)
5. Create a query to list all fields from all car_models in the year of 2010, and note the time somewhere (should have 13140 rows)

Compare the times of the queries before and after the table has been indexes.

Why are queries #4 and #5 not running faster?

Indexing on table create

1. Add your recorded indexing statements to the scripts/car_models.sql
2. Delete the car_models table
3. Run the provided scripts/car_models.sql script on the indexed_cars database
4. Run the provided scripts/car_model_data.sql script on the indexed_cars database 10 times
   there should be 223380 rows in car_models