Question

I’m developing an app that takes raw transit data and displays the upcoming train schedule. I use the following query to retrieve the list of stations (from a table containing approximately 50,000 entries for stop times):

SELECT DISTINCT SUBSTR(stops._id, 0, LENGTH(stops._id)) as _id, stops.stop_name 
FROM trips, stop_times, stops                                  
WHERE trips._id = 'B'
AND trips.trip_id = stop_times._id 
AND stop_times.stop_id = stops._id

I ran this query in an SQLite browser and it executed in about 0.8 seconds. But when I execute it in my DatabaseHelper class using rawQuery(), it takes in excess of 5 minutes on my physical device. I added some logging statements to see where the slowdown was occurring and it appears that it is the query itself which is taking a long time, not the inflation of the View. When I execute the same query without the DISTINCT option, it executes in about 1 second.

I could understand the Android library taking a bit longer to execute than the browser but it should not be several hundred times slower. Is there something about the DISTINCT option that is causing this?

Answer 1

DISTINCT takes time to process as it is a matter of sorting and selection. In case if you have huge data, DISTINCT will just make it even worse. I would suggest you to breakdown your query into smaller parts and/or do the filtering pragmatically