random-sampling.sql

-- Note: example given is for SQLite but can be adapted to other SQL

-- This is a fast query to sample random rows
-- It gets a list of existing row identifiers (rowids),
-- materializes, sorts, and picks 10.
SELECT * FROM test
    WHERE rowid IN
        (SELECT rowid FROM test
            ORDER BY random() LIMIT 10);

-- If the table has many, many rows, say hundreds of millions,
-- materializing the list and sorting it becomes expensive.
-- The size of the list of ids can be reduced. But don't
-- overdo it or you'll affect the sampling properties!
SELECT * FROM test JOIN
    (SELECT rowid as rid FROM test
        WHERE random() % 10 = 0  -- Reduce rowids by 10x
        ORDER BY random() LIMIT 10) AS srid
    ON test.rowid = srid.rid;

-- The magic reduction value can be calculated with count() of the table
-- and finding a reasonable number to reduce to a million (for example).
-- But you have to also be careful for tables with less than a million rows.
-- This is left as exercise.


-- To populate a table to play with this you can do
-- https://stackoverflow.com/a/28139886