Closed frabcus closed 10 years ago
For example, if more than (say) 100,000 rows, should take a sample at random, and all the facts make as much sense as now.
Example is Tweets matching 'banana': https://scraperwiki.com/dataset/bizrf5q/view/arxd2dy
This doesn't seem to be a problem in practice, although I'm not sure why not!
For example, if more than (say) 100,000 rows, should take a sample at random, and all the facts make as much sense as now.