Re: Query from HPM book to limit rows scanned doesn't appear to work as described

[Baron Schwartz]

>> But EXPLAIN is only a prediction.  If you look at the changes in the
>> Handler status variables, you'll see the second one reads fewer rows.
>>
>>
>
> Ok, I think I get it.  I first changed both of my queries to add
> "sql_no_cache" because without that, the Handler_read_rnd_next variable was
> zero in both cases.
>
> Before running each query, I ran "flush status", then the query, then "show
> session status like 'Handler%'".  The first one had a value of 207 for
> "Handler_read_rnd_next" and the second one had a value of 1.

Yes, that's it exactly.  The LIMIT stops the work as soon as you find
the first row.