On 11 Oct 2009, at 9:04am, Ron Arts wrote:
> CREATE TABLE company(id INTEGER PRIMARY KEY ASC, name)
>
> Then I insert 500000 records like this:
>
> INSERT INTO company (id, name) VALUES ('1', 'Company name number 1')
>
> (with consecutive values for the id value.)
I think you can remove the single quotes from around your value for
the id column, because both the column definition and the values
you're supplying for it are integers rather than strings. This might
speed up your INSERT even more.
> more about
> select speed
When testing SELECT speeds to work out how best to use the library,
include code to take the values that are returned and put them into
some variables in whatever language you're using. In some languages,
using some compilers, and with some optimisations turned on, the
SELECT command itself executes quickly but extracting the values to
variables takes longer.
Also, note that if you don't use the values from the variables some
compilers (e.g. recent versions of gcc) will optimise out the routines
which are meant to set the values. So if you're running comparative
speed tests and getting weird results try using the variables, for
example printing them to /dev/null.
Simon.
_______________________________________________
sqlite-users mailing list
[email protected]
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
