On 4 Sep 2013, at 1:43am, Jared Albers <jalb...@mymail.mines.edu> wrote:
> On my machine, when using relatively short table names like
> `TABLE_{table #}`, creation of a database with 10,000 tables takes
> approximately 14 seconds. These table names vary from 7 to a max of 11
> characters.
>
> When using relatively long table names like `TABLE_{table #}_{some
> unique identifying name that adds 120 or so characters}`, creation of
> a database with 10,000 tables takes approximately 60 seconds.
Although it might be nice to have the OP's own data, this is a consistently
reproducible characteristic. Just write a script or program to generate
sequential or random table names of different lengths. I used a spreadsheet to
generate corresponding 'CREATE TABLE' commands, then pasted them all in a text
file and used the shell tool.
I'm not certain this can really be characterised as a bug worth fixing.
Generating hashes will always take longer for longer text. Nobody expects
table names to be over 100 characters long. And the task described above is
obviously more suited to having all the data in one table, with an extra column
to distinguish which 'table number' is being used. But I'll leave that up to
the team.
Simon.
_______________________________________________
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
