> The best database performance advice I ever got still applies: think
> about how the DBMS will traverse the data, and minimize the work it
> will do. It's all about I/O. No number of additional threads will
> bring the data into memory sooner.
The only way to optimize I/O is to not do it.
On Fri, 15 Apr 2016 22:53:57 +0100
Simon Slavin wrote:
> On 15 Apr 2016, at 10:05pm, Cecil Westerhof
> wrote:
>
> * SQLite datatypes and how SQLite decides which datatype you want
> * SQLite uses affinities not column types
> * thinking you should index each column instead of indexes for
On 2016/04/15 11:53 PM, Simon Slavin wrote:
I'm with Simon in response to Cecil on the idea of adding a shortlist of
known "bite" possibilities, something like the whentouse.html or the
howtocorrupt.html. Perhaps named commonmistakes.html or the like.
Maybe we could ask for contributions
I was ?bitten? by the fact that in SQLite primary keys can contain NULL
values. As I understood it, I was certainly not the first person to who
this happened. It would not surprise me if there are more deviations that
could spell problems for users of SQLite. That is why I think it would be a
good
On 15 Apr 2016, at 10:05pm, Cecil Westerhof wrote:
> I was ?bitten? by the fact that in SQLite primary keys can contain NULL
> values. As I understood it, I was certainly not the first person to who
> this happened. It would not surprise me if there are more deviations that
> could spell
5 matches
Mail list logo