thx Ibrahim. Give me some time to digest yours and other suggestions.
But it seems I will end up with a adjunct list PLUS something. Of course
you are right: I need to store many other information for each animal. I
definitely need to use a database (sqlite of course). It will be used
for storing breeding informations for agriculture livestock (no lab mice
or so similar). So the generation cycle is not too short.
Bye
Jan
Ibrahim A schrieb:
> One further advice :
>
> The fastest solution for your Problem would be to create a Array with
> fixed size Entries to describe the relationship between animals.
>
> in C you would simply end up with a struct like :
>
> struct ancestors {
> integer id_father ;
> integer id_mother ;
> } ;
>
> If you have a continuing range of animals this would create for 1 Mio
> animals you would end up with an array of 8 MB size. You can store this
> as a seperate file or a blob in a database and get the fastest possible
> accesstime for your problem.
>
> The question is then if you have a continuing range of animal id's.
>
> If you have no continuous range of animal_id's than i'd advice to create
> a indextable (symboltable) which translates animal_id's into array
> indexes. Sqlite could be used for this too where you declare animal_id
> as a numerical primary key instead of rowid. (UNIQUE PRIMARY KEY). No
> other fields should be included in this translation table so you'll have
> a maximum amount of key-index pairs on each page of your database and
> the translation so will require less page reads to get the correct array
> index. Well designed hash tables would be a bit faster for this goal but
> the difference would be minimal.
>
> _______________________________________________
> sqlite-users mailing list
> sqlite-users@sqlite.org
> http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
>
_______________________________________________
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
