You're structure is bad for future growth (i.e. multiple languages) as the
query gets really nasty really quickly.
You should normalize this data and your query will never change.
pragma foreign_keys=ON;
CREATE TABLE buttons (
ID integer primary key autoincrement,
Key1 varchar no
On Tue, 15 Jan 2013 14:57:42 +0900, Yongil Jang wrote:
> SELECT t1.key1, t1.key2, t2.description FROM MASTERLANGUAGES as t1,
> MASTERLANGUAGES as t2 WHERE
> t1.ISOCode = 'DEU' and t2.ISOCode = 'ENG' and t1.key1 = t2.key1 and t1.key2 =
> t2.key2;
that is very close - just needed to add t1.descrip
like this?
sqlite> SELECT t1.key1, t1.key2, t2.description FROM MASTERLANGUAGES as t1,
MASTERLANGUAGES as t2 WHERE t1.ISOCode = 'DEU' and t2.ISOCode = 'ENG' and
t1.key1 = t2.key1 and t1.key2 = t2.key2;
FORM1|SAVE_BUTTON|Save
FORM1|HELP_BUTTON|Help
Sorry if my try is wrong.
2013/1/15 Kai Peters
Hello Dennis, Jay,
Thanks for you suggestions.
The ids are simply returned to the calling
application in a very large array. It's
the API we're writing against; so we can't
deviate from that unfortunately.
Thanks for clarifying the INSERT trigger
behavior.
I'll try your suggestions and see wha
Kervin L. Pierre wrote:
I'd appreciate any help or
pointers optimizing the SQL in the main
loop.
for( 'large number' ){
// query1
"SELECT id FROM table
WHERE attr1 = 'a', attr2 = 'b', ..."
if( 'query1 returns no rows' ){
if( 'create flag is on' ){
// query2
On 9/11/06, Kervin L. Pierre <[EMAIL PROTECTED]> wrote:
Basically, for every iteration we check
that an id exists ( ie. 'query1' ), if
it does we use it. But if the id does
not exist we insert a row ( ie. 'query2'
), then get the 'last_insert_rowid'
(ie. 'query3' ) and use that.
SQL works bes
6 matches
Mail list logo
