there is a hacking way I am using, leave it to be general list whenfirstly define the table, and then insert a sample data. in that way,q will recognize the table schema by checking the first record.here “I” means list of integers, this can be applied to “C” (string)as wellq)a:(c1:int$();c2:())q)insert [a; (1;1 2)],0q)meta ac | t f a–| -----c1| ic2| Iq)On Jun 4, 4:29?pm, Q_developer <qba…> wrote:> Hi ,>> I have declared an empty table as t:( c1:int$();c2:int$())> Now I want to insert compound column data like> c1 ? ? c2> – ? ? ? —> 10 ? ? 20 30> the statement t insert (10; 20 30) didn't work.But if I declare the table> without specifying null list of the type,then it is working.> i.e ?t:([] c1:int$();c2:()) works.> Is there any way to fix the type for the columns of the general lists> before inserting first item.>> Regards,</qba…>
To: personal-kdbplus@googlegroups.comX-Mailer: Apple Mail (2.1278)that’s no hacking. it is the way to define a column with nested data.On 4 Jun 2012, at 18:29, Sean wrote:> there is a hacking way I am using, leave it to be general list when> firstly define the table, and then insert a sample data. in that way,> q will recognize the table schema by checking the first record.> > here “I” means list of integers, this can be applied to “C” (string)> as well> > q)a:(c1:int$();c2:())\> q)insert [a; (1;1 2)]> ,0> q)meta a> c | t f a> --| -----> c1| i> c2| I> q)> > > On Jun 4, 4:29 pm, Q_developer <qba…> wrote:>> Hi ,>> >> I have declared an empty table as t:( c1:int$();c2:int$())>> Now I want to insert compound column data like>> c1 c2>> – —>> 10 20 30>> the statement t insert (10; 20 30) didn't work.But if I declare the table>> without specifying null list of the type,then it is working.>> i.e t:([] c1:int$();c2:()) works.>> Is there any way to fix the type for the columns of the general lists>> before inserting first item.>> >> Regards,> > – >