I’m confused about how aj works for non-time columns.t1:(a:(1 1); b:(4 5); d:(7 8))t2:(a:(2 1); b:(5 4); c:(7 9))aj[ab;t1;t2]a b d c-------1 4 7 91 5 8 9Why isn’t c null in the second row, there is no row in t2 with 1 and 5for a and b.
In case of aj[c1…cn;t1;t2], the last column cn is considered as weakcomparison, so for each record in t1, columns c1 to cn-1 shouldexactly match with columns in t2, and then value in cn is comparedwith the corresponding value of cn in the matching records. theresulting set contains all the matching records with values less thanor equal to the value of cn in t1 in row order, and only last matchingrecord is returnedThis is the reason why second record in t1 is matching with secondrecord in t2 as value of b(5 for t1 ) is more than corresponding valuein t2(4).On Nov 7, 5:25?am, Jack <bridgiema…> wrote:> I’m confused about how aj works for non-time columns.>> t1:(a:(1 1); b:(4 5); d:(7 8))> t2:(a:(2 1); b:(5 4); c:(7 9))>> aj[ab;t1;t2]>> a b d c> -------> 1 4 7 9> 1 5 8 9>> Why isn’t c null in the second row, there is no row in t2 with 1 and 5> for a and b.</bridgiema…>