User-Agent: Workspace Webmail 5.16.0Message-Id: <20160101141535.85f80dae80d1d2f2e266ec6278e6cbe8.09ddc238fa.wbe@email07.europe.secureserver.net>From: “David Demner (AquaQ)” <david.demner>To: personal-kdbplus@googlegroups.comSubject: RE: [personal kdb+] optimize ungroup columns operation in kdbDate: Fri, 01 Jan 2016 14:15:35 -0700Mime-Version: 1.0
You don’t need ungroup in this case. One approach is to create sub-tables, rename the col, and raze
q)n:1300000
q)a:([]cust:n?`8;v1:n?100;v2:n?100;v3:n?100)
q)\t t1:ungroup?[a;();0b;`cust`v!(`cust;(flip; enlist,`v1`v2`v3))]
1068
q)\t t2:raze{`cust`v xcol (`cust,y)#x}[a;] each cols[a] except `cust
13
q)asc[t1]~asc[t2]
1b
The order is different so it depends on whether that's ok or not.