*(Q) The data are in a "stacked" format according to time point like the following: ID TIME VAR1 VAR2 VAR3 1 1 A D G 1 2 B E H 1 3 C F I [...] * I'd like to re-arrange the data so that there is a single row for each id, creating columns for each variable (VAR1-VAR3) at each time point: * ID VAR11 VAR12 VAR13 VAR21 VAR22 VAR23 VAR31 VAR32 VAR33 1 A B C D E F G H I 2 ... * One thing to note is that there are more than 3 time points and more than 3 variables. * Any and all suggestions appreciated. * (A) Posted to spss newsgroup by rlevesque@videotron.ca on 2001/10/10. * SPSS web site ----> http://pages.infinit.net/rlevesqu/index.htm * Note naming the original variables var1, var2 and var3 does not work because if time =12 (say) then there wil confusion between var1 at time 11 which is var111 and var11 at time 1 which is var111. DATA LIST LIST /ID(F8.0) time1(F8.0) var1t(A8) var2t(A8) var3t(A8). BEGIN DATA 1 1 A E I 1 2 B F J 1 3 C G K 1 4 D H L 2 1 AA EE II 2 2 BB FF JJ 2 3 CC GG KK 2 4 DD HH LL END DATA. LIST. SET MPRINT=no. *//////////////////////////////////. DEFINE !transf(nbvars=!TOKENS(1) /nbtime=!TOKENS(1)) !DO !cnt=1 !TO !nbvars. STRING !CONCAT('var',!cnt,'t1') TO !CONCAT('var',!cnt,'t',!nbtime)(A8). VECTOR !CONCAT('var',!cnt)=!CONCAT('var',!cnt,'t1') TO !CONCAT('var',!cnt,'t',!nbtime). !DOEND VECTOR v=var1t1 TO !CONCAT('var',!nbvars,'t',!nbtime). !DO !cnt=1 !TO !nbvars. COMPUTE !CONCAT(v,'((',!cnt,'-1)*',!nbtime,'+time1)')=!CONCAT(var,!cnt,'t'). !DOEND. AGGREGATE /OUTFILE=* /BREAK=id !DO !cnt=1 !TO !nbvars /!CONCAT('var',!cnt,'t1') TO !CONCAT('var',!cnt,'t',!nbtime)=MAX(!CONCAT('var',!cnt,'t1') TO !CONCAT('var',!cnt,'t',!nbtime)) !DOEND. EXECUTE. !ENDDEFINE. *//////////////////////////////////. SET MPRINT=yes. * Call macro giving the number of variables and of time periods. * Note: these 2 varables could be determined by the syntax if required. !transf nbvars=3 nbtime=4.