How can I expand a table row variable array into multiple table rows

I have a table, something like this trivial example:

s(1).id="a"; s(1).datatype="this";s(1).val=zeros(34,1);
s(2).id="b"; s(2).datatype="that";s(2).val=zeros(12,1);
s(3).id="c"; s(3).datatype="other";s(3).val=zeros(42,1);
T=struct2table(s)

3×3 table

    id          val         datatype
    ___    _____________    ________

    "a"    [34×1 double]    "this"  
    "b"    [12×1 double]    "that"  
    "c"    [42×1 double]    "other"

I would like to expand the table such that the first row would expand to 34 rows, each with a successive value from the val table variable for the original row. I would end up with a table with 34+12+42 rows.

I've tried doing this with a split apply on the table, grouping into the datatype variable, but ultimately I'm iterating through the table and it's slow, on the order of 20 minutes for a table of 70K rows. It seemed that rowfun() would also be slow because it's also iterating through the table. I was wondering if there is some faster way to perform this operation.

Ultimately what I'm trying to do is treat the val variable in my trivial example as a date time value and then generate a timetable using the date times as the row times, in case that makes any difference.

Answer 1

[Peter Perkins]

Time to go old school:

>> s(1).id = "a"; s(1).datatype = "this"; s(1).val = (1:34)';
>> s(2).id = "b"; s(2).datatype = "that"; s(2).val = (34+(1:12))';
>> s(3).id = "c"; s(3).datatype = "other"; s(3).val = (34+12+(1:42))';
>> T = struct2table(s)
T =
  3×3 table
    id     datatype         val     
    ___    ________    _____________
    "a"    "this"      [34×1 double]
    "b"    "that"      [12×1 double]
    "c"    "other"     [42×1 double]
>> T2 = T(repelem(1:height(T),nvals),1:2);
>> T2.val = vertcat(T.val{:})
T2 =
  88×3 table
    id     datatype    val
    ___    ________    ___
    "a"     "this"      1 
    "a"     "this"      2 
    "a"     "this"      3 
    "a"     "this"      4 
    "a"     "this"      5 
    "a"     "this"      6 
    "a"     "this"      7 
    "a"     "this"      8 
  [snip]
    "c"    "other"     81 
    "c"    "other"     82 
    "c"    "other"     83 
    "c"    "other"     84 
    "c"    "other"     85 
    "c"    "other"     86 
    "c"    "other"     87 
    "c"    "other"     88

[Ron Cordell]

I'm not sure where nvals comes from. If set nvals=88 then height(T2) becomes 264...

[Peter Perkins]

oops:

nvals = cellfun('length',T.val)

[Elizabeth Ramsey]

Any tips on doing something similar to a table with multple "val" type columns? All tables within a row are the same size. I feel like there should be something simple but I'm at a loss

[Peter Perkins]

Varfun is the usual way to apply some operation across variables in a table. But of course there's no guarantee that the different vars will expand out the same.