Combine timetables vertically different variables

I have two timetables tt1 and tt2 each with different dates and I want to concatenate them vertically using tt12 = [tt1; tt2];

However, the issue is that tt1 and tt2 do not have all columns (variables) in common. I would need the variables that are present in both timetables to be combined (concatenated) vertically as usual, but also append any new variables to the final tt12, where missing dates in those new variables would be just NaN.

I cannot use synchronize as that creates tt12 with double number of variables (it does not merge the same variables into one column).

The new timetable tt12 would then have more columns (the unique union of all variables present in tt1 and tt2). The only thing I could think of is to manually prepare (expand) the timetables tt1 & tt2 so that they have the same variables (using ismember function on variable names, perhaps), before using the tt12 = [tt1; tt2];

Is there any easier function that would achieve this?

Answer 1

[Duncan Po]

You can use outerjoin with MergeKeys option set to true. You need to specify your common variable names (in addition to 'Time') as the keys. See this example:

>> t1 = timetable(hours(1:5)', (1:5)', (6:10)', 'VariableNames', {'foo', 'bar'});

>> t2 = timetable(hours(6:10)', (1:5)', (6:10)', 'VariableNames', {'foo2', 'bar'});

>> t12 = outerjoin(t1, t2, 'Keys', [{'Time'}, intersect(t1.Properties.VariableNames, t2.Properties.VariableNames)], 'MergeKeys', true)

t12 =

10×3 timetable

Time foo bar foo2

_____ ___ ___ ____

1 hr 1 6 NaN

2 hr 2 7 NaN

3 hr 3 8 NaN

4 hr 4 9 NaN

5 hr 5 10 NaN

6 hr NaN 6 1

7 hr NaN 7 2

8 hr NaN 8 3

9 hr NaN 9 4

10 hr NaN 10 5

[Martin Kabelka]

Thank you Duncan, this is exactly what I was looking for!