How to combine table vertically

I have two large tables I want to combine. They both have exactly the same column headings but different row numbers (>300) and the headings may be numbers, categoricals or text. I just want to combine them vertically so that the final table has just one column heading with all the data underneath.

[KALYAN ACHARJYA]

outerjoin(table1,table2,'MergeKeys', true);

Ans Link

[Wendy Cameron]

Thank you - this answer worked. I had been trying to use vertcat which works for an array and although one could convert the table to an arry as below, this solution here seems to work nicely in my situation. Thank you all very much or your ongoing assistance with Matlab.

Kind regards, Wendy

Answer 1

[madhan ravi]

table2array()

array2table()

https://www.mathworks.com/help/matlab/matlab_prog/create-a-table.html#d120e26567

Answer 2

[Peter Perkins]

Wendy, I'm not sure you are getting the correct advice. You have said you want to vertically concatenate. outerjoin does not do that. It's a completely different operation, closer to horizontal concatenation.

If you do want to vertically concatenate, you do that the same way with tables as with anything else in MATLAB: [t1; t2]. No conversion needed. However, the variable names in the two tables do need to be the same. You have said, "They both have exactly the same column headings", so that should not be a problem. You also said, "the headings may be numbers, categoricals or text.", which doesn't make sense, so I think maybe you meant to say, "the variables in the table may be ...".

So, if you want to vertically concatenate, you shouldn't use outerjoin, and you don't need to convert back and forth with numeric.

[Wendy Cameron]

Thanks Peter.

Yes you are right, the variables in the different columns are either numbers or text or categoricals. However, it seems that using outerjoin I got the result I was after where my two tables with 312 and 412 rows respectively and both with 42 columns were put into one table of 724 rows and 42 columns. It did seem to look correct. Maybe I am misunderstanding the word concatenate?

Kind regards, Wendy

[Peter Perkins]

I almost take back part of what I said about outerjoin. outerjoin happens to work in this case, because

1) you have identical variable names in both tables, and the default behavior is to use common vars as keys,

2) there's the flag to merge the keys in the output, and

3) apparently the two tables have no rows in common, so the outer join doesn't actually join anything, it treats every row as unmatched.

But this is pretty obscure. It will not work correctly as soon as the two tables have a row in common, and it isn't very obvious. A join is fundamentally a horzcat-like operation. What went wrong with vertcat?

On the other hand, if what you really want is, "vertcat the two tables but only keep one copy of common rows", then perhaps outerjoin is what you want. I would argue that vertcat+unique is more understandable, but that's up to you.

Hope this helps.

[Wendy Cameron]

Great, thanks for the feedback Peter. It seems it worked just out of luck for my situation. I checked a few columns to be sure. But your method works too and is of course simpler and I have noted that down to use next time. Thanks once again for helping me to better understand matlab.