Append datasets horizontally without generating _n

[Miji Lee]

Hi, I wanted to append two datasets horizontally. However, I can't do merge them with _n properly because I think the dataset is very large: 180,453,364 observations are included in each dataset, and the numbers generated by _n seem erroneous when they exceeds a certain number.

Could someone tell me how I can append datasets without an identifier?

[Clyde Schechter]

I don't understand what you are trying to do. Where is this _n being generated? It is not done by -merge-. And you don't need any such variable to use -merge-. To -merge- two data sets you need a variable, or set of variables, that serves as a merge key that matches up observations in one data set with the observations in another.

If, however, for other reasons, you want to generate a variable that contains the observation number, and your data set is large, then you have to specify its storage type as long:

gen long obs_num = _n