merging data

[ishak celik]

Hello, I have two data sets. Each data set has two common variables but not the case (ID) number. How can I merge these two data sets based on two of these common variables.? (The original data set is separated into several packages, I would like to merge two packages again)

[Daniel Feenberg]

If x and y combined uniquely identify each case, and dataset2 is in core and sorted by x and y then

use dataset1
sort x y
merge 1:1 x y using dataset2

but if x and y are sex and race, obviously they won't uniquely identify each case and the merge will fail. See
https://www.stata.com/manuals/u23.pdf for the -merge- command.

[ishak celik]

I tried but got a message : "variables x y do not uniquely identify observations in the master data" in this case I used gender and class level of the students. But it did not help, what can be the reason for this? Thanks..

[Andrew Musau]

Use

merge m:1

or

merge 1:m

You should not have duplicates of both variables at least in one dataset. If you do, you need to think more about how you want to combine the datasets.

[ishak celik]

Ok, in two datasets I have two same variables. it was actually one dataset. But now dataset is divided into two. I am missing variables in the main data set. The ather dataset has variables that I need. between two datasets I have two common variables (var_x and var_y) . In this case how to merge these two datasets to one? So I can use the other variables that I need in the second dataset.. Make this sense? thank you...

[Andrew Musau]

As I said in #4, as long as there are no duplicates in the using dataset:

use dataset1
sort x y
merge m:1 x y using dataset2

Otherwise, see why you have duplicates:

use dataset2, clear
duplicates tag x y, gen(dup)
sort x y
list if dup, sepby(x y)