Grouping variables with similar ending

[Hasan Isomitdinov]

I have a bunch of variables generated that are the means of other variables I have in a dataset. All the generated variables end with an ending "_z". I would like to be able to create a global variable which contains all of those variables so that I don't have to keep typing them. Also, when I am using the esttab function, I want to suppress those generated variables.
Any help would be appreciated.

[Hemanshu Kumar]

First, terminology: you are looking to create a macro, not a global variable. A local macro will likely suffice for most purposes, though you could also use a global macro. Many here would advise against the latter. You might want to see

help macro

Note that for your use, a macro may not be needed at all; you may be able to just use wildcards wherever variable lists are accepted. For instance

sum *_z

You might want to check

help varlist

If your variable list is more complicated than that, you might use a macro (here, a local macro):

local myvars var1 var2 *_z

and then use that macro myvars in a command:

sum `myvars'

[Nick Cox]

Pedantry is precision that the pedant prefers while the readership may not.

More pedantry, but the principle of using Stata terminology to discuss Stata is surely a good one.

I

I would like to be able to create a global variable which contains all of those variables

Following Hemanshu Kumar and taking one step further.

I would like to create a local macro which contains all of those variable names

when I am using the esttab function

esttab is a command, not a function. Unlike some other software, there is a sharp division in Stata between commands and functions.


If you want to suppress a group of variables, you need the names of the complementary set, and

ds *_z, not

shows some technique. See also findname from the Stata Journal, functionally a superset of ds.

To manipulate lists of variable names, check out e.g.

help macrolists

for how to get the difference between sets, and so on.