Create a set of dummies and rename them

[Dario Maimone Ansaldo Patti]

Dear All,

consider the following dataset:

* Example generated by -dataex-. To install: ssc install dataex
clear
input float(invest market stock) str5 company byte time float id
317.6 3078.5   2.8 "Firm1" 1 1
391.8 4661.7  52.6 "Firm1" 2 1
410.6 5387.1 156.9 "Firm1" 3 1
257.7 2792.2 209.2 "Firm1" 4 1
330.8 4313.2 203.4 "Firm1" 5 1
40.29  417.5  10.5 "Firm2" 1 2
72.76  837.8  10.2 "Firm2" 2 2
66.26  883.9  34.7 "Firm2" 3 2
 51.6  437.9  51.8 "Firm2" 4 2
52.41  679.7  64.3 "Firm2" 5 2
 33.1 1170.6  97.8 "Firm3" 1 3
   45 2015.8 104.4 "Firm3" 2 3
 77.2 2803.3   118 "Firm3" 3 3
 44.6 2039.7 156.2 "Firm3" 4 3
 48.1 2256.2 172.6 "Firm3" 5 3
12.93  191.5   1.8 "Firm4" 1 4
 25.9    516    .8 "Firm4" 2 4
35.05    729   7.4 "Firm4" 3 4
22.89  560.4  18.1 "Firm4" 4 4
18.84  519.9  23.5 "Firm4" 5 4
209.9 1362.4  53.8 "Firm5" 1 5
355.3 1807.1  50.5 "Firm5" 2 5
469.9 2676.3 118.1 "Firm5" 3 5
262.3 1801.9 260.2 "Firm5" 4 5
230.4 1957.3 312.7 "Firm5" 5 5
end

I would like to generate a set of dummies, one for each group under the variable id, Hence I type:

tab id, gen(c)

Then, I would like to rename each dummy with the corresponding name under the variable company. So dummy c1 should be Firm1, dummy c2 should be Firm2 and so on.

I have seen the following post:

https://www.statalist.org/forums/for...s-using-a-loop

The point is that my original dataset contains more than 150 groups and it is quite long to write down 150 names if I would like to use something like:

rename (c1-c5) (Firm1 Firm2 ....)

I would like to know if there is a quicker way to do so. Myabe using a loop? I have thought to use levelsof but I cannot figure out how to set a loop to assign the correct name to each dummy.

Thanks in advace for your kind help.

[Andrew Musau]

You need to ensure that the strings are valid Stata names. You can do something like:

gen id2= strtoname(company)
levelsof id2, local(ids)
foreach id in `ids'{
    gen `id'= id2=="`id'"
}

[Dario Maimone Ansaldo Patti]

Andrew Musau Thanks a lot. It worked fine.