Automate conditional variable creation

[Sebastien Dubois]

Hello,

I would like to automate the creation of categorical variables in Stata. My data look like this

nom	poste2011	rang2011	poste2012	rang2012
Durand	assistant	R1	assistant	R3
Lepetit	associé	R2	associé	R2
Legrand	assistant	R3	assistant	R2
Gros	assistant	R1	assistant	R3
Maigre	associé	R2	associé	R2
Dupuis	assistant	R3	assistant	R2

I would like to assign to each name (Durand, Lepetit etc.) a state following to two conditions (poste + rang) for each year in a new column.
For instance assistant + R1 = state A ; assistant + R2 = state B ; associé + R1 = state C etc.
Can I do this in Stata?
Thanks for your help!

[Andrew Musau]

Your data is in wide layout. See

help reshape

to get it into long layout. I think you want

egen wanted= group(poste rang), label
tab wanted

[Sebastien Dubois]

Thanks much for your answer. I thought of reshaping the data, but I have the issue that it would mean have a row by name and year? I have no "year" variable, can I extract the "year" variable from the "poste2011" (etc.) variable? Thanks again

[Andrew Musau]

For your future posts, review FAQ Advice #12 with a focus on how to use the dataex command to provide data examples.

[CODE]

* Example generated by -dataex-. To install: ssc install dataex
clear
input str7 nom str9 poste2011 str2 rang2011 str9 poste2012 str2 rang2012
"Durand"  "assistant" "R1" "assistant" "R3"
"Lepetit" "associé"  "R2" "associé"  "R2"
"Legrand" "assistant" "R3" "assistant" "R2"
"Gros"    "assistant" "R1" "assistant" "R3"
"Maigre"  "associé"  "R2" "associé"  "R2"
"Dupuis"  "assistant" "R3" "assistant" "R2"
end

reshape long poste rang, i(nom) j(year)
egen wanted= group(poste rang), label
tab wanted

Res.:

. tab wanted

 group(poste |
       rang) |      Freq.     Percent        Cum.
-------------+-----------------------------------
assistant R1 |          2       16.67       16.67
assistant R2 |          2       16.67       33.33
assistant R3 |          4       33.33       66.67
  associé R2 |          4       33.33      100.00
-------------+-----------------------------------
       Total |         12      100.00

. l, sepby(nom)

     +--------------------------------------------------+
     |     nom   year       poste   rang         wanted |
     |--------------------------------------------------|
  1. |  Dupuis   2011   assistant     R3   assistant R3 |
  2. |  Dupuis   2012   assistant     R2   assistant R2 |
     |--------------------------------------------------|
  3. |  Durand   2011   assistant     R1   assistant R1 |
  4. |  Durand   2012   assistant     R3   assistant R3 |
     |--------------------------------------------------|
  5. |    Gros   2011   assistant     R1   assistant R1 |
  6. |    Gros   2012   assistant     R3   assistant R3 |
     |--------------------------------------------------|
  7. | Legrand   2011   assistant     R3   assistant R3 |
  8. | Legrand   2012   assistant     R2   assistant R2 |
     |--------------------------------------------------|
  9. | Lepetit   2011     associé     R2     associé R2 |
 10. | Lepetit   2012     associé     R2     associé R2 |
     |--------------------------------------------------|
 11. |  Maigre   2011     associé     R2     associé R2 |
 12. |  Maigre   2012     associé     R2     associé R2 |
     +--------------------------------------------------+

.

[Sebastien Dubois]

Hello, many thanks. I did not know dataex, I'll use it next time. A last question: how can I manipulate the label command so that the labels in the new variable look like "etat1"; "etat2", "etat3" etc. Thanks again

[Andrew Musau]

Drop the label option in the egen command in #4 and see

help label

on defining your own labels.