Separating numeric and character parts of a string variable

[Chris Rooney]

Hello all,

I have a variable, s4mainocc, which is constructed in the following way (I provide a number of examples below). The s4mainocc variable is a string variable.

2412. Personnel and careers professionals
3433. Bookeepers
4115. Secretaries

Now I'd like to separate the occupational code (the first four digits - note ALL occupations are four digits) from the label into two separate variables. This example https://www.statalist.org/forums/forum/general-stata-discussion/general/1612510-split-a-string-variable-into-character-and-numeric-parts seemed to be exactly what I was looking for, so I tried the following:

gen numeric = substr(s4mainocc, 4, 4)

however I received the following error code:

type mismatch
r(109);

[Nick Cox]

numeric is a reserved word.

help [M-2] reswords

The error message is confusing, but I believe that to be the error if s4mainocc is a string variable. You need a different name.

[Chen Samulsion]

Firstly, please use -dataex- to provide data example, as there seems to be some storage type problem of your targeted vairable.

dataex s4mainocc

Secondly, if you want to extract the four digits of occupation code and the occupation details, you should modify you command like this:

gen numeric=substr(s4mainocc, 1, 4)
gen occupation=substr(s4mainocc, 6, .)

crossed: I can generate variable numeric in Stata 16 (why?)

[Chris Rooney]

Hi Chen Samulsion

Here is the dataex example - looks a bit weird to me. I should probably mention I used "numlabel, add" before doing this. Ah, I was looking at the variable above by mistake. I see this is an integer variable! Sorry, my mistake!

* Example generated by -dataex-. For more info, type help dataex
clear
input int s4mainocc
9211
7122
9131
9211
8334
9322
9211
9141
   .
5133
1314
   .
   .
9132
8290
8212
7136
9131
5220
   .
   .
   .
1315
9322
   .
9211
9211
3433
9211
4211
   .
4131
7136
   .
   .
9131
1311
   .
1311
   .
6112
   .
9211
9211
9211
9211
1227
   .
   .
7137
7213
   .
7245
3152
   .
   .
4211
1314
1314
   .
   .
   .
   .
9322
9322
   .
5169
   .
   .
9322
   .
   .
9333
   .
   .
7122
   .
   .
9322
8152
   .
   .
   .
9322
9322
9322
   .
9131
9211
   .
   .
9322
9322
   .
9322
   .
9322
   .
   .
   .
end
label values s4mainocc Q42OCCUPATION
label def Q42OCCUPATION 1227 "1227. Production and operations managers/department managers in business services", modify
label def Q42OCCUPATION 1311 "1311. General managers in agriculture, hunting, forestry and fishing", modify
label def Q42OCCUPATION 1314 "1314. General managers in wholesale and retail trade", modify
label def Q42OCCUPATION 1315 "1315. General managers of hotels, restaurants and other catering or accommodation services", modify
label def Q42OCCUPATION 3152 "3152. Safety, health and quality inspectors, Inspectors, safety and health", modify
label def Q42OCCUPATION 3433 "3433. Bookkeepers", modify
label def Q42OCCUPATION 4131 "4131. Stock clerks", modify
label def Q42OCCUPATION 4211 "4211. Cashiers and ticket clerks", modify
label def Q42OCCUPATION 5133 "5133. Home-based personal care workers", modify
label def Q42OCCUPATION 5169 "5169. Protective services workers not elsewhere classified, Rangers and game wardens", modify
label def Q42OCCUPATION 5220 "5220. Shop salespersons and demonstrators, Salespersons, Petrol pump and filling station attendants", modify
label def Q42OCCUPATION 6112 "6112. Tree and shrub crop growers (farm owners and skilled farm workers)", modify
label def Q42OCCUPATION 7122 "7122. Bricklayers and stonemasons (including apprentices/trainees)", modify
label def Q42OCCUPATION 7136 "7136. Plumbers and pipe fitters (including apprentices/trainees)", modify
label def Q42OCCUPATION 7137 "7137. Building and related electricians (including apprentices/trainees)", modify
label def Q42OCCUPATION 7213 "7213. Sheet-metal workers (including apprentices/trainees)", modify
label def Q42OCCUPATION 7245 "7245. Electrical line installers, repairers and cable jointers (including apprentices/trainees)", modify
label def Q42OCCUPATION 8152 "8152. Chemical heat-treating plant operators", modify
label def Q42OCCUPATION 8212 "8212. Cement and other mineral products machine operators", modify
label def Q42OCCUPATION 8290 "8290. Other machine operators and assemblers not elsewhere classified", modify
label def Q42OCCUPATION 8334 "8334. Lifting-truck operators", modify
label def Q42OCCUPATION 9131 "9131. Domestic helpers and cleaners", modify
label def Q42OCCUPATION 9132 "9132. Helpers and cleaners in offices, hotels and other establishments", modify
label def Q42OCCUPATION 9141 "9141. Building caretakers", modify
label def Q42OCCUPATION 9211 "9211. Farmhands and labourers", modify
label def Q42OCCUPATION 9322 "9322. Hand-packers and other manufacturing labourers", modify
label def Q42OCCUPATION 9333 "9333. Freight handlers", modify

[Chen Samulsion]

Aha, there it is. The variable s4mainocc is an integer, you cannot manipulate it using -substr()- which suits to string variable. And if you still want to get what you want in #1, that is to say, generate two variables, one stores the occupation code, and one stores the occupation titles or details, I recommend the following codes:

decode s4mainocc, gen(occupation)
generate occ_code=substr(occupation, 1, 4)
generate occ_titles=substr(occupation, 6, .)

help decode