Keeping observations that start with specific digits

[Alyse Polly]

Hello,

How can I keep observations that start with a certain set of numbers. Specifically, I want to keep observations that start with "3344" within a list of mostly 6-digit numbers. I found a previous post (http://www.stata.com/statalist/archi.../msg01191.html) but for some reason these options aren't working. I've tried the following but stata says I’m using invalid syntax.

keep if substr(string(naics02), 1, 2, 3, 4) == "3344"

I then tried using floor but it dropped all of my data (when in fact there are observations that start with 3344):

keep if floor(naics02/10000) == 3344

Please help! Thank you!

Alyse

[Nick Cox]

The first is a syntax problem. substr() has three arguments, the string expression, the start of the substring and the length of the substring. Hence what you need is something more like

keep if substr(string(naics02), 1, 4) == "3344"

See

help substr() 

The second is an arithmetic problem. If you want numbers like 3344?? to be truncated to 3344 you need a different divisor:

keep if floor(naics02/100) == 3344

as you want to lop off the last two digits. Check

 
di floor(334499/100) 
3344

but if some numbers are not 6 digits this may not deliver what you want in such cases. Hence the first solution is likely to be better.

[Alyse Polly]

Excellent - the following code worked. Thanks!
keep if substr(string(naics02), 1, 4) == "3344"