Collapse string variable to achieve a "count" within a postcode

[Kath Machray]

I have 1 string variable with 7913 observations. These are UK postcodes of either "AB23 8BB" or "AB1 1AA" format.

These have been generated from individual level data and I now need to get a count of people within each postcode. I have tried to use the collapse command but keep receiving a "type mismatch" error. I am using Stata 14.2.

Many thanks.

[Nick Cox]

No precise code to comment on here. FAQ Advice #12 applies. But by default collapse takes means, which may be your problem

At worst something like

gen one = 1 
collapse (count) one, by(postcode)

should work if you want a reduced dataset. Or see contract instead. Or consider

bysort postcode : gen freq = _N

for a variable alongside your existing dataset.

[Kath Machray]

Apologies for the lack of code.

Many thanks for that Nick - the best fit for my needs was:
gen one = 1 collapse (count) one, by(postcode)

[Adrien Bouguen]

Hi nick,

It seems strange to me that count does not count string variable. I don't really understand the rational behind it. I think this should be fixed.

Best