Drop observations in a string variable

[nawar han]

Hello Everyone, i'm a new user of Stata and having trouble with a really simple command. I can't seem to delete observation with a string variable. Can someone help me.
Thanks


AJCC |
Clinical N | Freq. Percent Cum.
------------+-----------------------------------
0 | 478,848 64.58 64.58
1 | 1,625 0.22 64.80
X | 260,978 35.20 99.99
88 | 61 0.01 100.00
------------+-----------------------------------
Total | 741,512 100.00

. drop if (TNM_CLIN_N =="1")
(0 observations deleted)

[ben earnhart]

Looks like you probably have a numeric variable with labels. try:

tab TNM_CLIN_N, nol

and see.

[Nick Cox]

It's not obvious that TNM_CLIN_N has any relation at all to a variable with label "AJCC Clinical N".

TNM_CLIN_N is evidently a string variable, but it can not be what you are tabulating as otherwise 1625 observations would have been dropped.

In any case, there is a radical confusion of categories here. Observations can't have string variables. Perhaps what you mean is that you want to drop observations for which there is either

(a) entirely non-numeric characters in a string variable

or

(b) a value label with non-numeric characters for a numeric variable.

We can't tell. You must tell us more.

[nawar han]

Thanks for reply
The variable name is TNM_CLIN_N and it's labeled AJCC Clinical N.
The type is str4 and Format is %4s

When I tab the variable the categories are 0 - 1 - X - 88

I want to drop all the observation that have either "1" or "88"

is this clearer?

thanks again

[Nick Cox]

If what you say is true then

 
drop if (TNM_CLIN_N =="1")

should not do what you report, namely drop nothing.

There may be leading or trailing spaces. See what happens with

replace TNM_CLIN_N = trim(TNM_CLIN_N)

[Robert Picard]

One explanation is that your values include spaces. For example

clear
input str4 TNM_CLIN_N
" 0"
" 1"
x
88
end
label var TNM_CLIN_N "AJCC Clinical N"
tab TNM_CLIN_N
drop if TNM_CLIN_N == "1"
drop if trim(TNM_CLIN_N) == "1"

[nawar han]

It worked!

thank you very much guys