Re: st: replace part of observations in a string variable

[Nick Cox <njcoxstata@gmail.com>]

It's nice of Muhammad to quote a posting of mine, but he doesn't spell
out how it answers the question here.

First off, <redacted> writes of a variable -new2- but applies code to a
variable -addnew2-, so that confusion may be the answer.

Second, <redacted> is testing for _exact_ equality and if that fails one
possibility is that there are leading and trailing spaces so that

replace addnew2=" " if trim(addnew2) == "Pasir Mas"

or

replace addnew2=" " if strpos(addnew2, "Pasir Mas")

would be work-arounds in that case.

Either way, a broad strategic answer is "find out about string
functions", which was I suspect Muhammad's implicit point.

Nick

On Wed, Oct 17, 2012 at 11:08 AM, Muhammad Anees <anees@aneconomist.com> wrote:

> This can be for you.
>
> http://www.stata.com/statalist/archive/2011-06/msg00393.html

On Wed, Oct 17, 2012 at 3:02 PM, <redacted> wrote:

>> I want to replace part of observations in the string variable to
>> missing values. For example, I want to replace "Pasir Mas" to either
>> "." or " " from a string variable named as "new2"
>>
>> ------------------------------------
>> new2
>> ------------------------------------
>> Tumpat
>> Tumpat
>> Kota Bharu
>> Kota Bharu
>> Pasir Mas
>> Kota Bharu
>> Kota Bharu
>> --------------------------------------
>>
>> what I did was:
>>
>> . replace addnew2=" " if addnew2 == "Pasir Mas"
>> (0 real changes made)
>>
>> Observation "Pasir Mas" was not replaced by " " . What have I done
>> wrong? What is the better way of working with string variable?
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/