Count frequency of observation within a variable

[Xueqian Chen]

Hi,

I would like to count the frequency of an observation within a variable.
Thus, I would like to create a variable "a" which refers to variable"x" in the following pattern:

variable "x" 1 1 1 2 2 2 2 2 6 7 7 7 8 8
variable "a" 3 3 3 5 5 5 5 5 1 3 3 3 2 2

Basically "a" counts how many times "x" is 1, how many times "x" equals 2 etc.

Does someone know how to do that?
I was unable to find anything in that direction..

Thanks,
Xueqian

[Dave Airey]

See -contract-.

[daniel klein]

Say your original dataset is called mydata, then

tempfile mycopy
save `mycopy'
contract x , freq(a)
mer 1:m x using `mycopy'

will do what you want.

Best
Daniel

[Robert Picard]

All you need is to group observations by x and use _N (see help _variables) as the count. Also see [U] 11.5 by varlist: construct and

SJ-2-1 pr0004 . . . . . . . . . . Speaking Stata: How to move step by: step
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . N. J. Cox
Q1/02 SJ 2(1):86--102 (no commands)
explains the use of the by varlist : construct to tackle
a variety of problems with group structure, ranging from
simple calculations for each of several groups to more
advanced manipulations that use the built-in _n and _N

* Example generated by -dataex-. To install: ssc install dataex
clear
input byte x
1
1
1
2
2
2
2
2
6
7
7
7
8
8
end

* the number of observation in a group is _N
bysort x: gen a = _N

list, sepby(x)

[Xueqian Chen]

Thanks to all of you! The contract command followed by the merge command works perfectly.
I was not really able to use Robert's example as I have over 15.000 observations so that the code would be very long and unwieldy.

[Robert Picard]

It's just a single line, please read the references I just added. All you need is

bysort x: gen a = _N

[Xueqian Chen]

Oh, I got it now. I was thinking that I would need the upper part of your code as well, but it also works with a variable which is already in the dataset. Sorry for the confusion!