Statalist The Stata Listserver

[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: RE: Count presence of item on different vars

From   "Nikolaos A. Patsopoulos" <>
Subject   Re: st: RE: Count presence of item on different vars
Date   Mon, 09 Oct 2006 18:53:03 +0300

Nick Cox wrote:
Well, now you tell us. Possibly some variant on this:
quietly { foreach v of var v6-v146 { levelsof `v', local(this) local all : list all | this }

local all : list sort all
foreach text of local all { local count = 0 foreach v of var var6-v146 { count if `v' == "`text'" local count = `count' + r(N)
noi di as txt "`text' `count'" }
Nikolaos A. Patsopoulos

Nick Cox wrote:

Check out -tabm- from -tab_chi- on SSC.
Alternatively, an approach from first
principles is to -reshape long- so that this reduces to a two-way tabulation.
Nikolaos A. Patsopoulos

I've got a set of variables v6-v146, each one containing string values (unique values ~4500):

v6 v7 v8 v9 v10...
text1 text1 text3 text45 text1...
text5 text2 text1 text455 text1

I want to count how many times each of the values appears
in at least
one of the v6-v146 in each row [observation], e.g. text1
appears at
least once in 7365obs.

How can achieve this?

I tried :
tabm v7-v146

I had to kill the process after few minutes (it's a 150MB database)

Then used:

tabm v7-v8 // just to see what will come up but that's what I got:
too many values

The unique values for each var is 4364 from a total of 25725 observations
* For searches and help try:

Nick I believe that you answered my problem, however I cannot test it since " levelsof `v', local(this) " exceeds 800 items (if i get it right).


* For searches and help try:

© Copyright 1996–2021 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index