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

RE: st: do group identifiers identify the same groups?

From   "Nick Cox" <>
To   <>
Subject   RE: st: do group identifiers identify the same groups?
Date   Thu, 10 Jul 2008 15:14:41 +0100

I'd tweak this so that you can see all observations in any group that is problematic. 
The logic is laid out in the FAQ I previously referred to in this thread. 

   bysort g1 (g2): generate byte same1 = g2[_N] == g2[1]
   bysort g2 (g1): generate byte same2 = g1[_N] == g1[1]
   tab1 same1 same2
   list if !same1 | !same2 


Svend Juul

I see. A combination of CÚlio's and my ideas comes here:

   bysort g1: generate same1 = (g2[_n]==g2[1])
   bysort g2: generate same2 = (g1[_n]==g1[1])
   tab1 same1 same2
   list if same1==0 | same2==0

Jeph Herrin 
Svend's idea will not work at all, as my two identifiers
are not meant to take the same values. That is, g1 will
likely never equal g2.
CÚlio's suggestion does get at what I want, especially as
it found no errors. However, I'm curious if there is a way
that would also let me identify which records or values
produced the errors;  -assert- is uninformative in that

*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index