st: Slight problem with merge documentation / merge suggestions

[Fred Wolfe <fwolfe@arthritis-research.org>]

I have a file that contains duplicates in that 6 id variables are missing.

I merge a master file with the data set with missings

. merge patkey newenc using ${sql}demo1, nokeep keep(ulcermd) unique

There is no problem

I then do it but remove the -nokeep- option

. merge patkey newenc using ${sql}demo1,  keep(ulcermd) unique
variables patkey newenc do not uniquely identify observations in the using data
r(459);

What Stata is doing is not to see the duplicates in the using set if the 
-nokeep- option results in their not becoming a part of the merged data set.

This is very reasonable behavior, but one could not predict this in advance 
from the documentation which reads:

"unique specifies that the match variable(s) uniquely identify the 
observations in the master data and in the using data."

As this could cause some problems with users who might not understand the 
fine technicality, perhaps another line could be added to the help file to 
explain that duplicates excluded by the -nokeep- option do not trigger the 
errors message generated by the unique options.

merge suggestion

Stata's new merge program is a major improvement and fixed the most 
important problem addressed by Jeroen Weesie's -mmerge- series. Jeroen's 
program also automatically dropped _merge. I understand that Stata doesn't 
want to do that (though they drop xi _var), but would it be possible to add 
a simple option to -merge- to drop _merge & to ignore it if it exists in 
the using data set?

Jeroen's program provided a friendly warning when you merged non-unique 
observations. Stata's merge give the warning and generates an error if the 
-unique- options are set. But suppose the unique options are not set. It 
would still be very helpful to have the message "variables keyvar(s) do not 
uniquely identify observations in the using data" displayed even when the 
merge program does not generate an error message, either automatically or 
as a -message- option. Seems like a simple addition.

Fred Wolfe




---------------------------------------------------------------------------- 
------------------------
Fred Wolfe                                                          Tel 
(316) 263-2125
National Data Bank for Rheumatic Diseases         Fax (316) 263-0761
Wichita, 
Kansas                                    fwolfe@arthritis-research.org
---------------------------------------------------------------------------- 
-------------------------

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/