A colleague sent me two Stata data sets created in Stata 13 with Spanish variable labels in ASCII and asked me to merge the two files and save them as Stata 13 files. I have Stata 14. I opened one file and used -merge 1:1 etc.- to bring it together with a file on disk, then used -saveold etc., version(13)-, renaming the new merged file.
State returned the following:
(saving in Stata 13 format)
note: variable label "N�mero de Integrante" contains unicode and thus may not display well in Stata
13.
note: variable label "�Cu�l es el sexo del(de la) entrevistado(a)?" contains unicode and thus may not
display well in Stata 13.
note: variable label "�Cu�ntos a�os cumplidos tiene actualmente? " contains unicode and thus may not
display well in Stata 13.
..............................................
and much more in the same vein.
My data source reports that the ASCII labels with accents were legible to her when she sent them to me, though I read them in Stata 14 (as I expected) as shown above, and that they were legible (i.e. in ASCII) when she opened the merged file I sent back to her. I performed no translations of ASCII to Unicode or viceversa in the files.
Are the erroneous warnings I received after using -saveold- a bug, since the ASCII labels appeared to be identified as unicode?
Steve Rothenberg
National Institute of Public Health
Cuernavaca, Mexico
State returned the following:
(saving in Stata 13 format)
note: variable label "N�mero de Integrante" contains unicode and thus may not display well in Stata
13.
note: variable label "�Cu�l es el sexo del(de la) entrevistado(a)?" contains unicode and thus may not
display well in Stata 13.
note: variable label "�Cu�ntos a�os cumplidos tiene actualmente? " contains unicode and thus may not
display well in Stata 13.
..............................................
and much more in the same vein.
My data source reports that the ASCII labels with accents were legible to her when she sent them to me, though I read them in Stata 14 (as I expected) as shown above, and that they were legible (i.e. in ASCII) when she opened the merged file I sent back to her. I performed no translations of ASCII to Unicode or viceversa in the files.
Are the erroneous warnings I received after using -saveold- a bug, since the ASCII labels appeared to be identified as unicode?
Steve Rothenberg
National Institute of Public Health
Cuernavaca, Mexico