The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Data » Merging data files » v001 v002 v003 doesnot uniquely identify observation
v001 v002 v003 doesnot uniquely identify observation [message #3072] Sun, 12 October 2014 23:32 Go to next message
kalfikirsisay@gmail.com is currently offline  kalfikirsisay@gmail.com
Messages: 11
Registered: October 2014
Location: NL
Member
Hi All,
I am using Ethiopian DHS2011 for my analysis. I need to merge birth recode file (using file) with HH member recode(master file). I tried to see the DHS merging guide and related topics in the FAQ. I was aware of the fact that HIDX is the household line number and used the following STATA command to merge,
gen gen v001= hv001
gen v002 = hv002
gen v003= hvidx
sort v001 v002 v003
merge m:1 v001 v002 v003 using "the birth recode file"
neverthless, I could'nt succed in merging and STAT propmted me the messgae " v001 v002 v003 doesnot uniquely identify the observation in using file". would you please quickly tip me on how to merge the two recodes?
Thanks in advance,
Kal
Re: v001 v002 v003 doesnot uniquely identify observation [message #3075 is a reply to message #3072] Mon, 13 October 2014 08:47 Go to previous messageGo to next message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 3017
Registered: February 2013
Senior Member
Dear User,

Please reference: post # 3011

Thanks - Bridgette-DHS
Re: v001 v002 v003 doesnot uniquely identify observation [message #3077 is a reply to message #3075] Mon, 13 October 2014 09:31 Go to previous messageGo to next message
kalfikirsisay@gmail.com is currently offline  kalfikirsisay@gmail.com
Messages: 11
Registered: October 2014
Location: NL
Member
Dear Bridgette,
thanks for your help, but still I tried both methods(gen long v001 = hv001,gen v002 = hv002,gen v003 = hvidx)and the rename approach, both turned out not to work and STATA gave me similar results as before"variables v001 v002 v003 do not uniquely identify observations in the using data". What is wrong with my case? your quick elaboration is highly appreciated,
Regards,
Kalkidan Sisay
Re: v001 v002 v003 doesnot uniquely identify observation [message #3080 is a reply to message #3077] Mon, 13 October 2014 10:22 Go to previous messageGo to next message
Trevor-DHS is currently offline  Trevor-DHS
Messages: 787
Registered: January 2013
Senior Member
You should be using a 1:1 merge, not many to one - try that. If not, send your exact code you are using.
Re: v001 v002 v003 doesnot uniquely identify observation [message #3081 is a reply to message #3080] Mon, 13 October 2014 11:43 Go to previous messageGo to next message
kalfikirsisay@gmail.com is currently offline  kalfikirsisay@gmail.com
Messages: 11
Registered: October 2014
Location: NL
Member
dear Trevor,
thanks for quick response
I tried using 1:1 merge command only to get simialr message from STATA as before. I couldnot figure out why this is so, the following was the exact code I used.
use "C:\Users\mademsei\Desktop\CSA data and reports\Biomaker manual\merging files for bio\HH member recode\ETPR61FL.DTA", clear
gen v001 = hv001
gen v002 = hv002
gen v003= hvidx
sort v002 v002 v003
merge 1:1 v001 v002 v003 using "C:\Users\mademsei\Desktop\birth 1 to 1 merge.dta"
variables v001 v002 v003 do not uniquely identify observations in the using data

The master file I am using is HH member and the using file is birthrecode. Neverthless, I used the 1:m merge and get matched data albeit many observations(over 66000) were not matched. This is the table I obtained from 1:m merge command.
merge 1:m v001 v002 v003 using "C:\Users\mademsei\Desktop\birth 1 to 1 merge.dta"

Result # of obs.
-----------------------------------------
not matched 66,848
from master 66,848 (_merge==1)
from using 0 (_merge==2)

matched 45,540 (_merge==3)
-----------------------------------------

Is this the right way? I am a bit sceptic of the merging approach, I thought either m:1 or 1:1 to be proper, ofcourse I merged all observations in my using file and this what I need to appear in the maser file. your last comment on this is of prior value,Thanks in advance,
Kalkidan Sisay

[Updated on: Mon, 13 October 2014 11:58]

Report message to a moderator

Re: v001 v002 v003 doesnot uniquely identify observation [message #3082 is a reply to message #3072] Mon, 13 October 2014 12:07 Go to previous messageGo to next message
Trevor-DHS is currently offline  Trevor-DHS
Messages: 787
Registered: January 2013
Senior Member
Your merge doesn't work as you have it because you are trying to match the births recode (BR) file to the PR file. For each respondent there are many births, so consequently the 1:1 merge doesn't work. If you tried merging using the IR file it would work. What are you trying to achieve? If you can explain what you want your final dataset to look like then I can tell you the best way of achieving that.
Re: v001 v002 v003 doesnot uniquely identify observation [message #3083 is a reply to message #3082] Mon, 13 October 2014 12:41 Go to previous messageGo to next message
kalfikirsisay@gmail.com is currently offline  kalfikirsisay@gmail.com
Messages: 11
Registered: October 2014
Location: NL
Member
Dear Trevor,
Many thanks for your extended helps,
My objective is to have a data set containing HIV test result, anemia test result ,other biomarkers(nutrition and vaccine) and other diseasses like diarrhoea prevalence and treatements (I also needed malaria but thanks to DHS experts, one expert told me the Eth. DHS has no data on it).I choose birth recode, hiv recode and HH memeber recode files as candisate files to merge them so that I can have one file contining the aforementioned inforamtion for my analysis (disease prevalence and treatements).
Thanks in advance
Kalkidan S.

[Updated on: Mon, 13 October 2014 12:43]

Report message to a moderator

Re: v001 v002 v003 doesnot uniquely identify observation [message #3084 is a reply to message #3083] Mon, 13 October 2014 13:44 Go to previous messageGo to next message
Trevor-DHS is currently offline  Trevor-DHS
Messages: 787
Registered: January 2013
Senior Member
HIV test results are only available for women and men age 15-49. Diarrhoea prevalence and treatment, and nutrition and vaccine information are only available for children under 5. Anemia data are available for children under 5 and for women age 15-49 and for men age 15-49. I suggest constructing separate datasets for your analysis for children and for women and men. I also recommend thoroughly reviewing the questionnaires first to understand who has been asked which set of questions before deciding how to structure your data.
Re: v001 v002 v003 doesnot uniquely identify observation [message #3086 is a reply to message #3084] Mon, 13 October 2014 15:42 Go to previous message
kalfikirsisay@gmail.com is currently offline  kalfikirsisay@gmail.com
Messages: 11
Registered: October 2014
Location: NL
Member
Dear Trevor,
I thank you very much for your helpfull advices, I will structure the data and do the anlaysis separately.
Regards,
Kalkidan S.

[Updated on: Mon, 13 October 2014 15:44]

Report message to a moderator

Previous Topic: Merging child's malaria results from HH member recode dataset with other child data in Child recode
Next Topic: Can continuous DHS be merged with standard DHS?
Goto Forum:
  


Current Time: Fri Mar 29 12:00:57 Coordinated Universal Time 2024