Appending variables within the same dataset

[davidoh91]

Hi,

I have a dataset that looks something like this:

ID	v1a	v1b	v1c
1	1	1	2
2	1	3
3	2
4	5
5	3	3	1

Now, I want to append v1b and v1c on top of v1a in order to combine it into one variable of v1. The problem is this is within the same dataset, so I can't simply append using another DTA.

Is there an easier way to do this other than to create new DTA files and then append?

Thanks.

[Maarten Buis]

yes, see help reshape

// create the example data
clear
input ID     v1a     v1b     v1c
1     1     1     2
2     1     3   .    
3     2     .   .    
4     5     .   .    
5     3     3     1
end

// do the reshape
reshape long v1, i(ID) j(version) string

// admire the result
list, sepby(ID)

[davidoh91]

Thanks so much!

[Linda Lin]

I have a similar dataset and want to use this command. However, when I enter

insert id_pat_no v1 v2 v3

, I get the error message "variable id_pat_no already defined". The ID is not a single number as was the case for David, but rather a series of 6 numbers. I would love some advice on how to proceed.

Thanks!
Yi

[Rich Goldstein]

what is the -insert- command and where does it come from?

[Linda Lin]

what is the -insert- command and where does it come from?

Sorry, I mean -input-

[Jorrit Gosens]

Please provide an example of your data by using dataex, so that people will be able to answer this much easier. See more on how and why to use daatex in the FAQ: https://www.statalist.org/forums/help#stata

[Linda Lin]

Please provide an example of your data by using dataex, so that people will be able to answer this much easier. See more on how and why to use daatex in the FAQ: https://www.statalist.org/forums/help#stata

I have tried the dataex command but my data set is too large. I know snapshots of data sets are generally discouraged but it is the only way I can think of. My data looks like this

I'd like to calculate a rate (case/risk) for each disease for each country automatically using a loop command that applies the

gen rate_disease = *case/*risk

where the wildcard (*) is the various diseases included in the data.

I really appreciate your help!

[Jorrit Gosens]

You can select numbers of variables and observations to create a data example with dataex by doing e.g.,:

dataex Country Disease1_case Disease1_risk Disease2_case Disease2_risk in 1/20

Staring from your example, you coudl do:

foreach n of numlist 1/4 {
gen rate_disease_`n' = Disease`n'_case / Disease`n'_risk
}

edit: in the above code, replace the 4 with the number of diseases in your dataset

However, you could also consider whether your data is in the correct format for your analyses. Most Stata analyses would require long formats. You could get there by doing:

reshape long Disease, i(Country) j(disease) string
ren Disease value
gen var = substr(disease,3,.)
replace disease = substr(disease,1,strpos(disease, "_")-1)
destring disease, replace
reshape wide value, i(Country disease) j(var) string
ren valuecase case
ren valuerisk risk
gen rate_disease = case/risk

Edit no 2: You will also need to think what values need to be used when dividing by zero, which happens a lot with your data example