Merge two datasets wih the same variable names

[Mario Jimenez]

Hi there,

I have several datasets, each with the prediction error of different forecast models. I would like to merge the datasets in order to compare those errors and determine which predicts better. My problem is that I can´t merge them since the varname is the same in every dataset, so when I try to merge them Stata will automatically delete the merging dataset.

Any idea how can I merge them without having to change the varnames one by one, I am talking about 256 different datasets.

Thanks for your help beforehand!

Mario

[Wouter Wakker]

I think append would be a better option. If for some reason you need the data as seperate variables you can still reshape wide after append to get the same result.

[Mario Jimenez]

Thanks Mr. Wakker,

I am also trying to make a loop with it, any idea how?

[Jenny Williams]

Here is one way:

sysdir

/*LOAD YOUR FIRST FILE*/
use "C:\[INSERT YOUR FILE PATH FOR STATA]\auto.dta", clear

rename price price0

/*LOOP THROUGH ALL OTHER DATA SETS AND RENAME THE DESIRED VARIABLE (PRICE) ACCORDING TO THE LOOP NUMBER*/
/*OBVIOUSLY, PRICE WILL BE DUPLICATED HERE BUT WILL BE DIFFERENT FOR EACH OF YOUR DATA SETS*/

local iteration=0
forvalues filenumber=1/256 {
local iteration=`iteration'+1

/*ASSUMING YOUR FILES ARE NAMED SEQUENTIALLY, YOU WILL REFER TO THEM USING THE NUMBERING IN FORVALUES*/
*quietly merge 1:1 make using "C:\Program Files (x86)\Stata13\auto`filenumber'.dta", keepusing(price)

/*BUT FOR THIS EXAMPLE, THE SAME DATASET WILL BE MERGED REPEATEDLY*/
quietly merge 1:1 make using "C:\Program Files (x86)\Stata13\auto.dta", keepusing(price)

drop _merge
rename price price`iteration'
}

browse

[Mario Jimenez]

Thanks Jenny, exactly was I was looking for!

[Wouter Wakker]

Or:

cd "C:\Users\Your\Path\Here"
clear
append using `: dir . files "*.dta"', generate(source)
reshape wide whatever, i(id) j(source)