How to calculate row mean

[Stefano Amato]

Dear all,

I need to create a new row according to one condition.

My data look like this:

* Example generated by -dataex-. To install: ssc install dataex
* dataex ID I1 I2 I3 
clear
input str61 ID double(I1 I2 I3)
"A" -.2512 -.2951 -.2552  
"B" -.2757 -.3962 -.6286  
"C"  -.3005 -.3327 -.2896  
"D" -.1695  .0448 -.1167 
"E" -.4392 -.6512 -.7259  
end

What I need is a new row called "F" for instance as mean between "D" and "E".

Thank you in advance for your help.

Best,
Stefano

[Marcos Almeida]

There must be a neat code for that, but you can use - egen - with rowmean() for D and E, then generate the mean value if str61 is equal to "D" or "E".

[Clyde Schechter]

tempfile copy
save `copy'

keep if inlist(ID, "D", "E")
collapse (mean) I1 I2 I3
gen ID = "F"
append using `copy'
sort ID
list, noobs clean

By the way, in Stata we do not speak of "rows" and "columns." They are referred to as "observations" and "variables," respectively. Stata is not a spreadsheet; rows and columns are spreadsheet concepts. If you speak about Stata in spreadsheet terms, you will tend to think about Stata as if it were a spreadsheet, and that will make it harder for you to use Stata effectively.

Actually, as I think about it, what you are trying to do here, add a new observation that is the mean of existing observations, is a very spreadsheet-like thing to do and is very un-Stata-ish. In fact, unless all you are doing is preparing the data for a visual display, you are probably going to really mess up future analyses by doing this. The circumstances where a Stata data set is actually improved by having an observation that summarizes the information in other observations are extremely few and far between.

[Joro Kolev]

There many solutions to this task, and there are many critiques of this task, here
https://www.statalist.org/forums/for...in-the-dataset

[Stefano Amato]

Thank you all for your help!

[Stefano Amato]

Clyde Schechter thank you for your valuable suggestions.