Create a variable that contains the name of the datafile

[Ronny Kalinka]

Dear Stata community,

I am struggling with a probably simple operation. I have data that contains the date and number of users holding a stock. The stock ticker however is not a variable, but part of the filename (e.g. "A.dta").

Now I want to create a ticker variable that contains the filename without the ending (e.g. "A").

Has anyone a hint on how I can achieve this?

Best regards!



PS: Ideally I want this step to be a part of a script that runs through a whole folder, as I have thousands of separate stock files.

[Nick Cox]

See the results of creturn list -- which point to one solution, namely

gen filename = subinstr("`c(filename)'", ".dta", "", 1)

[Ronny Kalinka]

Thank you Nick!

May it be the case, that this only works, if your code is put directly after loading the file?

Because when I edit the dataset after loading, and only then try to generate the new variable, I get an empty column

[Nick Cox]

You can answer your own question. The result depends on a filename being present in c(filename) which you can check with a display. When I looked at your question my Stata has some sandbox data in that didn't come from a .dta file, but then I read in one such, and c(filename) is now populated.

. di "`c(filename)'"


. sysuse auto, clear
(1978 automobile data)

. di "`c(filename)'"
C:\Program Files\Stata18\ado\base/a/auto.dta

That said, editing the data shouldn't make a difference. (I just checked and it doesn't.)

[Hemanshu Kumar]

If you are reading in data from a large number of files in a folder, don't you already have a macro with the name of the file in it?

I would imagine some code of the following type:

local data_dir "path/to/data/folder"
local data_files: dir `"`data_dir'"' files "*.dta"

foreach this_file of local data_files {

    use `"`data_dir'/`this_file'"', clear
    ...
}

In this case, you don't need the macro c(filename) at all, that purpose is served by the macro this_file in the above toy code. The macro is available while you're within the foreach loop, but doesn't need the ticker data file to actually be in memory.