Unrecognized characters in filename when importing using: "local myfiles: dir "directory" files "*.csv"

[Adi JJahic]

Hello!

I have a file named "abc+def+mm+dd+yyyy.csv" in directory "$path/survey responses". The file is being downloaded directly from Qualtrics and I want to not have to change the file name.

However, when I run the following code:

local myfilelist : dir "$path/survey responses" files "*.csv"
foreach file of local myfilelist {
    
    di in red "`file'"
    import delimited using "`file'" , clear varnames(1)
}

the code is unable to find the correct file to import because this line of code "local myfilelist : dir "$path/survey responses" files "*.csv"" removes the plus signs and interprets it as "abc def mm dd yyyy.csv".
Is there anything I can do to solve this?
Perhaps, is there another way I can import an arbitrary file name from a specific directory?

Thanks for the help everyone!

[Clyde Schechter]

I cannot replicate the problem you are having in my setup:

. dir *.csv
   0.3k   1/02/23 15:53  abc+def+mm+dd+yyyy.csv
   1.6k   4/09/16 10:37  abortions_by_state_2009-2012.csv
   0.0k  11/19/14  4:20  as_strings.csv    
   0.1k   9/23/17 10:35  commatest.csv     
  42.9k  12/30/21  9:39  dca.csv           
   0.0k  11/19/14  4:20  formatted.csv     
  10.3k   1/26/18  4:50  SAT.2004.csv      
   6.6k  11/30/15  6:36  state_table.csv   
   0.3k  12/10/18  7:55  test.csv          
   0.0k  11/19/14  4:20  unformatted.csv   

.
. local myfilelist: dir "." files "*.csv"

. display `"`filelist'"'


.
. foreach file of local myfilelist {
  2.     display in red "`file'"
  3.     import delimited using "`file'", clear varnames(1)
  4. }
abc+def+mm+dd+yyyy.csv
(encoding automatically selected: UTF-8)
(15 vars, 4 obs)
abortions_by_state_2009-2012.csv
(encoding automatically selected: ISO-8859-1)
(6 vars, 56 obs)
as_strings.csv
(encoding automatically selected: UTF-8)
(1 var, 3 obs)
commatest.csv
(encoding automatically selected: ISO-8859-1)
(5 vars, 2 obs)
dca.csv
(encoding automatically selected: ISO-8859-1)
(10 vars, 750 obs)
formatted.csv
(encoding automatically selected: UTF-8)
(1 var, 3 obs)
sat.2004.csv
(encoding automatically selected: ISO-8859-1)
(51 vars, 61 obs)
state_table.csv
(encoding automatically selected: ISO-8859-1)
(17 vars, 56 obs)
test.csv
(encoding automatically selected: UTF-8)
(15 vars, 4 obs)
unformatted.csv
(encoding automatically selected: UTF-8)
(1 var, 3 obs)

I am running Stata version 17 on Windows 10.

[William Lisowski]

There is a difference between post #1 and post #2.

The dir returns only the filename portion, you need to include $path in your import command. The example in post #2 worked because the data was in the current working directory.

. local myfilelist : dir "~/Downloads" files "*.png"

. macro list _myfilelist
_myfilelist:    "ugreen.png" "Better World Books.png"

[Adi JJahic]

Hello! Thank you Clyde. This is very interesting-
This is the output I received at first --

. local myfilelist : dir "$path/survey responses" files "*.csv"

. foreach file of local myfilelist {
  2.         di in red "`file'"
  3.         import delimited using "`file'" , clear varnames(1) 
  4. }
abc def mm dd yyyy.csv
file abc def mm dd yyyy.csv not found

Then, when I made these changes to make my code look more like yours:

. version 17
. cap cd "$path/survey responses"

. dir "*.csv"
   2.2k   1/02/23 14:11  ABC+def+Mm+dd+yyyy.csv

. local myfilelist : dir . files "*.csv"

. foreach file of local myfilelist {
  2.         di in red "`file'"
  3. 
  4.      import delimited using "`file'" , clear varnames(1) 
  5. }
abc+def+mm+dd+yyyy.csv
(encoding automatically selected: ISO-8859-1)
(22 vars, 4 obs)

Not quite sure why it works now, but it seems to be. Thank you. I'm also working on version 17 windows 10. It is possible that earlier I was working on windows 14/15 since I may have ran a do-file wit the "version 14" command.

Thank you for your help @ClydeSchechter