All variables are strings but I need it in numeric

[Herman Bloemhof]

Dear all,

I want to do a research on Swedish stocks. I have gathered a lot of data for this and now I am trying to work with it (do a time series analysis). But the problem is that all the data is still in string data. When I am trying to destring data and replace (destring variable_name, replace) It says: WC01001: contains nonnumeric characters; no replace. But there are no nonnumeric characters that I can see. I added in the Attachments a small part of my data because the whole file was too big to upload. I need to have all these variables in numerical variables so I can do regressions etc.
Anyone know how to resolve this?

Herman

[Nick Cox]

Please see FAQ Advice #12

https://www.statalist.org/forums/help#stata

on not posting spreadsheet attachments (please) and on using dataex (please).

You were asked to read that before posting.

Your problem is the spreadsheet-style header in observation 1 -- which contains many non-numeric characters. Stata doesn't want that stuff as data. It wants it as metadata. Here it seems that the headers would work fine as variable names in Stata.

 
orgno
name
year
WC01001
WC01051
WC01151
WC01100
WC01101
WC01250
WC01253
WC01254
WC01266
WC18191
WC01251

You can go back to import excel or import delimited and specify so that the first observation (row) becomes variable names. The dialog box versions

db import excel
db import delimited

make that easy.

Or you can fix the problem by something like this.

foreach v of var * {
    local newname = `v'[1]
    capture rename `v' ``newname'
    label variable `v' "`newname'"
}

drop in 1

and then try

destring, replace