st: Re: Bizzare return code in Mata

["Thomas, Anthony" <anthony_h_thomas@brown.edu>]

Also if it is helpful, here is the code up to that point:

get = fopen("`1'", "rw")
line = "HI"
while(strlen(line) <= `3'){
         line = fgetnl(get)
}

t = tokeninit("", ",", `""""')
tokenset(t, line)

variables = tokengetall(t)

variables = select(variables, strmatch(variables, ","):!=1)
variables[`2'] = char(select(ascii(variables[`2']), ascii(variables[`2']):!=34))
variables[`2'] = "#!!@$*&!&" + variables[`2']

variables = invtokens(variables, ",")
---> This is where error happens fseek(get, 0, 1)
whereAmI = ftell(get)


On Mon, Mar 17, 2014 at 8:48 AM, Thomas, Anthony
<anthony_h_thomas@brown.edu> wrote:
> Hi All,
>
> I need Stata to always import certain columns of a set of CSV files as
> string. Insheet (Stata 11 - Unix) does not provide any options for
> coercing variables to string so I wrote a small Mata utility which
> appends some non-numeric characters to the first observation in a
> specified column, copies the observation to the end of the file,
> insheets the data (which will now be read as string) and then
> truncates the file before the new line to restore the file to its
> original status.
>
> In one CSV (the program works fine for all others), the call fseek(fh,
> 0, 1) to seek the end of the file, aborts with the error r(1458296728)
> - Stata returned error. I strongly suspect this is not a valid return
> code... Also I'm a bit confused as to why Stata needs to be involved
> at all. I assumed Stata called Mata's file routines, not the other way
> around. Stata will happily insheet the file, and a hexdump of its
> contents reveals no irregularities. Does anyone have any suggestions
> as to what to look for or why I am getting this strange return code?
>
> Thanks,
>
> Anthony
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/