Hi Alex,

I decided to add another function signature to the
convert:binary-to-string function [1]: If you set the third parameter
to true(), invalid XML characters will automatically be replaced with
\uFFFD (the so-called Unicode replacement character):

  declare
    %rest:path("/upload")
    %rest:POST("{$body}")
  function local:store-csv($body as xs:base64Binary)
  {
    let $string := convert:binary-to-string($input, 'CP1251', true())
    return csv:parse($string)
  };

A new stable snapshot is available [2]; the final version of BaseX 8.5
will be released in June or July.

Hope this helps,
Christian

[1] http://docs.basex.org/wiki/Conversion_Module#convert:binary-to-string
[2] http://files.basex.org/releases/latest/

On Sun, Jun 12, 2016 at 11:06 PM, Alexander Shpack <shadowkin@gmail.com> wrote:
> Oh.. right. I forgot about it
>
> On Mon, Jun 13, 2016 at 12:02 AM, Christian Grün <christian.gruen@gmail.com>
> wrote:
>>
>> This line tells you the difference:
>>
>> > let $content := file:read-text($fName, "cp1251")
>>
>> You are reading the file contents as text (i.e., string). You could
>> also use file:read-binary. Using the browser, you can also upload
>> images or other binary contents.
>>
>>
>> > let $csv := csv:parse($content, map{ 'separator': ';', 'header': false()
>> > } )
>> > return $csv
>> >
>> > What the difference between post data and raw file?
>> >
>> >
>> > On Sun, Jun 12, 2016 at 11:45 PM, Alexander Shpack <shadowkin@gmail.com>
>> > wrote:
>> >>
>> >> Did you always upload the same file?
>> >>
>> >> Yeah
>> >>
>> >>>
>> >>> As the error indicates, it seems that your CSV input seems to contain
>> >>> characters that are not valid in XML. There are various ways to tackle
>> >>> this; one looks as follows:
>> >>>
>> >>>  (: interpret client data as Base64 :)
>> >>>  declare
>> >>>     %rest:path("/upload")
>> >>>     %rest:POST("{$body}")
>> >>>     function local:store-csv($body as xs:base64Binary)
>> >>>   {
>> >>>     (: replace invalid characters with a question mark :)
>> >>>     let $input := bin:from-octets(
>> >>>       bin:to-octets($body) ! (if(. >= 32 or . = (9, 10, 13)) then .
>> >>> else
>> >>> 63)
>> >>>     )
>> >>>     (: convert to XQuery Unicode string; convert to XML :)
>> >>>     let $string := bin:decode-string($input, 'CP1251')
>> >>>     return csv:parse($string)
>> >>>   };
>> >>>
>> >>> There may be easier solutions as well (I’ll give you an update once I
>> >>> remember them ;).
>> >>
>> >>
>> >> Strange solution... Ok, I'll wait for a better one.
>> >>
>> >> --
>> >> s0rr0w
>> >
>> >
>> >
>> >
>> > --
>> > s0rr0w
>
>
>
>
> --
> s0rr0w