Re: csv read problem

[Marco Atzeri]

On 18/02/2016 18:31, Dave Cottingham wrote:
> I sounds like this file is in a multi-byte unicode charset, e.g. utf-16 or
> ucs-2, in which case you could solve your problem by translating it to
> something octave can read, like utf-8.
>
> Most linux distros come with the iconv utility. You could try experimenting
> with something like
>
> iconv -f utf-16 -t utf-8 <bad.csv >good.csv
>
> and see if it works better.

On my iconv the coding is called
  CP1200 UCS-2LE UCS2LE UNICODELITTLE

$ iconv -f CP1200 -t utf-8 <bad.csv >good.csv

$ head good.csv
IdProvincia;IdEstacion;Fecha;Año;Dia;Precipitación (mm)
3;2;01/01/2000;2000;1;0
3;2;02/01/2000;2000;2;0
3;2;03/01/2000;2000;3;0
3;2;04/01/2000;2000;4;0,2
3;2;05/01/2000;2000;5;0,2
3;2;06/01/2000;2000;6;0,2
3;2;07/01/2000;2000;7;0
3;2;08/01/2000;2000;8;0
3;2;09/01/2000;2000;9;0

Moreover the line termination is CR LF.

Regards
Marco