Am 12.12.2013 12:48, schrieb "Juan C. Estévez":
En 11/12/2013 22:26, Cottrell, Allin escribiu:
> On Mon, Dec 9, 2013 at 10:18 AM, Cottrell, Allin <cottrell(a)wfu.edu>
>> On Mon, Dec 9, 2013 at 9:45 AM, Sven Schreiber <svetosch(a)gmx.net> wrote:
>>> here's an (alleged) SPSS file downloaded from Eurostat which won't
>>> imported into gretl (TIME field has invalid length or somesuch is the
>>> I have no idea whether that's a gretl import bug or the file is not
>>> according to spec. Maybe somebody out there has SPSS and can tell.
>> It may be the big endianness of the file that's a problem (or
>> possibly the
>> fact that it was created using pspp rather than SPSS). I'll see what
>> I can
>> find out.
> There was a bug in the facility for handling big-endian SPSS files on
> little-endian systems. That's now fixed in CVS. However, this file is
> strange and arguably mal-formed. I opened it in pspp (I don't have
> SPSS) and it looked just as strange there as in gretl.
I have SPSS and Allin is right. SPSS reads 380 rows with strange data.
As you can see in image attached, the variables 'SEX' and 'AGE' always
have the same codes, and
'value' has strange codes when there's no data.
That is actually kind of correct, SEX and AGE are constant metadata. The
colon ":" as the missing code also appears in other output formats
(remember that is data directly from Eurostat).
But it certainly is more like a database dump wrapped in some
superficial spss format, not really like a dataset that is ready to go.
thanks for looking at this,