Date:         Wed, 28 Oct 2009 09:08:09 -0400
Reply-To:     Gerhard Hellriegel <gerhard.hellriegel@T-ONLINE.DE>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Gerhard Hellriegel <gerhard.hellriegel@T-ONLINE.DE>
Subject:      Re: convert SPSS to SAS
Content-Type: text/plain; charset=ISO-8859-1

Hi Andre,
interesting! I think it is important simply to know how the formats look
like, because normally they are used to build groups.
Thanks,
Gerhard



On Wed, 28 Oct 2009 10:15:28 +0100, Andre Wielki <wielki@INED.FR> wrote:

>Gerhard,
>
>Yes Spss has also 'formats' called value labels  and in matter of
>missing value has 3 possibility
>those labels are of 200$. long. All this is internal to the table.
>
>The version 17 of spss has a well working export to SAS   data + saspgm
>with proc format + data step
>to make the users formats  internal to the table.
>
>the proc import of SAS 9.1.3  (9.2 not yet tested in that matter)
>is importing the spss file and normally creating and adapting the
>formats names
>
>Now what is the sas rule for a format name  : not ending with a number
>and starting with a $ incase of   and last 31 maxi for length.
>
>The initial asker did not precise well the context of his problem : a
>message of proc import or
>a later problem in matter of user
>
>Stattransfer too is creating a data + the same kind of sas pgm like Spss.
>The advantage over sas proc import is that the length of numeric
>variable is adapted
>the result is relative low inflation of size from spss file to sas file.
>
>In matter of proc export(9.1.3), is Sas adapting the future spss data
>with applying old strength rules
>spss previous versions and the result is that you have a lot of things
>changed.
>What about the lengths : i don't remember well but i think SAS is
>maintaining a length of 8!!!!!
>for all numeric variable while the others solutions are not doing this!
>
>The import from spss17 is working mostly well  EXCEPT with a table with
>user format
>and the programmatic solution of spss is not working on  sas table +
>format table!
>
>(not yet tested under spss18 which was announced with a (fuller!) import
>of sas tables)
>
>So the stattransfer is the optimized solution for this kind of work
>and in any cases Stattransfer may also produce a .dat + correct pgm  to
>read it in sas stata spss...
>
>HTH
>Andre
>
>
>Gerhard Hellriegel a �crit :
>> that is also my problem: if that is a message of proc import, that thing
>> can only come from the source (the output table is new!). That implies
>> that SPSS also has formats (?) like SAS. I don't know SPSS, so I can't
>> say. If it is similar to SAS, there must also be a format catalog (?)
with
>> user-formats. Maybe that can also be imported.
>> Gerhard
>>
>>
>>
>> On Tue, 27 Oct 2009 18:18:09 +0100, Andre Wielki <wielki@INED.FR> wrote:
>>
>>
>>> Darren
>>> see inyour work
>>> normaly you have a format catalog
>>>
>>> then
>>> to open your newfilename sas table (probably no more in the work) you
>>> must before  submit this
>>>
>>> options fmtsearch=(work.formats);
>>>
>>> did this solve your problem
>>> or was this message a message of proc import?
>>> Andre
>>>
>>> darren Brust a �crit :
>>>
>>>> I used the following code to convert SPSS file to SAS dataset
>>>>
>>>> PROC IMPORT DATAFILE='temp.sav'
>>>> OUT=work.newfilename
>>>> DBMS=SAV replace;
>>>> RUN;
>>>>
>>>> but I cannot get the formats loaded.
>>>>
>>>> the error message: 'ERROR: The format D27B was not found or could not
be
>>>> loaded.'
>>>>
>>>> please help!
>>>>
>>>>
>>>>
>>> --
>>>   Andr� WIELKI
>>>   INED (Institut National d'Etudes D�mographiques)
>>>   Service Informatique
>>>   133 Boulevard Davout      75980 Paris Cedex 20
>>>   m�l : wielki@ined.fr      t�l :  33 (0) 1 56 06 21 54
>>>
>>
>>
>
>--
>   Andr� WIELKI
>   INED (Institut National d'Etudes D�mographiques)
>   Service Informatique
>   133 Boulevard Davout      75980 Paris Cedex 20
>   m�l : wielki@ined.fr      t�l :  33 (0) 1 56 06 21 54