Text Importing

[John Darrington]

On Mon, Nov 12, 2007 at 01:33:52PM -0800, Ben Pfaff wrote:
     Jason Stover <address@hidden> writes:
     
     > SPSS has a "text import wizard" that walks the user through the
     > process of reading text data. It asks the user questions about
     > delimiters and whether variable names are at the top of the file,
     > etc., then runs DATA LIST to read the data. This is a good function to
     > have: Any frustration at the beginning of learning to use a new

The version I've got access to runs GET DATA /TYPE=txt.  Here's an
example:

GET DATA  /TYPE = TXT
 /FILE = 'F:\text.txt'
 /DELCASE = LINE
 /DELIMITERS = " "
 /QUALIFIER = "'"
 /ARRANGEMENT = DELIMITED
 /FIRSTCASE = 2
 /IMPORTCASE = ALL
 /VARIABLES =
 x A3
 y A2
 z A7
 name A4
 .
CACHE.
EXECUTE.
DATASET NAME DataSet3 WINDOW=FRONT.

     
     I have some ideas on how to guess the format of a text file, so
     that in many cases it would be unnecessary to specify the format.
     The UI would be like this: specify that you want to read a file
     as text, then it would bring up a dialog that shows a few lines
     of text from the file and how they would be interpreted by
     default, with a set of controls that allow you to override the
     defaults.  If done competently, I think that this would indeed be
     "as easy as possible".

Sounds good to me.  FYI, I've posted some screenshots of spss' "text
import wizard" at http://darrington.wattle.id.au/SS

I don't think it's necessary to replicate that interface exactly.  But
it might give you some ideas about what spss users would be familiar
with.  

I don't think that spss makes any attempt to "guess" the format.  It
just gathers that information from the user, by asking a series of very
verbose questions.  If we can come up with some sort of AI heuristic
to guess the most likely format, then that might pleasantly suprise
some people. So long as there's a GUI widget to easily override the
guess, should it be wrong.

J'

-- 
PGP Public key ID: 1024D/2DE827B3 
fingerprint = 8797 A26D 0854 2EAB 0285  A290 8A67 719C 2DE8 27B3
See http://pgp.mit.edu or any PGP keyserver for public key.

signature.asc
Description: Digital signature