Re: Import excel file

[John Darrington]

On Tue, Apr 12, 2011 at 02:31:14AM -0300, Michel Boaventura wrote:
     Hello,
     
     Talking with Ben I decide to try to make an excel data import to pspp.
     I've found libxls (http://libxls.sourceforge.net/) an LGPL3
     library to help read this kind of file.
     
     I was able to make a small program that open an excel file and returns
     it as a csv. It would be very nice if you could help me to
     improve/plug it into
     pspp. Where do I start?


Importing XLS files would certainly be a good thing to have.   It sounds
like you already have a working proof of concept.

Where to start?  src/language/data-io/get-data.c contains the parser
for the GET DATA command.  Currently  we have the code: 

  if (lex_match_id (lexer, "GNM"))
    return parse_get_gnm (lexer, ds);
  else if (lex_match_id (lexer, "TXT"))
    return parse_get_txt (lexer, ds);
  else if (lex_match_id (lexer, "PSQL"))
    return parse_get_psql (lexer, ds);


So you'll have to add a "XLS" to that.  The options for importing an
excel spreadsheet are likely to be similar (if not identical) to those
for gnumeric.  So parse_get_gnm could probably be renamed to 
parse_get_spreadsheet (and suffer a few very small modifications).

Then,  you will need to implement the code to read the stuff in. This
would entail taking your program and turning it into an implementation
of a casereader which grabs rows from the excel file. Like this:

struct casereader * excel_open_reader (struct excel_read_opts *, struct 
dictionary **);

The code in src/data/gnumeric-reader.c can probably be used as a template.  
Again, much
of the code might well be identical to what you need.


So.  Get the syntax :

        GET DATA /TYPE=XLS /FILE= ....

all working and take it from there.

Post a message here, or shout on #pspp if you need help.


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