Re: How to build a program that can read the content of a very simple MS Excel file?

[John Machin <sjmachin lexicon net>]

On 13/06/2007 11:24 AM, Huge Mountain wrote:
> Hello everyone.
> I'm trying to build a program to read the content of a Microsoft Excel 
> file, just read a very simple Excel file (a file with text in cell A1, 
> A2,..., don't care about formula, chart,...). I simply want to read the 
> content of cell A1.  With that, I can get the binary format of MS Excel 
> file more clearly.
> Ex:if cell A1 have the content: "text message", my program can show 
> "text message" as result.
> Could someone tell me how to do that, or where to find the references? I 
> downloaded source, file " gnumeric-1.6.3.tar.gz" from 
> http://www.gnome.org/projects/gnumeric/, but I don't know how to use, 
> which to compile. I'm stray  with >1000 files (some file I don't know 
> how to open).
> Please help me!
> Thank in advance!
>
> PS:  I explored a simple MS Excel file (with "abc" in cell A1, this file 
> was saved by MS Excel 2000). This file have 3 stream: SummaryInfomation 
> stream, DocumentSummaryInfomation stream and Workbook stream. The only 
> stream I care is Workbook stream, started from byte 512 to byte 
> (512+4096) (this stream has the 4096 bytes length). So, where are bytes 
> which contain the "abc" string of cell A1 in this 4096 bytes?
> Please help me!

The format of an XLS file is extremely complicated. If you are desperate 
enough to want to do it yourself, see 
http://sc.openoffice.org/excelfileformat.pdf and note that the Workbook 
stream is not necessarily contiguous; you'll nead to read 
http://sc.openoffice.org/compdocfileformat.pdf.

If you want to try a pre-written Excel-reading library:
(1) Python: http://www.lexicon.net/sjmachin/xlrd [I'm the author]
(2) perl: look for Spreadsheet::Excel (or something like that) in cpan
(3) Java: http://poi.apache.org/hssf/index.html

HTH,
John