Re: Excel/POI - Here We Go Again -- RPG400-L

Strangely, the 'open' returns an address of ZERO for 'book' but does not
halt. The exception occurs only when the HSSF_GETSHEET fails because the
'book' address is 0.

That's not strange at all. In an earlier discussion someone (I thought itwas you, actually?) was trying to handle situations where hssf_open()could fail. I think he was trying to open a file that might not exist,and wanted a way to handle the error properly. My memory is fuzzy, but itwas something like that.

Hang on, I'll look up the message in the archives... yes, it was you.Here are the key messages in the thread where we discussed it.

http://archive.midrange.com/rpg400-l/200608/msg00733.html
http://archive.midrange.com/rpg400-l/200608/msg00745.html

I probably didn't tell you that I changed it, though.. my fault. Theidea, though, is that instead of crashing (and taking the JVM down) when afile can't be opened for some reason, it now returns *NULL. Your programand therefore check for that.


     if HSSF_open( '/path/to/myfile.xls' ) = *NULL;
          ErrMsg ='Unable to open Excel document';
          ShowErrorToUser(Errmsg);
          iter;  // try next document.
     endif;

The following part of your message is the REAL problem, the reason whyit's failing and therefore returning zero:

  received Java exception
"org.apache.poi.hssf.record.RecordFormatException:
  Unable to construct record instance, the following" when calling method
  "<init>" with signature
  "(Lorg.apache.poi.poifs.filesystem.POIFSFileSystem;)V" in class
  "org.apache.poi.hssf.usermodel.HSSFWorkbook".

Excel documents are made up of specially formatted "records" (they aren'tdatabase records in the manner that we're used to thinking of, but theyare a record of sorts.)

When the Java code (HSSF) is reading the file, it needs to create a newJava record object to handle the record that it just loaded from the Exceldocument. However, for some reason it's unable to do that.

I would suspect that either it doesn't understand the format of the record(maybe it's a feature that HSSF doesn't support) or there's a bug in theApache HSSF code. Something like that.

Unfortunately, that's outside of my area of expertise. You could tryupgrading POI to see if it's fixed in a newer version, or you might trysearching (Google) for the Java error to see if anyone else hasencountered it and found a workaround.