For this kind of thing, I would use awk. There is also sqlite which
might be ported to your OS, you didn't specify, and this supports outer
joins.

I am not aware of a matrix operation that will sort/uniq muliple ranges
of data which would be neat.

Attached is a simple vlookup based approach you might use. You just need
to be able to have a complete set of student ids for the result sheet.

I am not sure how to get that complete set of student ids inside
gnumeric, but in the unix shell it is very simple.

Suppose you have 2 csv files of exam results, both with the first column
containing student ids.

cut -d, -f1 exam1.csv exam2.csv | sort -u > studentids.csv



On Fri, 2005-09-09 at 20:47 -0400, Alexander Kirillov wrote:
> Hi guys: 
> maybe one of the experts here can provide some advice...
> 
> I have two spreadsheets containing grades for two different exams
> (midterm 1, midterm2) for  the same class. For simplicity, we can
> assume that each contains the following filed:
> ID Name Grade 
> 
> I need to merge them creating a single spreadsheet with all grades for
> all students who took at least one of the exams. If the list of
> students was the same, I could just copy-and-paste a column. However,
> the lists are slightly different: some students missed one midterm,
> others the second midterm, some dropped the class, etc.
> 
> Is there any simple way of merging them, short of manually matching
> (there are about 900 students)? Of course, if using a database - such
> as MySQL - it would be exactly one command [AFAIK, it is called "full
> outer join"]. But I do not know MySQL, and learning it, converting the
> data, and then converting back seems too much of a hassle for such a
> simple thing.
> 
> Is there any way to do it in Gnumeric? Or in OpenOffice.org? In Excel
> if there is no other way? [BTW: Microsoft Access database doesn't
> support outer join. Typical for point-and-click products)
> 
> Thanks,
> Sasha
> 
> PS: I know the right way would be to keep all the data in a single
> spreadsheet, avoiding this problem. However, for reasons you do not
> want to know, it was impossible - and it is too late now, anyway.
> _______________________________________________
> gnumeric-list mailing list
> [email protected]
> http://mail.gnome.org/mailman/listinfo/gnumeric-list
> 

Attachment: eg-exam.gnumeric
Description: application/gnumeric

_______________________________________________
gnumeric-list mailing list
[email protected]
http://mail.gnome.org/mailman/listinfo/gnumeric-list

Reply via email to