Re: Import of NIST ANOVA data sets
- To: mathgroup at smc.vnet.net
- Subject: [mg100068] Re: Import of NIST ANOVA data sets
- From: Bill Rowe <readnews at sbcglobal.net>
- Date: Fri, 22 May 2009 18:50:20 -0400 (EDT)
On 5/21/09 at 12:17 AM, morejunk at theward.com (Adrian) wrote:
>I am using Mathematica 7 and if I bring the data set into Excel 2007
>then save it as an .xls file, Mathematica imports the data fine but
>the then the ANOVA fails (Error: "ANOVA::arg1:The 1st argument has
>unequal columns or rows."). If I save the .xls file as a CSV file,
>import it into Mathematica and run the ANOVA, it runs successfully.
>Any ideas on why the xls import is breaking the ANOVA?
You really haven't provided enough information here to identify
exactly what is going on.
Clearly, there must be a difference between what you get when
you import the .csv file and the .xls file. So, the obvious
thing to do is look at those results in more detail. You can use
Mathematica tools like Depth, Dimensions and Length to determine
if the overall structure is the same for both.
Or you could do something like
Position[MapThread[SameQ, {csvData,xlsData}], False]
This will do a row by row comparison and return the position of
any rows that mismatch. That will give you more information as
to what the difference is between the two imports.
But while it should not be difficult to isolate the difference
and determine what is going on, why bring the data set into
Excel in the first place? NIST data sets are made available as
ASCII files. Mathematica can easily work with these directly and
in a far more flexible manner than Excel can.