I agree with returning zero for the absolute frequencies. For the relative frequencies, I would argue for returning NaN as is done in BivariateRegression when there is insufficient data.
Brent Worden -----Original Message----- From: Phil Steitz [mailto:[EMAIL PROTECTED] Sent: Wednesday, August 11, 2004 3:45 PM To: Jakarta Commons Users List Subject: RE: [Math] Getter methods in Frequency when there is no data Thanks for reporting this. I agree that getCumFreq should return 0, but the pct and cumPct are more problematic. One could argue that an IllegalStateException would be more appropriate for these. If there have been no values added, percentages don't really make sense. Thoughts? Phil -----Original Message----- From: Shing Hing Man [mailto:[EMAIL PROTECTED] Sent: Wed 8/11/2004 7:41 AM To: commonMath Cc: Subject: [Math] Getter methods in Frequency when there is no data Hi, I have tested all the getter methods of org.apache.commons.math.stat.univariate.Frequency when no data has been added. (If anyone is interested, the JUnit test file is attached.) Here is a summary of the result. Method Result getCount returns 0 (as expected) getCumFreq java.util.NoSuchElementException getCumPct java.util.NoSuchElementException getPct returns NaN f.getSumFreq return 0 (as expected) Is it reasonable to expect all the getter methods to return 0, when there is no data ? My version of org.apache.commons.math is dated 10th August, 2004. regards, Shing ===== Home page : http://uk.geocities.com/matmsh/index.html ___________________________________________________________ALL-NEW Yahoo! Messenger - all new features - even more fun! http://uk.messenger.yahoo.com --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
