Difference between revisions of "Calc/ODFF Implementation/Examine functions"

From Apache OpenOffice Wiki
Jump to: navigation, search
(Other Functions)
(Statistical Functions)
Line 81: Line 81:
 
! NORMDIST
 
! NORMDIST
 
| Last Parameter has type Logical and is optional || Last Parameter has type Number and is required || || ||
 
| Last Parameter has type Logical and is optional || Last Parameter has type Number and is required || || ||
 +
|-
 +
| || ||Accuracy is insufficient for x<−2 because cancellation.||New implementation with ERFC.||{{Bug|69069}} and for implementation of ERFC {{Bug|97091}}||
  
 
|-
 
|-
 
! LEGACY.NORMSDIST
 
! LEGACY.NORMSDIST
 
| "This is exactly NORMDIST(x)." Should be "This is exactly NORMDIST(x;0;1)." because mean parameter and standard deviation parameter are not optional.|| || ||Mail to OASIS [http://lists.oasis-open.org/archives/office-comment/] ||
 
| "This is exactly NORMDIST(x)." Should be "This is exactly NORMDIST(x;0;1)." because mean parameter and standard deviation parameter are not optional.|| || ||Mail to OASIS [http://lists.oasis-open.org/archives/office-comment/] ||
 +
|-
 +
| || ||Accuracy is insufficient for x<−2 because cancellation.||New implementation with ERFC.||{{Bug|69069}} and for implementation of ERFC {{Bug|97091}}||
  
 
|-
 
|-

Revision as of 19:48, 10 December 2008

Examine Changes in Functions for ODFv1.2

This is a collection of the differences I found and what I know is already done or in progress. I also note, when I found problems in the draft spec itself. You are invited to add your observations. Regina 19:40, 10 October 2008 (CEST)

Statistical Functions

Function ODFv1.2 CALC Action Progress
BETADIST Draft Spec misses faktor 1/(b-a) in densitiy case not yet implemented Inform Eike Rathke Mail to OASIS [1]
Has density case and optional parameter, to choose between density and cumulative not yet implemented expand function to density case and additional parameter tracked in Issue 91547 , function code is ready, finishing implementation depends on Issue 91602
BINOMDIST Parameter C is integer with constraint to 0 and 1. It should be Logical as in other distributions C=0 is density, all other values of C are treated as cumulative
LEGACY.CHIDIST Returns 1 for x <0 x < 0 returns Err:502 expand domain to x < 0 internal patch in Issue 94555
CHISQDIST density function and cumulative left tail of chi-square distribution not yet implemented prepare Implementation patch in Issue 94555
CHISQINV inverse of CHISQDIST(...TRUE()) not yet implemented prepare Implementation patch in Issue 94555
EXPONDIST Parameter Cumulative has type Logical and is optional. Last Parameter C has type Number and is not optional. C=0 returns denisity, other values for C return cumulative function.
FDIST density function and cumulative left tail of F-distribution not yet implemented prepare for implementation first draft, depends on BETADIST Issue 91547
Name conflicts with actual FDIST, which calculates right tail. get new UI-name have brought problem to dev@sc
FINV inverse of new FDIST not yet implemented
Name conflicts with actual FINV, which calculates inverse for right tail. get new UI-name
GAMMA Parameter is constraint to >= 0 Gamma is definied for all real numbers, but non-positive integers Inform Eike Rathke
not yet implemented prepare implementation using full domain. Add UI texts. patch in Issue 94555
is in subgroup Mathematical functions will go to subgroup Statistical functions Mail to OASIS [2]
GAMMALN is in subgroup Mathematical Functions is in subgroup Statistical Functions
GAMMADIST Last Parameter has type Logical Last Parameter has type Number Change type internal to bool. Change UI texts. patch in Issue 94555
Domain for Parameter x includes x < 0 x < 0 returns Err:502 in OOo2.* Change implementation integrated in OOo3.0
GROWTH Last parameter has type Logical (in Excel too) Last parameter has type Number
is in subgroup Statistical Functions (in Excel too) is in subgroup Array Functions
HYPGEOMDIST Optional, logical parameter to calculate "exact" or "at most" no parameter, calculates always "exact"
INTERCEPT Current, "Calculates the point at which a line will intersect the y-values by using known x-values and y-values." (BTW: That is literally from Excel help.) Problems:1. It calculates no point but a single value. 2. The line does not intersect the y-values but the y-axis. 3. It does not say, that the line is the linear regression of the known values. Mail to OASIS [3]
NORMDIST Last Parameter has type Logical and is optional Last Parameter has type Number and is required
Accuracy is insufficient for x<−2 because cancellation. New implementation with ERFC. Issue 69069 and for implementation of ERFC Issue 97091
LEGACY.NORMSDIST "This is exactly NORMDIST(x)." Should be "This is exactly NORMDIST(x;0;1)." because mean parameter and standard deviation parameter are not optional. Mail to OASIS [4]
Accuracy is insufficient for x<−2 because cancellation. New implementation with ERFC. Issue 69069 and for implementation of ERFC Issue 97091
PERCENTRANK It has an optional parameter Significance not yet implemented
PHI Description is unclear. PHI(x)=NORMDIST(x;0;1;FALSE()). It is the density function of the standard normal distribution Mail to OASIS [5]
POISSON Parameter Cumulative has type Logical and is optional. Parameter 'C' is required. 'C'=0 is density function, all others are cumulative change in calc work in progress (Ivyue) Issue 96837
QUARTILE Example calculates with Quart=1.5 although there is a constraint INT(Quart)=Quart Mail to OASIS [6]
RANK Description "If not 0, Data is ranked in descending order." has to be "If not 0, Data is ranked in ascending order." Mail to OASIS [7]
SKEWP not yet implemented
STDEVA The text "cells with text are converted to 0;" conflicts with "The handling of strings is implementation defined." Mail to OASIS [8]
TDIST Draft spec shows the density function. Calc (and Excel) evaluates the cumulative function. Inform Eike Rathke Mail to OASIS [9]
TREND The optional parameters knownX and newX may be empty, denoted as ;; knownX and knownY cannot be empty
Parameter AllowOffset is Logical Parameter is 0 or not 0
TRIMMEAN The limits in the definition formula are wrong. Either count i from 0, then sum from i=cutOff to n−1−cutOff, or count i from 1, then sum from i=cutOff+1 to n−cutOff Mail to OASIS [10]
TTEST (1) All formulas miss the degrees of freedom. It is necessary for TDIST. In case unequal variance there must be defined, whether the degrees of freedom are made integer by FLOOR or whether they remain real. (2) case paired: is wrong because it is already contained in . (3) case equal variance: definition of s² is missing. (4) case unequal variance: definition does not meet usual definitions. In fact it is the definition for the equal variance case, with pooled variance.(5) As TDIST actually returns the values of the CDF, this formulas has to be revised together with TDIST. Mail to OASIS [11]
WEIBULL Parameter Cumulative is logical. Parameter is 0 or not 0
Alpha and Beta are exchanged in the formula, compared with the current implementation in OOo, Excel and Gnumeric. Mail to OASIS [12]
ZTEST Calc calculates z-value wrong. Read discussion in Issue 90759 Correct sigma-bug Patch in Issue 90759
Uses instable algorithm for standard deviation. Change to stable version or use STDEV directly.
Calculates Beside the sigma-bug OOo calculates for z>0 and for z<0 Mail to OASIS [13]
Functionname ODF Calc Action Progress

Other Functions

Function ODFv1.2 CALC Action Progress
GCD "Return the largest value that can be evenly divided (no remainder) into the given numbers." Should be "Return the largest value that can divide the given numbers evenly (no remainder)." - - Mail to OASIS [14]
LOG Base (third parameter) is optional with default 10. All parameters are required. Change in Calc work in progress (Ivyue) Issue 96835
Functionname ODF Calc Action Progress

Personal tools