Opened 7 years ago
Closed 7 years ago
#15306 closed defect (fixed)
partial_fraction_decomposition over QQ[] wrong
Reported by:  pong  Owned by:  

Priority:  major  Milestone:  sage6.2 
Component:  algebra  Keywords:  
Cc:  robertwb, dkrenn, rws  Merged in:  
Authors:  Robert Bradshaw, Marc Mezzarobba  Reviewers:  Marc Mezzarobba, Robert Bradshaw 
Report Upstream:  N/A  Work issues:  
Branch:  5f4a857 (Commits)  Commit:  5f4a8579c7453dc3c2de8ecae9b455ab36c47f64 
Dependencies:  Stopgaps: 
Description
The method partial_fraction_decomposition() does not work correctly.
Compare this
sage: p = 2*x; q = x^2 + 6*x +9; r = p/q sage: r.partial_fraction() Out[2]: 2/(x + 3)  6/(x + 3)^2
with
sage: R.<x> = PolynomialRing(QQ) sage: p = 2*x; q = x^2 + 6*x +9; r = p/q sage: r.partial_fraction_decomposition() Out[5]: (0, [2*x/(x^2 + 6*x + 9)])
Change History (18)
comment:1 Changed 7 years ago by
 Cc dkrenn added
comment:2 Changed 7 years ago by
 Summary changed from partial_fraction v.s. partial_fraction_decomposition to partial_fraction_decomposition wrong
comment:3 Changed 7 years ago by
 Summary changed from partial_fraction_decomposition wrong to partial_fraction_decomposition over QQ[] wrong
comment:4 Changed 7 years ago by
 Cc rws added
comment:5 Changed 7 years ago by
 Status changed from new to needs_info
comment:6 Changed 7 years ago by
Hum... I don't understand, shouldn't the output of r.partial_fraction() and r.partial_fraction_decomposition() be "equivalent"? They are not in the example that I gave. For the rational function 2x/(x^{2} +6*x +9), the method partial_fraction() gives the correct answer while partial_fraction_decomposition() does not.
comment:7 Changed 7 years ago by
 Cc robertwb added; Robert Bradshaw removed
It looks like partial_fraction_decomposition
does not even attempt to return numerators of degree smaller than that of the irreducible polynomial in the denominator. The reason for that (I guess) is that partial_fraction_decomposition
is defined for all elements of parents belonging to QuotientFields()
, "the category of quotient fields over an integral domain".
But while trying to fix the issue I realized that I am a bit confused:
 Is partial fraction decomposition really defined over the field of fractions of any integral domain R? (I agree that it is when R is a PID, but what about other cases?)
 In any case, isn't its computation going to require a Euclidean domain?
 But in the Euclidean case, the stronger partial fraction decomposition pong asks for is always defined, isn't it?
 What is the correct way to define methods that apply to all elements of the field of fractions of a Euclidean domain? Conversely, should there be a category of quotient fields over an integral (as opposed to Euclidean) domain in Sage at all?...
comment:8 Changed 7 years ago by
Please ignore my previous comment.
comment:9 Changed 7 years ago by
 Branch set to u/robertwb/ticket/15306
 Created changed from 10/19/13 14:52:12 to 10/19/13 14:52:12
 Modified changed from 01/27/14 18:15:37 to 01/27/14 18:15:37
comment:10 Changed 7 years ago by
 Commit set to 18528939b11702517e6240eaa15ffad5052116c2
Branch pushed to git repo; I updated commit sha1. New commits:
1852893  Larger example.

comment:11 followup: ↓ 15 Changed 7 years ago by
 Status changed from needs_info to needs_review
I've updated the implementation of partial_fraction_decomposition to do the full decomposition into prime powers. I also uncovered and fixed a bug in the case of nonunqiue remainders satisfying the Euclidean function inequality (e.g. ZZ).
comment:12 Changed 7 years ago by
 Commit changed from 18528939b11702517e6240eaa15ffad5052116c2 to 6c18589378d45119e4ab2545f0cda7ea92e9d297
comment:13 Changed 7 years ago by
 Milestone changed from sage6.1 to sage6.2
comment:14 Changed 7 years ago by
 Branch changed from u/robertwb/ticket/15306 to u/mmezzarobba/15306partial_fraction
 Commit changed from 6c18589378d45119e4ab2545f0cda7ea92e9d297 to 5f4a8579c7453dc3c2de8ecae9b455ab36c47f64
New commits:
5f4a857  #15306 rev. patch: simplifications, more tests

comment:15 in reply to: ↑ 11 Changed 7 years ago by
Replying to robertwb:
I've updated the implementation of partial_fraction_decomposition to do the full decomposition into prime powers. I also uncovered and fixed a bug in the case of nonunqiue remainders satisfying the Euclidean function inequality (e.g. ZZ).
Positive review to your changes. (They do not answer my metaphysical questions, but they do fix the bug.) I prepared a small reviewer patch, though: can you have a quick look? Thanks.
comment:16 Changed 7 years ago by
 Status changed from needs_review to positive_review
Your patch looks fine.
comment:17 Changed 7 years ago by
 Reviewers set to Marc Mezzarobba, Robert Bradshaw
comment:18 Changed 7 years ago by
 Branch changed from u/mmezzarobba/15306partial_fraction to 5f4a8579c7453dc3c2de8ecae9b455ab36c47f64
 Resolution set to fixed
 Status changed from positive_review to closed
Can you please elaborate on what output is expected? The output is correct in ZZ, and I don't see it different in QQ.