Timeline for Problem with SeriesData
Current License: CC BY-SA 3.0
Post Revisions
13 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Mar 14, 2018 at 6:36 | vote | accept | SonerAlbayrak | ||
| Mar 10, 2018 at 20:31 | comment | added | SonerAlbayrak | Thank you Daniel. | |
| Mar 10, 2018 at 16:29 | comment | added | Daniel Lichtblau |
It is sort of implied by Series ref guide page bullet item "Series can construct standard Taylor series, as well as certain expansions involving negative powers, fractional powers, and logarithms.". Yes, there are some differences in what Series does vs. operations on an explicit SeriesData object. But the limitations for what one can expect to be correctly handled are as above. I should also note that SeriesData, being an internal form of a series, is not likely to be as fully documented as the `Series function.
|
|
| Mar 10, 2018 at 2:29 | comment | added | SonerAlbayrak | Thank you very much Daniel. I may be missing something but I still think the issue is not with Series but with SeriesData. In your example you take symbolic power of output of Series which is itself SeriesData. What I was trying to say is that Series does not treat its input as algebraic (as can be seen in the example I gave in my comment above), whereas SeriesData apparently does, but I did not consider that this was a feature. I could not find in the documentation regarding the restrictions on the realm of SeriesData, do you happen to know any explanation somewhere? | |
| Mar 9, 2018 at 22:32 | comment | added | Daniel Lichtblau |
SeriesData is using upcode in the Series world. You will see the same issue with InputForm[Series[-x^2,{x,0,2}]^a] Out[7]//InputForm= (-x)^(2*a)*SeriesData[x, 0, {1}, 0, 1, 1]. It's basically doing what it is supposed to do, but operating outside the realm of algebraic functions. Getting it to behave inside that realm, e.g. with rational powers, has been dicey already.
|
|
| Mar 9, 2018 at 18:06 | comment | added | SonerAlbayrak | I actually do not have problem with Series: It indeed does not calculate what I want however it does not give me incorrect results either. SeriesData on the other hand does give me incorrect result, that is why I think it is a bug. As an example, Series[(-x (x - a))^b, {x, Infinity, 0}] does not evaluate but replacing $x$ with $x+O[1/x]^0$ gives incorrect answer, whereas it should be Mathematically equivalent. Please see Carl Woll's answer here: mathematica.stackexchange.com/questions/80107/… | |
| Mar 9, 2018 at 15:53 | comment | added | Daniel Lichtblau |
Series does not do well with symbolic powers. It's not a bug per se, its a limitation since this falls outside the scope of what it will handle.
|
|
| Mar 9, 2018 at 8:38 | answer | added | ha loa | timeline score: 3 | |
| Mar 9, 2018 at 8:28 | comment | added | SonerAlbayrak | My apologies, thanks for pointing that out. | |
| Mar 9, 2018 at 7:58 | comment | added | J. M.'s missing motivation | Please do not use the bugs tag until other users have confirmed your observations. | |
| Mar 9, 2018 at 7:57 | history | edited | J. M.'s missing motivation |
edited tags
|
|
| Mar 9, 2018 at 7:34 | review | First posts | |||
| Mar 9, 2018 at 9:48 | |||||
| Mar 9, 2018 at 7:31 | history | asked | SonerAlbayrak | CC BY-SA 3.0 |