@marc005 That restriction on the Matrix brower appears to be present in 17. But it's not difficult to write custom procedures which do similar things.

For examination of the entries by value, you could just use a Data (table) Component. If the Matrix already is assigned to a name in the session, then inserting such an embedded component from the Components palette will very nicely prompt whether you wish to use it.

The attached worksheet took just a few moments to write. A little longer and it could be spiffier still.

quandlfun.mw

@marc005 The message "Unable to display structured data" arises from the Matrix browser (also available via right-click context-menu), when either dimension exceeds a certain value. It's not a consequence of that `GetCSV` procedure, per se.

The limiting value for either dimension appears to be 10001.

The Matrix browser has a few quirks. For example, the `image` view does not seem to work when either row or column dimension is just 1.

@Carl Love Yes, I get agreement (except on that one mentioned subexample). When an assumption is placed, including as additional qualification, a new assumed local is generated and assigned to the global name. It's always been problematic (a FAQ of sorts) that people get into trouble if they make assignments using some intermediate stage assumptions and then combine expressions involving the mix of assumed names. I don't know whether the word "intended" connotes to everyone, but that is the longstanding implementation and, presumably, the early design.

I find it less error prone (and easier to keep straight) to use `assuming` rather than `assume`, and this kind of thing is one reason.

@Carl Love Yes, I get agreement (except on that one mentioned subexample). When an assumption is placed, including as additional qualification, a new assumed local is generated and assigned to the global name. It's always been problematic (a FAQ of sorts) that people get into trouble if they make assignments using some intermediate stage assumptions and then combine expressions involving the mix of assumed names. I don't know whether the word "intended" connotes to everyone, but that is the longstanding implementation and, presumably, the early design.

I find it less error prone (and easier to keep straight) to use `assuming` rather than `assume`, and this kind of thing is one reason.

So your y1,...yn are floats? If so then what happens if you raise Digits in Maple? (Ie. to 15, and then to 30?) Numeric roundoff error can be different for different (equivalent) forms of an expression, even without the kind of algebraic cancellation that Carl described. Applying `simplify` will not necessarily reduce the numeric roundoff effects for subsequent floating-point substituion and evaluation.

Why use `subs` instead of `eval`, btw?

acer

So your y1,...yn are floats? If so then what happens if you raise Digits in Maple? (Ie. to 15, and then to 30?) Numeric roundoff error can be different for different (equivalent) forms of an expression, even without the kind of algebraic cancellation that Carl described. Applying `simplify` will not necessarily reduce the numeric roundoff effects for subsequent floating-point substituion and evaluation.

Why use `subs` instead of `eval`, btw?

acer

@Carl Love Using your code above I get,

  > about(x1);

Originally x, renamed x~:
  is assumed to be: RealRange(Open(1),infinity)

even after the additional qualification `x>2` is made.

I get that in Maple 11 through 17, on Windows.

Above your results show,

  > about(x1);

Originally x, renamed x~:
  is assumed to be: RealRange(Open(2),infinity)

which I don't obtain.

@Carl Love Using your code above I get,

  > about(x1);

Originally x, renamed x~:
  is assumed to be: RealRange(Open(1),infinity)

even after the additional qualification `x>2` is made.

I get that in Maple 11 through 17, on Windows.

Above your results show,

  > about(x1);

Originally x, renamed x~:
  is assumed to be: RealRange(Open(2),infinity)

which I don't obtain.

Is your acquaintance running 32bit or 64bit Maple on Windows?

acer

@Carl Love Carl, you wrote in a Comment above that, "Most interesting is that all copies end up with the same assumptions."

In which version did you obtain those results from the `about` command, may I ask?

@Carl Love Carl, you wrote in a Comment above that, "Most interesting is that all copies end up with the same assumptions."

In which version did you obtain those results from the `about` command, may I ask?

@emma hassan Well, you already have a table, as V;

eval(V);

If you want you can turn V into a Matrix (and then right-click and browse as one way to view its entries), or call plots:-surfdata on that.

matdata:=Matrix(M,N,V);

plots:-surfdata(matdata,axes=box);

@emma hassan Well, you already have a table, as V;

eval(V);

If you want you can turn V into a Matrix (and then right-click and browse as one way to view its entries), or call plots:-surfdata on that.

matdata:=Matrix(M,N,V);

plots:-surfdata(matdata,axes=box);

Are a,b,c,... each something numeric? Are F1 and F2 numeric? Are they floating-point?

acer

@Carl Love I had written an explanation of what my answer was supposed to convey, but lost it entirely when hitting the Submit button (the site game me that new message about there being a problem with the submission...). No back action available in my chrome tag. Extreme wtf moment.

Anyway, my earlier intention was to wait until finding out something more about the OP's motivation in asking why it didn't "work" before spending time fiddling with evalf[d] or ilog10 (or what have you) to accomodate the region about zero. But being unable to use the site's editor as usual made me lose interest.