Odd. But ParametricPlot[{f[t], g[t]}, {t, 0,1}] works, yes?

Nope I get a blank plot

Then something went wrong.

ctrlpts is a list of coordinates, yes?

I am trying.. I have been trying other stuff.. so I need couple of minutes to get your suggestion working

test it's list of coordinates

All numbers?

ya

btw, could please explain to me how this snippet works : (Dot @@ #[[{1, -1}]]) & /@ Partition[pts, n, n - 1] ? I know you have used pure function but I do not understand how this is working

Graphics[BSplineCurve[ctrlpts, SplineDegree -> m, SplineKnots -> knots]] shows a curve, doesn't it?

(One problem at a time, please. :) )

ok..let me try :)

I get the curve back

Add Point[pts] (if pts is the name of your original set of points) and check that the curve passes through all of the points.

where should I add Point[data1]?

Ah… Graphics[BSplineCurve[ctrlpts, SplineDegree -> m, SplineKnots -> knots], Epilog -> {{Blue, Line[ctrlpts]}, {AbsolutePointSize[4], Point[data1]}}]

Ya the curve passes through the points

Can you post a picture of that here?

I have been trying to do that. Is it allowed?

supposedly. otherwise, upload it to imgur.com and link to it here.

imgur.com/254hxmv

Okay, look here again.

Execute the ff. in sequence:

{xu, yu} = Transpose[ctrlpts];

f[t_] = xu.Table[BSplineBasis[{3, knots}, i - 1, t], {i, Length[ctrlpts]}];

g[t_] = yu.Table[BSplineBasis[{3, knots}, i - 1, t], {i, Length[ctrlpts]}];

It worked .. I think I forgot to clear something

I was just working on tons of other stuff and it polluted this work

So you can compute the arclength from those two now?

yep

Okay, I'll be away for now; I must have my lunch.

Thanks a lot before you go could please explain to me how this snippet works : (Dot @@ #[[{1, -1}]]) & /@ Partition[pts, n, n - 1] ? I know you have used pure function but I do not understand how this is working

@Guesswhoitis. Enjoy your lunch.. I need to crash now. I will work on it tomorrow and I will bother you if I get lost :) You have a lot to explain :) Thanks a lot again.

It's surprisingly difficult to answer questions about Mathematica when you do not have a computer. Who knew? :D

This is going to get closed it seems. But there is a question within the question that is unanswered: Why doesn't dataset[All, {"c" -> Function[x,x]}] work? Is it a bug?

What happens if you replace Function[x, x] with Identity?

@Pickett It works for me. What do you get?

@Pickett Why does dataset[All, {"a"->Sqrt}] give floating point results instead of symbolic ones?

Or even Dataset@{1, Sqrt[2], Sqrt[3]}.

No, this is just a display issue, the underlying data is preserved.

Why again are people supposed to be using Dataset instead of lists?

@Guesswhoitis. I'd put easily working with column names at the top of the convenience list for Dataset, but currently the performance hit is pretty extreme. As a funny aside, I tried to export a Dataset as an XLSX and it just prints "PATCH!".

I think a list of asssociations can give you most of the column name support. So it's more for simplified query syntax that is supposed to push people over to dataset I think.

So, it's Mathematica's attempt at something like a DB, except slower. :D

(I swear, the logo for the next version should be a spikey with kitchen sinks on all its faces.)

@Szabolcs @Guesswhoitis. Sorry for taking up your time. I looked at this for a good while with the OP's function yesterday, and didn't notice that he had a "#c" in there, not "#". It works as expected...

@Guesswhoitis. Well, database languages don't have nearly as many functions built-in as WL does. So I guess it is for people who want database like functionality with access to the WL API.

I appreciate the effort they're taking, but I'm sure they also know that for DB users a query should be far faster than an eye blink. :)

After looking through docs and examples on main, it looks like a nice toy, but still a toy.

@Guesswhoitis. It also competes with things such as "dataframes" from R. Anyone coming from R is immediately going to ask, "Where's my dataframe?! Even Python has it!" It's e.g. possible to reference columns by name

But I'm not experienced with these.

Well, with R you expect a large fraction of its users to be categorizing, so they certainly want labeled columns. :)

But then you'd hope that you can access a list as data[[All, "number of kitties"]] or something.

@Guesswhoitis. What's the time over there?

1 AM here, I'm the only one awake at where I'm staying. :D

@Guesswhoitis. I'm studying R right now, and I really miss something similar to Dataframe in Mathematica. Handler columns by name is a "must have" that you can't do in a native way in Mathematica yet. Dataset is taking to long to do basic stuffs, like work with simple array off data, that is all DB people needs to start something useful.

@Guesswhoitis. I have overload some functions, and created my way to do this in Mathematica, but it's not a acceptable way for a new user, or for mass use of MMA for DataScience. It's like to say: "off course Mathematica can do your data analysis, you just have to create your own tools from scratch". :p

Hey, I've been doing "making tools from scratch" in Mathematica for years! :D

(…and then, both amazed and disappointed when it becomes built-in in the next version.)

@Murta It seems like serious stuff is only a side effect now. The main goal of WRI is, shiny stuff for mass audience. I don't blame developers, human resources seem to be distributed in such way it is how it is.

@Kuba I agree with you. WR has really amazing developers, but sometimes forced to develop in wrong directions...

The height of this ridiculousness for me is the tweet a code thing.

@Guesswhoitis. I don't mind but I had serious doubts about Ebola outbreak modeling.

@Guesswhoitis. @Kuba For me, the biggest proof that WL is not concerned about developer community, is the lack of Workbench for V10, launched for almost one year. Or the lack of a minimal debugger in the notebook interface.

@Kuba, huh, epidemiology. Where did they do this?

@Murta prerelease is launched but I agree. Well, at least we can release our frustration here :)

@Murta, hah I can imagine. :) (even if I don't have a computer)

@Guesswhoitis. blog.wolfram.com/2014/11/04/… It's not bad topic but I have mixed feellings.

Ok, I have to go for an hour or so, see you guys.

@Kuba bye!

At least it was only slightly more sophisticated than SIR. :)