Re: How to use Save/Get for interpolation of unstructured
- To: mathgroup at smc.vnet.net
- Subject: [mg125445] Re: How to use Save/Get for interpolation of unstructured
- From: Ted Sariyski <tsariysk at craft-tech.com>
- Date: Wed, 14 Mar 2012 00:38:39 -0500 (EST)
- Delivered-to: l-mathgroup@mail-archive0.wolfram.com
- References: <jj7dri$2ds$1@smc.vnet.net> <201203110906.EAA19348@smc.vnet.net> <jjkehc$1bh$1@smc.vnet.net> <201203130803.DAA12761@smc.vnet.net>
On 3/13/2012 4:03 AM, David Bailey wrote:
> On 12/03/2012 09:10, Ted Sariyski wrote:
>> On 3/11/2012 5:06 AM, David Bailey wrote:
>>> On 07/03/2012 10:38, Ted Sariyski wrote:
>>>> Hi,
>>>>
>>>> I posted this question a couple of days ago but did not get any
>>>> response. Meanwhile I noticed that the Save-d interpolation data contains
>>>> "NDSolve`FEM`ElementMesh", which may relates to my problem. Let me
>>>> restate the question.
>>>>
>>>> I have 2D unstructured data in the form {{{1.22001, 0.00024605},
>>>> 0.816237}, {{1.22334, 0.000250332},0.436219},... } (connectivity is also
>>>> available but I do not know how to use it with Mathematica).
>>>>
>>>> Interpolation issues a warning:
>>>>
>>>> "Interpolation::udeg: Interpolation on unstructured grids is currently
>>>> only supported for InterpolationOrder->1 or InterpolationOrder->All.
>>>> Order will be reduced to 1.>>",
>>>>
>>>> The warning is not an issue for me, the problem is that I cannot use
>>>> Save/Get for this interpolation function. After Get the interpolation
>>>> function instead of a number returns:
>>>>
>>>> intp[1., .8]
>>>> InterpolatingFunction[{{0.000905256237704326, 1.9993734385285347}, {0.00024605021944692736, 0.9989678166248219}},
>>>> {4, 4225, 0, {10880, 2}, {2, 2}, 0, 0, 0, 0, Automatic},
>>>> {NDSolve`FEM`ElementMesh[{{1.220009724772626, 0.00024605021944692736},
>>>> {1.223336731504134, 0.0002503318707254048}, {1.2273720422942693, ...
>>>>
>>>> which is the non-evaluated content of the Save-d file.
>>>>
>>>> What is the right way to use Save/Get for interpolation of unstructured
>>>> data?
>>>>
>>>> Any help will be highly appreciated.
>>>> Thanks,
>>>> --Ted
>> data = Table[{{Random[], Random[]}, Random[]}, {i, 1, 10}];
>> intp = Interpolation[data];
>> intp[0.1, 0.2]
>> Out[]: 1.110174
>>
>> Save["intp.mat", intp];
>> Quit[]
>>
>> Get["intp.mat"];
>> intp[0.1, 0.7]
>> Out[]: intp[0.1, 0.7]
>> InterpolatingFunction[{{0.141559,0.972136},{0.0916656,0.927415}},{4,4225,0,{10,2},{2,2},0,0,0,0,Automatic},{NDSolve`FEM`ElementMesh[{{0.141559,0.972136},{0.0916656,0.927415}},{NDSolve`FEM`TriangleElement[<12>]}]},{0.106611,0.195957,0.105951,0.417478,0.901584,0.324786,0.696795,0.424483,0.192741,0.949504},{Automatic}][0.1,0.7]
>>
> I looked at the file intp.mat, and that appears to be syntactically
> valid, and if you split the above into separate cells, you can see that
> the syntax error does not come from the Get command. However, for some
> reason, the term starting NDSolve`FEM`ElementMesh..... seems to print
> out in skeletal form with <13> - which is obviously invalid. I presume
> you cut and pasted that back to create new input - hence the skeleton
> expression will not parse and gives a syntax error.
>
> I would use the InterpolatingFunction object in the kernel without
> copy/pasting it.
>
> David Bailey
> http://www.dbaileyconsultancy.co.uk
>
> I would use the InterpolatingFunction object in the kernel without
> copy/pasting it.
The reason I want to use Save/Get is that Interpolation takes a couple of minutes per variable. If using "the InterpolatingFunction object in the kernel without copy/pasting" will bypass re-Interpolation at every restart, I am very interested. Could you explain or point me to an example how to do it, please?
- References:
- Re: How to use Save/Get for interpolation of unstructured data?
- From: David Bailey <dave@removedbailey.co.uk>
- Re: How to use Save/Get for interpolation of unstructured
- From: David Bailey <dave@removedbailey.co.uk>
- Re: How to use Save/Get for interpolation of unstructured data?