[Paraview] PV 4.1.0: save the data saves only 1/2 of a circle

Stephen Wornom stephen.wornom at inria.fr
Wed Mar 11 12:16:52 EDT 2015



----- Original Message -----
> From: "Kenneth Moreland" <kmorel at sandia.gov>
> To: "Stephen Wornom" <stephen.wornom at inria.fr>
> Cc: "ParaView list" <paraview at paraview.org>
> Sent: Wednesday, March 11, 2015 4:55:06 PM
> Subject: Re: [Paraview] PV 4.1.0:  save the data saves only 1/2 of a circle
> 
> Stephen,
> 
> I got your data and tried again. Once again I cannot replicate your error.
> When I open up the csv file in Excel, I get the number of rows I expect
> and the bounds of Y are what I expect. I tried both on a serial process
> and using a parallel job (in case there was a problem with writing CSV in
> parallel, which there has been in the past but has since been fixed).
> 
> When you sent the file to me, you mentioned some further instructions I
> didn't understand about modifying the csv file and loading it indirectly
> through an inp file. I don't know anything about the inp format (and there
> appear to be three readers of different formats that expect that
> extension), but perhaps the real problem is in the inp reader? Did you try
> looking at the csv file directly in a text editor or spreadsheet program?




I will look at the inp file and see if how I read the csv file (recall that I replace the commas with blanks) in fortran could be a problem.

I forgot to mention that I am running on my linux workstation the 64-bit binary versions of PV 4.1.0 and 4.3.1 that I downloaded from paraview web site. Are you using the binary or an installation from source on your workstation?.

It is strange that  for a single cylinder I get the correct number of values for the upper and lower parts of the cylinder.

The problem occurs when there two cylinders. I will look at the excel files and if it is correct, maybe I can save it in text format that I can read into gnuplot.
I have no experience with excel nor reading csv in gnuplot scripts.
Will keep you posted.
Stephen
> 
> -Ken
> 
> 
> 
> On 3/11/15, 6:57 AM, "Stephen Wornom" <stephen.wornom at inria.fr> wrote:
> 
> >Moreland, Kenneth wrote:
> >> Is there any way you can share your data? That might help diagnose the
> >> problem.
> >Yes, I will send it. I did an additional test with a single cylinder and
> >the file is written correctly. So it appears that the problem occurs
> >only with two tandem cylinders in the mesh.
> >Stephen
> >>
> >>
> >> On 3/10/15, 11:01 AM, "Stephen Wornom" <stephen.wornom at inria.fr> wrote:
> >>
> >>> I tried both PV 4.10 and PV 4.31 and get the same result.
> >>> - extract surface
> >>> - connectivity
> >>> - threshold (defines the surface of the cylinder)
> >>>
> >>> information shows ymax= 0.5, ymin=-0.5 which is correct as the radius
> >>>of
> >>> the circle is 0.5.
> >>> When I save data, the csv files has all of the y< 0 values. For y>0
> >>>most
> >>> values are missing.
> >>> For the csv file, ymin= 0.5 but ymax = 0.24.
> >>>
> >>> There are actually two circles, could this pose a problem?
> >>>
> >>> I really need to solve this problem of missing values.
> >>> Give me ideas to try, alternative ways to write the data (how do I
> >>>write
> >>> only the values for the 1st cylinder to see if this is the problem?).
> >>> Thanks for your help,
> >>> Stephen
> >>>
> >>>
> >>> ----- Original Message -----
> >>>> From: "Kenneth Moreland" <kmorel at sandia.gov>
> >>>> To: "Stephen Wornom" <stephen.wornom at inria.fr>, "ParaView list"
> >>>> <paraview at paraview.org>
> >>>> Sent: Monday, March 9, 2015 6:28:45 PM
> >>>> Subject: Re: [Paraview] PV 4.1.0:  save the data saves only 1/2 of a
> >>>> circle
> >>>>
> >>>> Stephen,
> >>>>
> >>>> The CSV writer should write out all the points. I tried repeating your
> >>>> instructions with ParaView 4.3.1 using both the Cylinder source and
> >>>>the
> >>>> disk_out_ref dataset to write out data for a hollow cylinder source
> >>>>to a
> >>>> csv file and it looks like it works fine. I can only think of the most
> >>>> basic things to try:
> >>>>
> >>>> * Are you sure your cylinder is centered in the x-y plane?  Maybe the
> >>>> circle is actually sitting below the y axis. Check that with either
> >>>>the
> >>>> cube axes or the information panel.
> >>>> * Is the connectivity/threshold filters really the best way to get a
> >>>> circle from a cylinder. You might consider instead to use the slice
> >>>> filter
> >>>> to get the circle. (I don't know why that would fix any I/O issues,
> >>>> though.)
> >>>> * Is there any chance you can try using the latest version of
> >>>>ParaView.
> >>>> I
> >>>> don't recall any such problems with the CSV reader in 4.1, but you
> >>>>never
> >>>> know.
> >>>>
> >>>> -Ken
> >>>>
> >>>>
> >>>>
> >>>>
> >>>> On 3/9/15, 10:47 AM, "Stephen Wornom" <stephen.wornom at inria.fr> wrote:
> >>>>
> >>>>> I have a cylinder ( a circle in the x-y plane, z is the span
> >>>> direction).
> >>>>> I do:
> >>>>> - extract surface
> >>>>> - connectivity
> >>>>> - threshold (defines the surface of the cylinder)
> >>>>>
> >>>>> The result is the circle shown in that attached figure.
> >>>>>
> >>>>> I want to write the circle values to a file so I save the data.
> >>>>>   The problem is that  PV4.10 saves only 1/2 half of the data (y<0).
> >>>> For
> >>>>> y> 0 only 4 points are written to the csv file, 124 points for y <0.
> >>>>>
> >>>>> How can I save the data so that all values are written?
> >>>>> Hope the question is clear.
> >>>>> Stephen
> >>>>>
> >>>>> --
> >>>>> stephen.wornom at inria.fr
> >>>>> 2004 route des lucioles - BP93
> >>>>> Sophia Antipolis
> >>>>> 06902 CEDEX
> >>>>> 		
> >>>>> Tel: 04 92 38 50 54
> >>>>> Fax: 04 97 15 53 51
> >>>>>
> >>>>
> >
> >
> >--
> >stephen.wornom at inria.fr
> >2004 route des lucioles - BP93
> >Sophia Antipolis
> >06902 CEDEX
> >		
> >Tel: 04 92 38 50 54
> >Fax: 04 97 15 53 51
> >
> 
> 


More information about the ParaView mailing list