[nlinder]

Hi!

I ran an interFoam case with the velocity BC

        type            uniformFixedValue;
        uniformValue    table
        (
            (0.0 (0.0 0.0 0.0))
            (1 (1.0 0.0 0.0))
            (3 (4.0 0 0))
        );

Results are as expected. Opening the case, paraview complains

Error reading line 55 of /case/0/U: Expected a number, found (

which is the line of the first entry of the table. I read, opening it with paraFoam should resolve that issue. However, it does not for me.

Is it possible to either substitute each occurrence of that BC by a fixedValue easily (lots of timesteps)? Or does anyone know another solution? I use OF 2.2.x

Thanks in advance!
Nicklas

[nlinder]

managed to do it with sed:

find ./ -name U -type f -exec sed -i "/\(uniformFixedValue\)/{n;N;N;N;N;N;N;N;N;d}" {} \;
find ./ -name U -type f -exec sed -i "s/uniformFixedValue/fixedValue;value uniform (0 0 0)/g" {} \;

and yes - you might can do that in one command. But if anyone has an easier solution I'd be interested.

Greetings

[wyldckat]

Greetings Nicklas,

I'll quote myself from almost a year... er, 7 months ago:

Originally Posted by wyldckat

The internal ".foam" reader is allergic to the boundary condition "uniformFixedValue". Therefore, before running paraview, you'll need to run this command:

sed -i -e "s=uniformFixedValue=fixedValue=g"  [0-9][0-9]*/{U,p,k,nut,epsilon}

It will replace all occurrences of "uniformFixedValue" with "fixedValue", for all of the essential field files in the time folders after "0".

The problem is that you should not use these time folders for continuing the simulation afterwards, since this hack could damage the simulation characteristics for continuing simulating. But it should not damage the results for post-processing.

edit: By the way, the same happens in ParaView 3.12.0.

Best regards,
Bruno

[yvesc]

thanks to you both. I have been trying to experiment with time-varying BC, and I have had the exact same problem; I am reassured that your analysis is the same as mine, it's a "bug" where openFoam files just recopy the uniformFixedValue in its entirety, and paraFoam is unable to read that format. Would be smoother and more logical if each time-file had a fixed value entry with the correct value for that time.

1) should this be reported as a bug? Doesn't just everybody meet this problem and lose time over it?

2) Bruno, your one-liner

 sed -i -e "s=uniformFixedValue=fixedValue=g"  [0-9][0-9]*/{U,p,k,nut,epsilon}

doesn't work for me:
a - "-i -e" : that has the result of generating "U-e" etc backup files -- maybe that's my version of sed, I'm on Mac OsX
b- "changing "uniformFixedValue" to "fixedValue" is not enough, you have to erase the following lines too (in my case, an inline table), otherwise paraFoam chokes.

I guess this is a job for awk, I will have to brush up on the subject.

[wyldckat]

Greetings Yves,

The problem is with ParaView's internal ".foam" reader. It is not a problem with OpenFOAM's own ".OpenFOAM" reader plug-in for ParaView.

Personally I haven't reported this to the ParaView or VTK bug tracker, as there are several more bugs that need fixing in the internal reader, which may or may not have already been fixed in the latest "vtkPOFFReader": http://openfoamwiki.net/index.php/Co...r_for_ParaView

Best regards,
Bruno

[frantov]

Hello Yvesc
Did you manage to fix this with awk or something else?
Im experiencing a similar issue and I need to get rid of three tables (three outlets)
regards
Fran

[yvesc]

Hi francisco.
No, I just gave my personal problem a quick SED treatment.

#!/bin/bash
LC_ALL=C sed -i "" -e "/uniformFixedValue/{n;N;N;N;N;N;N;N;N;d;}" [0-9][0-9]*/Solid2/T
LC_ALL=C sed -i "" -e "s/uniformFixedValue/fixedValue; value uniform 300/g" [0-9][0-9]*/Solid2/T

in my case, the files to modify were T in directories xxxxxx/Solid2 except of course directory 0, hence the first [0-9] in the path.
The first line deletes the lines after uniformFixedValue; the number of N's had to be adapted to the number of lines to be deleted *in my case*.
The second line replaces the remaining "uniformFixedValue" by "fixed value".
The "" after -i was necessary to remove the problem I pointed out in my previous post.

LC_ALL=C was necessary to avoid an "illegal byte sequence" error, my files probably are not UTF8, didn't check.

[lithos]

Originally Posted by yvesc

maybe that's my version of sed, I'm on Mac OsX

Hi,

yes, this is exactly the problem. I would suggest you to install gsed via MacPorts, Fink, or any other source you might want to use. Standard 'sed' by Apple is a weird, limited edition. I had already several problems using linux scripts on Mac due to this. Hope it helps.

Cheers, Michael

[Rodrigo Correa]

Dears,

I'm also facing that problem with Paraview.
I'd like to know what specific thing in the following command posted by Bruno:

sed -i -e "s=uniformFixedValue=fixedValue=g" [0-9][0-9]*/{U,p,k,nut,epsilon}

what does [0-9] [0-9] means?

Best Regards,

Rodrigo Correa

[wyldckat]

Originally Posted by Rodrigo Correa

what does [0-9] [0-9] means?

Quick answer:

• http://openfoamwiki.net/index.php/In...with_the_Shell -->
  • http://linuxcommand.org/learning_the_shell.php -->
    • http://linuxcommand.org/lts0050.php#wildcards

[fabian_roesler]

Hi,

I gave the uniformFixedValue-with-paraview-problem a bug treatment.
After a short discussion, Henry Weller made the behaviour of uniformFixedValueFvPatchField consistent between OpenFOAM-2.4.x and the latest version in OpenFOAM-dev.

0001867: uniformFixedValue BC not read by praview
http://www.openfoam.org/mantisbt/view.php?id=1867

Cheers

Fabian

[wyldckat]

Hi Fabian,

Many thanks for reporting this! I kept forgetting about looking into this

Best regards,
Bruno