The OPAL Discussion Forum

[Andreas Adelmann <andreas.adelmann AT psi.ch>]

I changed this to a StepAttribute that should solve the problem and

was also my original intention.

AA

On Apr 27, 2012, at 3:16 PM, Achim Gsell wrote:

Hi,

I try to fix the ParallelPlateFurman1 test and get the following error:

ParallelTTracker {+0}> 14.20.41 Step      0 at    2.500 [mm] t=  2.000e-12 [s] E=    0.003 [keV]
[proc 0] E: H5WriteFileAttribString: Cannot create attribute 'OPAL_flavour' for '/'.
[proc 1] E: H5WriteFileAttribString: Cannot create attribute 'OPAL_flavour' for '/'.
[proc 2] E: H5WriteFileAttribString: Cannot create attribute 'OPAL_flavour' for '/'.
[proc 3] E: H5WriteFileAttribString: Cannot create attribute 'OPAL_flavour' for '/'.

The problem is that the attribute 'OPAL_flavour' is already set and H5hut doesn't allow to overwrite the value. This is one of the ancient decisions former H5Part developers made. Unfortunately there is no API call get query the existence of an attribute. To solve the problem we can

1) implement a H5hut function H5GetFileAttribExists() and adopt the OPAL code accordingly

2) or change the H5WriteXxxAtribXxx() calls.

I prefer the second. Usually data in a file opened for writing can be overwritten. In append mode data can only be appended but not be overwritten. Applied to H5hut means, it is possible to add new time-steps, add new attributes, add new datasets/field-data to steps etc. but you cannot overwrite an existing attribute, dataset etc.

Comments, suggestions?

Achim


_______________________________________________
Opal mailing list
Opal AT lists.psi.ch
https://lists.web.psi.ch/mailman/listinfo/opal

------
Dr. sc. math. Andreas (Andy) Adelmann
Staff Scientist
Paul Scherrer Institut WBBA/311 CH-5232 Villigen PSI
Phone Office: xx41 56 310 42 33 Fax: xx41 56 310 31 91
Phone Home: xx41 62 891 91 44
-------------------------------------------------------
Friday: ETH CAB H 8x.x  +41 44 632 xx xx
============================================
The more exotic, the more abstract the knowledge, 
the more profound will be its consequences.
Leon Lederman 
============================================