[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [h5md-user] box and observables

From: Felix Höfling
Subject: Re: [h5md-user] box and observables
Date: Mon, 23 Sep 2013 18:15:47 +0200
User-agent: Opera Mail/12.15 (Linux)

Am 23.09.2013, 17:51 Uhr, schrieb Peter Colberg <address@hidden>:

On Mon, Sep 23, 2013 at 10:27:28AM +0200, Felix Höfling wrote:
From the point of view of a single subgroup, this is
easy, the time grids are congruent, and the information is complete.

This is not the case.

The data within a subgroup can be sampled at different time grids.

As noted again and again (indicating that the specification is not
explicit enough as to the use of "time" and "step" datasets), a reader
must lookup the "value" index corresponding to a time/step in the
"time" or "step" dataset, for any value dataset.

As an optimization, a reader can check if two time/step datasets point
to the same HDF5 object (i.e. they are hard-linked); then, and only
then, can the time/step information be reused for multiple value

(Since I am mentioning this so often, it would be my task to point
this out in the specification…)


You are right, this applies for the position data only. I was referring to the following statement in the box spec: "A specific requirement for box groups inside particles is that the step and time datasets exactly match those of the corresponding position groups".

For all other data fields (e.g., velocity), one has to check whether the step dataset coincide (or start searching for the corresponding step). Our motivation to restrict the match to the position data only was that the box data are mainly needed to wrap/unwrap the positions.


reply via email to

[Prev in Thread] Current Thread [Next in Thread]