Illegal pair_coeff command




Thanks for the comment.








I fixed it and ran again.



I got a another error message.



"Invalid bond type in Bonds section of data file"



I think It may be a simple error just like before, though I don’t know what it is now.



Here’s my question.



If I have a grammatical mistake in input file, this can be easily solved by checking the log file.



But how do I know if the grammatical problem is in data file?






--------- 원본 메일 ---------


> 보낸사람: Chang Woon Jang <changwoonjang@…24…>
> 받는사람 : Ray Shan <rayshan819@…24…>
> 참조 : 윤홍민 <1234dndd@…1333…>,LAMMPS Users Mailing List [email protected]
> 날짜: 2013년 6월 14일 금요일, 23시 52분 03초 +0900
> 제목: Re: [lammps-users] Illegal pair_coeff command
>
> There are pair coefficients defined in the data file, but no space between the word “Pair Coeffs” and the numbers in the data file.
>
> C. W. Jang







Thanks for the comment.

I fixed it and ran again.

I got a another error message.

"Invalid bond type in Bonds section of data file"

I think It may be a simple error just like before, though I don't know what it is now.

Here's my question.

If I have a grammatical mistake in input file, this can be easily solved by checking the log file.

But how do I know if the grammatical problem is in data file?

currently, the only way is to read the data file and compare it with
the documentation of the format in the read_data section of the lammps
manual.

axel.

The read_data command throws many kinds of errors
if your data file in incorrectly formatted. The error
message you got is clear on where the problem is.
Look at the bonds section and you will find an invalid
bond type.

Steve

The read_data command throws many kinds of errors
if your data file in incorrectly formatted. The error
message you got is clear on where the problem is.

Look at the bonds section and you will find an invalid
bond type.

steve,

i somewhat disagree on this one. you may be right, but having debugged
a parser written in Tcl for data files and having seen a lot of error
messages for data files being thrown at the most unexpected places, it
is equally possible (or even more likely) that the _real_ error is
located somewhere else, e.g. missing mandatory empty lines, problems
with incorrect text more transfers between operating systems with
different end-of-line conventions, corrupted files due to errors in
the data file generating code. or simply just not providing the
correct atom style in the input file.

the latter is a problem that could be somewhat contained, if the data
file had a line (optional?) in the header section that would specify
which atom style the data file was generated for and then this can be
compared against the atom style choice in the input file and then lead
to a warning (e.g. about when this field is not provided) or an error
(when an incompatible atom style is signaled).

the, IMO better, alternative would be to support a more
self-descriptive file format (with the option to encode it in hdf5,
please for handling very large systems). this parser could assign
symbolic values to per atom/bond/angle/whatever properties and thus it
would be sufficient, if the data file has a superset of the data
required by the choice of atom style (and potentials).

axel.

this might be a good topic for discussion at the developer
session of the LAMMPS workshop. I agree that the data
file reader can get confused (by bad data files) in many
different ways. I think if the Atoms section is bad, due
to a mis-matched atom style, you will get a quick error,
e.g. if each atom line has the wrong # of values.

Steve