write_data command


write_data file keyword value ...
  • file = name of data file to write out

  • zero or more keyword/value pairs may be appended

  • keyword = pair or nocoeff or nofix or nolabelmap

    nocoeff = do not write out force field info
    nofix = do not write out extra sections read by fixes
    nolabelmap = do not write out type labels
    types value = numeric or labels
    pair value = ii or ij
      ii = write one line of pair coefficient info per atom type
      ij = write one line of pair coefficient info per IJ atom type pair


write_data data.polymer
write_data data.*


Write a data file in text format of the current state of the simulation. Data files can be read by the read data command to begin a simulation. The read_data command also describes their format.

Similar to dump files, the data filename can contain a “*” wild-card character. The “*” is replaced with the current timestep value.

Data in Coeff sections

The write_data command may not always write all coefficient settings to the corresponding Coeff sections of the data file. This can have one of multiple reasons. 1) A few styles may be missing the code that would write those sections (if you come across one, please notify the LAMMPS developers). 2) Some pair styles require a single pair_coeff statement and those are not compatible with data files. 3) The default for write_data is to write a PairCoeff section, which has only entries for atom types i == j. The remaining coefficients would be inferred through the currently selected mixing rule. If there has been a pair_coeff command with i != j, this setting would be lost. LAMMPS will detect this and print a warning message unless pair ij is appended to the write_data command. This will request writing a PairIJCoeff section which has information for all pairs of atom types. In cases where the coefficient data in the data file is incomplete, you will need to re-specify that information in your input script that reads the data file.

Because a data file is in text format, if you use a data file written out by this command to restart a simulation, the initial state of the new run will be slightly different than the final state of the old run (when the file was written) which was represented internally by LAMMPS in binary format. A new simulation which reads the data file will thus typically diverge from a simulation that continued in the original input script.

If you want to do more exact restarts, using binary files, see the restart, write_restart, and read_restart commands. You can also convert binary restart files to text data files, after a simulation has run, using the -r command-line switch.


Only limited information about a simulation is stored in a data file. For example, no information about atom groups and fixes are stored. Binary restart files store more information.

Bond interactions (angle, etc) that have been turned off by the fix shake or delete_bonds command will be written to a data file as if they are turned on. This means they will need to be turned off again in a new run after the data file is read.

Bonds that are broken (e.g. by a bond-breaking potential) are not written to the data file. Thus these bonds will not exist when the data file is read.

Use of the nocoeff keyword means no force field parameters are written to the data file. This can be helpful, for example, if you want to make significant changes to the force field or if the force field parameters are read in separately, e.g. from an include file.

Use of the nofix keyword means no extra sections read by fixes are written to the data file (see the fix option of the read_data command for details). For example, this option excludes sections for user-created per-atom properties from fix property/atom.

The nolabelmap and types keywords refer to type labels that may be defined for numeric atom types, bond types, angle types, etc. The label map can be defined in two ways, either by the labelmap command or in data files read by the read_data command which have sections for Atom Type Labels, Bond Type Labels, Angle Type Labels, etc. See the Howto type labels doc page for the allowed syntax of type labels and a general discussion of how type labels can be used.

Use of the nolabelmap keyword means that even if type labels exist for a given type-kind (Atoms, Bonds, Angles, etc.), type labels are not written to the data file. By default, they are written if they exist. A type label must be defined for every numeric type (within a given type-kind) to be written to the data file.

The types keyword determines how atom types, bond types, angle types, etc are written into these data file sections: Atoms, Bonds, Angles, etc. The default is the numeric setting, even if type label maps exist. If the labels setting is used, type labels will be written to the data file, if the corresponding label map exists. Note that when using types labels, the nolabelmap keyword cannot be used.

The pair keyword lets you specify in what format the pair coefficient information is written into the data file. If the value is specified as ii, then one line per atom type is written, to specify the coefficients for each of the I=J interactions. This means that no cross-interactions for I != J will be specified in the data file and the pair style will apply its mixing rule, as documented on individual pair_style doc pages. Of course this behavior can be overridden in the input script after reading the data file, by specifying additional pair_coeff commands for any desired I,J pairs.

If the value is specified as ij, then one line of coefficients is written for all I,J pairs where I <= J. These coefficients will include any specific settings made in the input script up to that point. The presence of these I != J coefficients in the data file will effectively turn off the default mixing rule for the pair style. Again, the coefficient values in the data file can be overridden in the input script after reading the data file, by specifying additional pair_coeff commands for any desired I,J pairs.


This command requires inter-processor communication to migrate atoms before the data file is written. This means that your system must be ready to perform a simulation before using this command (force fields setup, atom masses initialized, etc).


The option defaults are pair = ii and types_style = numeric.