Tables and plotting

Variables can be printed in Table form using the command


The values of each variable are printed in one column, so all variables used must be defined for the same range, and corresponding elements should belong together. For example, if in a calculation one has stored R(i), THETA(i), ECI(i) for each geometry i, one can print these data simply using


By default, the number of rows equals the number of elements of the first variable. This can be changed, however, using the RANGE subcommand.

The first ten columns of a table may contain string variables. For instance,


prints a table with the SCF, CCSD, and QCI results in the first, second, and third row, respectively. For other use of string variables and tables see the examples in Tables and Macro definitions using string variables.

The apparence of the table may be modified using the following commands, which may be given (in any order) directly after the the TABLE card:

  • HEADING,head1, head2,… Specify a heading for each column. By default, the names of the variables are used as headings.
  • FORMAT,format Specify a format for each row in fortran style. format must be enclosed by quotes. Normally, the program determines automatically an appropriate format, which depends on the type and size of the printed data.
  • FTYP,typ1, typ2, typ3, … Simplified form to modify the format. This gives the type (A, F, or D) for each column (sensible defaults are normally used).
  • DIGITS,dig1, dig2, dig3, … Give the number of digits after the decimal points to be printed for each column (sensible defaults are normally used).
  • TYPE Specify a data format for the table. The default is TEXT which gives a plain text file. Other possibilities are CSV (comma-separated fields suitable for a spreadsheet), LATEX (a LaTeX table environment), MATHEMATICA (Mathematica code that assigns the table to an array), MATLAB (Matlab code that assigns the table to an array), MAPLE (Maple code that assigns the table to an array), HTML (an HTML construction), PYTHON (a Python script that contains the data and generates a 2-dimensional plot using matplotlib), and XML (an XML document containing a tree representing the table. The actual format is XHTML).
  • SAVE,file,status Specify a file on which the table will be written. If status is NEW, the file is rewound, otherwise it is appended. In the case of Python format, unless status is NEW, any edits that have been made to a pre-existing file will be preserved, and only the table data will be replaced. If file has a suffix that is one of txt, csv, tex, m, mpl, py, html, xml, and a TYPE command is not specified, then the type will be set to that which is conventionally appropriate for the suffix. If file is omitted, then a file name is automatically generated, with the form input.table$n$.ext: input is the basename of the input file (or molpro if running from standard input); $n$ is a sequence number that is incremented by one each time a table is produced; ext is a suffix appropriate to the file format, eg txt, html, etc.
  • TITLE,title Specify one line of a title (several TITLE cards may follow each other). Note that titles are only displayed in the SAVE file, if the SAVE command is given before the TITLE card.
  • SORT,col1,col2,… Sort rows according to increasing values of the given columns. The columns are sorted in the order they are specified.
  • PRINT,key1,key2,… Specify print options (TABLE, HEADING, TITLE, WARNING, FORMAT, SORT). The default is print for the first three, and noprint for the last three.
  • NOPRINT,key1,key2,… Disable print for given keys.
  • NOPUNCH Don’t write data to the punch file (data are written by default).
  • RANGE,start,end Specify start and end indices of the variables to be printed.
  • STATISTICS Print also linear regression, upper and lower bounding lines, and quadratic fits of the data columns. The slopes and intercepts of these lines are saved in the Molpro variables



Construct input for a plotting program using the table as data. PLOT is a subcommand of TABLE and must follow TABLE or any of its valid subcommands given in the previous section. More than one PLOT command can be included within a single TABLE, and each invocation generates a new plot. However, PLOT must appear after all other TABLE subcommands.

col1, col2,$\dots$ are the names of the table columns to be plotted. These must be an exact subset of those given on the TABLE command. The first column is taken as abscissa, and the values of the remainder will be plotted against it. If no columns are specified, then the entire table is plotted; if a single column is specified, it will be used as abscissa, and all other columns in the table will be plotted as ordinate. options can be chosen from the following.

  • CMD=unix_plot_command unix_plot_command consists of the system command needed to start the plotting program, followed by any required options. The whole thing should normally be enclosed in quotation marks to preserve lower-case letters. The default is ’xmgrace’. At present, the //Grace// program (also known as xmgrace, grace, gracebat), with only numerical data, is supported. The output is also compatible with the portable drop-in replacement for Grace, //AptPlot//, and if Grace is not found on the system, Molpro will attempt to use AptPlot as default instead.
  • FILE=plotfile By default the input file for the plotting program is saved in input.table$n$.plot$m$.agr, where $m$ is an automatically generated sequence number. The name of the plotfile can be modified using the FILE option.
  • INTERACTIVE By default, the plot is not shown on the screen but all plot data are saved in the given file. The plotting program can be started interactively by giving the INTERACTIVE option.
  • TYPE=type If TYPE is specified, type should be set to one of pdf, svg, png, jpeg or eps. The result is that the gracebat program is executed on the plot input file to generate the graph output file in the desired format. This feature depends on the availability of gracebat, and on it supporting the requested output format (for example, at present pdf is supported under Mac OS X, but not in some Linux systems).
  • BACKGROUND=rgb rgb should be a string of six hexadecimal digits specifying the red-green-blue colour to use for the background of the plot.
  • BACKGROUND=TRANSPARENT The background of the plot is made transparent (currently implemented only for TYPE=svg).
  • NOSPLINE Prevents spline interpolation of data points
  • CURVE=type Specifies type of curve to be drawn with points. Possibilities are SPLINE (default; spline interpolation); NONE (equivalent to NOSPLINE); REGRESSION (linear regression line); UPPERBOUNDMAX (maximum gradient line that bounds points from above); UPPERBOUNDMIN (minimum gradient line that bounds points from above); LOWERBOUNDMAX (maximum gradient line that bounds points from below); LOWERBOUNDMIN (minimum gradient line that bounds points from below).
  • NSPLINE=number Number of interpolation points (default 20)
  • LEGEND=’x, y Position legend at $(x,y)$ on plot.
  • LEGEND=OFF Do not draw legend; this behaviour is chosen automatically when there is only a single ordinate dataset.
  • PCOMMAND=’command Insert arbitrary Grace command into the plot file; for details, consult

The following additional directives can be given before the PLOT directive:

  • COLOUR,icolour1, icolour2,… Colour map to be used for columns 1,2,…; zero means to use default values (colours black, blue, red, green cycle)
  • COLOUR,rgb1, rgb2,… Absolute colours (6-hex-digit rgb values) to be used for columns 1,2,…;
  • SYMBOL,isymb1, isymb2,… Symbol types to be used for columns 1,2,…; -1 means no symbols; zero means to use default values.
  • LINEWIDTH,width1, width2,… Line widths to be used for columns 1,2,…; omit to use default values.
  • LINESTYLE,style1, style2,… Line styles to be used for columns 1,2,…; omit to use default values.

For the case that a table contains one or more potential energy functions for a diatomic molecule, with the first column containing bond lengths in Bohr or Ångstrom, it is possible to calculate spectroscopic constants using


The data are fitted to a polynomial of degree $n$ (default is number of points minus 1, ie interpolation), and spectroscopic constants calculated using reduced mass $m$ expressed in u. Note that it is possible to constrain which bond lengths are used through the use of the RANGE subcommand.

D$_{\rm e}$ and D$_0$ are computed from the energy at the fitted minimum distance, and the energy at the longest distance along the potential curve. If the longest distance is less than 10 Å, then D$_{\rm e}$ and D$_0$ are not computed, and instead a value of 0 is printed.