Previous: Command-Line Format of the mpl Programs
Up: User's Manual for the mpl Scientific Toolkit
Next: Using the Toolkit at APS
Previous Page: Command-Line Format of the mpl Programs
Next Page: Using the Toolkit at APS

Limitations of mpl Format

In spite of its convenience and simplicity, mpl format has its limitations. It is easy to think of examples where storing data in a two-column format is inadequate. Even so, when the data is used it is most frequently by reduction to forms that permit graphical display. In most cases, ``graphical display'' means plotting two-column data.

The solution I have adopted is to employ more complicated data formats for storage of such data. Three such formats are available: SDDS format[6], awe format[3], and so-called column format. All of these are self-describing formats for storage of tables of data. Each is accompanied by one or more programs to allow manipulation of the data and extraction of data pairs into mpl format.

awe format uses binary storage of numbers to allow efficient storage of large amounts of data; it will not be discussed further here. column format uses ASCII only, so that files may be easily generated by a program, edited, etc. column format files may also be generated from mpl format files using the program mpl2col (see the manual entry for details).

SDDS format is functionally a superset of both awe and column formats, and should be used for new applications. Conversion programs are available to translate awe and column format data into SDDS format.

The program column is discussed in a manual entry below. As mentioned, the column data format is self-describing, which means that it consists of a descriptive header specifying what data is in the file, followed by the data itself. The header identifies the columns of the data table by name, and also gives units, short descriptions, and a C-style format string for printing the data. The header also optionally contains the names, units, descriptions, and values of so-called ``auxiliary variables''; these are quantities that, while not part of the data table, are relevant to it (e.g., a parameter of the simulation that produced the results).

The data file format is as follows, where exclamation point introduce comments: ! numbers of columns, lines per row, and auxiliary variables:
n_columns n_lines_per_row n_auxiliaries
! backslash-separated names, units, descriptions, and C-style formats for columns
name[0]unit[0]description[0]format[0]
name[1]unit[1]description[1]format[1]
.
.
.
name[n_columns-1]unit[n_columns-1]description[n_columns-1]format[n_columns-1]
! two lines of descriptive text:
description line 1
description line 2
! backslash-separated names, units, descriptions, and values for auxiliary variables:
! (present only if n_auxiliaries is nonzero)
Aname[0]Aunit[0]Adescription[0]Avalue
Aname[1]Aunit[1]Adescription[1]Avalue
.
.
.
Aname[n_auxiliaries-1]Aunit[n_auxiliaries-1]Adescription[n_auxiliaries-1]Avalue
! number of rows of data:
number_of_rows
! first row of table, spread over n_lines_per_row lines:
line[0] of row[0]
line[1] of row[0]
.
.
.
line[n_lines_per_row-1] of row[0]
! second row of table:
line[0] of row[1]
line[1] of row[1]
.
.
.
line[n_lines_per_row-1] of row[1] .
.
.



Previous: Command-Line Format of the mpl Programs
Up: User's Manual for the mpl Scientific Toolkit
Next: Using the Toolkit at APS
Previous Page: Command-Line Format of the mpl Programs
Next Page: Using the Toolkit at APS

borland@aps.anl.gov
Thu Dec 23 15:13:43 CST 1993