Table data are arranged in columns and rows, and stored in MIDAS files with the extension .tbl. The entry at a given row and column may be either a single value or an array. The items in one row may describe different properties of the same object or feature. All elements in a given column must be of the same type and thus be associated with the same property. For instance, a table with stellar data could contain the following items in each row: identification, right ascension, declination, magnitude, and spectral type. The first column would then contain all the stellar identifications, the second the right ascensions, etc.
The supported column data types are numerical data (8/16/32-bit integers or 32/64-bit reals) and character strings.
Each column is tagged with a user-defined label, a display format and optional physical units and can be referred to either by its absolute number or its label.
An item in a table is accessed by giving its column and row in addition to the table name. The row number can either be given as an absolute value (i.e. the sequence number) or indicated by the value in a previously defined reference column.
In addition to the normal columns all tables contain a SELECT and a SEQUENCE column:
If an element is not defined it will be a NULL entry and will be listed as a "*" for all data types except for character strings. In that case it will be listed as an empty field.
The tables may be physically stored on disk in two formats: by records corresponding to the natural way of storing sequentially the rows and transposed , where all the values of a given column are stored together( default mode). A table can be always expanded in the sense that its number of columns and rows is automatically increased when the allocated space is exceeded.