Computer Data Types

The computer has different types of data, each with different properties:

• Integer--whole or counting numbers.  Since the computer stores in binary, it is easier to store integers compared to floats, and they take less storage.
• Float  (or floating point or real number)--numbers with a decimal point.  These are actually stored with a mantissa and exponent, and will never be exact: 2 will be 1.99999999 or 2.00000001. (√2)² will often not be 2.
• String or character--just text, can include numbers.  In GIS you cannot do arithmetic on strings.

When a vector data database is created, each field will have a specific type which limits what you can do with the database.  For instance, if your enter "14495 feet" as the elevation for Mount Whitney, the computer will store it as a string, because you cannot mix the number with its units.  You cannot do arithmetic with that data, or color it by elevation since the computer cannot compare "14495 feet" and "13298 feet".  One solution is to put the units in the field name--this could be ELEV_FT.  The other option is to force the user to look up the units in the metadata that should accompany the database.

Raster data sets also have types; the values in the grids can be stored as integers or floats.  This determines the precision of the values, and the required storage.

Last revision 7/22/2015